joaovissoci / Preference-Mturk

Analysis for the Preference manuscript

1 stars 1 forks source link

create mock tables and graphics #5

Open rpietro opened 12 years ago

rpietro commented 12 years ago

Problem : trying to cross three categorical variables using the crosstable code, but is not working, specifically because it is not recognizing the recoded variable Code and error : CrossTable(Attention, Mturk*new.Age, chisq=TRUE, missing.include= TRUE, format="SAS", prop.r=FALSE)

         | Mturk * new.Age

Attention	NA	Row Total
0	61	61
	0.151
-------------	-----------	-----------
1	343	343
	0.849
-------------	-----------	-----------
Column Total	404	404
-------------	-----------	-----------

Warning message: In Ops.factor(Mturk, new.Age) : * not meaningful for factors

OBS : this code works when used with two recoded variables, and it works with 3 not recoded variables, but doesn't work with 3 variables being 1 recoded

tkooyen commented 12 years ago

Hi Ricardo, I've built these tables based on the analysis we've discussed previously.

external validity in relation to 2010 US census tract (Table 1)
MturkUS censusageunder 35 = 53%under 18 = 24% 18 to 44 = 36.5% 45 to 64 = 26.4% 65 and over = 13% http://goo.gl/BqL7ogendermale 46 %male 49.2% http://goo.gl/slCCLracewhite = 81.8%white 72.4% african american 12.6% native american 0.9% asian 4.8% pacific islander 0.2% some other race 6.2% two or more races 2.9% http://goo.gl/nbP4sethnicityeducation2 year college or more 56.6%population 25 years and older high school graduate or more 83.4% some college or more 50.1% bachelor degree or more 25.2% http://goo.gl/ajPAqincome over 20,000 70.4% http://goo.gl/U7Moe http://goo.gl/ZYnDMmarital status non single 50.6%married 54.4% widowed, divorced or separated 18.5% never married 27.1% http://goo.gl/F6Ukpcomorbiditiesno 63.3%http://goo.gl/8b5Zr
feasibility

yesnoattention question84.2%15.8%understanding100%missing dataaccepted due duration91%9%

*validation MturkHangoutTTOSliderTTOSliderWill to live22.67021.81818.62521.333

these values i got from analyzing the data in excel ... On Jun 15, 2012 11:19 PM, "Ricardo Pietrobon" < reply@reply.github.com> wrote:

Reply to this email directly or view it on GitHub: https://github.com/joaovissoci/Preference-Mturk/issues/5

rpietro commented 12 years ago

got, but what was the specific question for me?

On Thu, Jul 19, 2012 at 3:15 PM, tkooyen < reply@reply.github.com

wrote:

Hi Ricardo, I've built these tables based on the analysis we've discussed previously.

**

external validity in relation to 2010 US census tract (Table 1)

MturkUS censusageunder 35 = 53%under 18 = 24% 18 to 44 = 36.5% 45 to 64 = 26.4% 65 and over = 13% http://goo.gl/BqL7ogendermale 46 %male 49.2% http://goo.gl/slCCLracewhite = 81.8%white 72.4% african american 12.6% native american 0.9% asian 4.8% pacific islander 0.2% some other race 6.2% two or more races 2.9% http://goo.gl/nbP4sethnicityeducation2 year college or more 56.6%population 25 years and older high school graduate or more 83.4% some college or more 50.1% bachelor degree or more 25.2% http://goo.gl/ajPAqincome over 20,000 70.4% http://goo.gl/U7Moe http://goo.gl/ZYnDMmarital status non single 50.6%married 54.4% widowed, divorced or separated 18.5% never married 27.1% http://goo.gl/F6Ukpcomorbiditiesno 63.3%http://goo.gl/8b5Zr

feasibility

yesnoattention question84.2%15.8%understanding100%missing dataaccepted due duration91%9%

*

*validation MturkHangoutTTOSliderTTOSliderWill to live22.67021.81818.62521.333

these values i got from analyzing the data in excel ... On Jun 15, 2012 11:19 PM, "Ricardo Pietrobon" < reply@reply.github.com> wrote:

Reply to this email directly or view it on GitHub: https://github.com/joaovissoci/Preference-Mturk/issues/5

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7109517

tkooyen commented 12 years ago

what are the statistical tests I have to do for these comparisons ?

On Thu, Jul 19, 2012 at 5:17 PM, Ricardo Pietrobon reply@reply.github.com wrote:

got, but what was the specific question for me?

On Thu, Jul 19, 2012 at 3:15 PM, tkooyen < reply@reply.github.com

wrote:

Hi Ricardo, I've built these tables based on the analysis we've discussed previously.

**

external validity in relation to 2010 US census tract (Table 1)

MturkUS censusageunder 35 = 53%under 18 = 24% 18 to 44 = 36.5% 45 to 64 = 26.4% 65 and over = 13% http://goo.gl/BqL7ogendermale 46 %male 49.2% http://goo.gl/slCCLracewhite = 81.8%white 72.4% african american 12.6% native american 0.9% asian 4.8% pacific islander 0.2% some other race 6.2% two or more races 2.9% http://goo.gl/nbP4sethnicityeducation2 year college or more 56.6%population 25 years and older high school graduate or more 83.4% some college or more 50.1% bachelor degree or more 25.2% http://goo.gl/ajPAqincome over 20,000 70.4% http://goo.gl/U7Moe http://goo.gl/ZYnDMmarital status non single 50.6%married 54.4% widowed, divorced or separated 18.5% never married 27.1% http://goo.gl/F6Ukpcomorbiditiesno 63.3%http://goo.gl/8b5Zr

feasibility

yesnoattention question84.2%15.8%understanding100%missing dataaccepted due duration91%9%

*

*validation MturkHangoutTTOSliderTTOSliderWill to live22.67021.81818.62521.333

these values i got from analyzing the data in excel ... On Jun 15, 2012 11:19 PM, "Ricardo Pietrobon" < reply@reply.github.com> wrote:

Reply to this email directly or view it on GitHub: https://github.com/joaovissoci/Preference-Mturk/issues/5

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7109517

Reply to this email directly or view it on GitHub: https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114035

rpietro commented 12 years ago

Talitha, could you just list side by side the variable names of what you are trying to compare? i tried to go through the code but i am somewhat confused

On Thu, Jul 19, 2012 at 4:20 PM, tkooyen < reply@reply.github.com

wrote:

what are the statistical tests I have to do for these comparisons ?

On Thu, Jul 19, 2012 at 5:17 PM, Ricardo Pietrobon reply@reply.github.com wrote:

got, but what was the specific question for me?

On Thu, Jul 19, 2012 at 3:15 PM, tkooyen < reply@reply.github.com

wrote:

Hi Ricardo, I've built these tables based on the analysis we've discussed previously.

**

external validity in relation to 2010 US census tract (Table 1)

MturkUS censusageunder 35 = 53%under 18 = 24% 18 to 44 = 36.5% 45 to 64 = 26.4% 65 and over = 13% http://goo.gl/BqL7ogendermale 46 %male 49.2% http://goo.gl/slCCLracewhite = 81.8%white 72.4% african american 12.6% native american 0.9% asian 4.8% pacific islander 0.2% some other race 6.2% two or more races 2.9% http://goo.gl/nbP4sethnicityeducation2 year college or more 56.6%population 25 years and older high school graduate or more 83.4% some college or more 50.1% bachelor degree or more 25.2% http://goo.gl/ajPAqincome over 20,000 70.4% http://goo.gl/U7Moe http://goo.gl/ZYnDMmarital status non single 50.6%married 54.4% widowed, divorced or separated 18.5% never married 27.1% http://goo.gl/F6Ukpcomorbiditiesno 63.3%http://goo.gl/8b5Zr

feasibility

yesnoattention question84.2%15.8%understanding100%missing dataaccepted due duration91%9%

*

*validation MturkHangoutTTOSliderTTOSliderWill to live22.67021.81818.62521.333

these values i got from analyzing the data in excel ... On Jun 15, 2012 11:19 PM, "Ricardo Pietrobon" < reply@reply.github.com> wrote:

Reply to this email directly or view it on GitHub: https://github.com/joaovissoci/Preference-Mturk/issues/5

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7109517

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114035

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114104

tkooyen commented 12 years ago

the code is really confused because i was doing a bunch of tests to understand how the program worked i'm going to take down the codes that i did that have nothing to do with what i'm trying to analyze than i'll send you a list of the variable names

On Thu, Jul 19, 2012 at 7:35 PM, Ricardo Pietrobon reply@reply.github.com wrote:

Talitha, could you just list side by side the variable names of what you are trying to compare? i tried to go through the code but i am somewhat confused

On Thu, Jul 19, 2012 at 4:20 PM, tkooyen < reply@reply.github.com

wrote:

what are the statistical tests I have to do for these comparisons ?

On Thu, Jul 19, 2012 at 5:17 PM, Ricardo Pietrobon reply@reply.github.com wrote:

got, but what was the specific question for me?

On Thu, Jul 19, 2012 at 3:15 PM, tkooyen < reply@reply.github.com

wrote:

Hi Ricardo, I've built these tables based on the analysis we've discussed previously.

**

external validity in relation to 2010 US census tract (Table 1)

MturkUS censusageunder 35 = 53%under 18 = 24% 18 to 44 = 36.5% 45 to 64 = 26.4% 65 and over = 13% http://goo.gl/BqL7ogendermale 46 %male 49.2% http://goo.gl/slCCLracewhite = 81.8%white 72.4% african american 12.6% native american 0.9% asian 4.8% pacific islander 0.2% some other race 6.2% two or more races 2.9% http://goo.gl/nbP4sethnicityeducation2 year college or more 56.6%population 25 years and older high school graduate or more 83.4% some college or more 50.1% bachelor degree or more 25.2% http://goo.gl/ajPAqincome over 20,000 70.4% http://goo.gl/U7Moe http://goo.gl/ZYnDMmarital status non single 50.6%married 54.4% widowed, divorced or separated 18.5% never married 27.1% http://goo.gl/F6Ukpcomorbiditiesno 63.3%http://goo.gl/8b5Zr

feasibility

yesnoattention question84.2%15.8%understanding100%missing dataaccepted due duration91%9%

*

*validation MturkHangoutTTOSliderTTOSliderWill to live22.67021.81818.62521.333

these values i got from analyzing the data in excel ... On Jun 15, 2012 11:19 PM, "Ricardo Pietrobon" < reply@reply.github.com> wrote:

Reply to this email directly or view it on GitHub: https://github.com/joaovissoci/Preference-Mturk/issues/5

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7109517

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114035

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114104

Reply to this email directly or view it on GitHub: https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7117405

rpietro commented 12 years ago

perfect, in that way i can get into your code and add examples, and then create some other instructional material for you to interpret the results that will look like http://goo.gl/jKPjy

On Thu, Jul 19, 2012 at 8:18 PM, tkooyen < reply@reply.github.com

wrote:

the code is really confused because i was doing a bunch of tests to understand how the program worked i'm going to take down the codes that i did that have nothing to do with what i'm trying to analyze than i'll send you a list of the variable names

On Thu, Jul 19, 2012 at 7:35 PM, Ricardo Pietrobon reply@reply.github.com wrote:

Talitha, could you just list side by side the variable names of what you are trying to compare? i tried to go through the code but i am somewhat confused

On Thu, Jul 19, 2012 at 4:20 PM, tkooyen < reply@reply.github.com

wrote:

what are the statistical tests I have to do for these comparisons ?

On Thu, Jul 19, 2012 at 5:17 PM, Ricardo Pietrobon reply@reply.github.com wrote:

got, but what was the specific question for me?

On Thu, Jul 19, 2012 at 3:15 PM, tkooyen < reply@reply.github.com

wrote:

Hi Ricardo, I've built these tables based on the analysis we've discussed previously.

**

external validity in relation to 2010 US census tract (Table 1)

MturkUS censusageunder 35 = 53%under 18 = 24% 18 to 44 = 36.5% 45 to 64 = 26.4% 65 and over = 13% http://goo.gl/BqL7ogendermale 46 %male 49.2% http://goo.gl/slCCLracewhite = 81.8%white 72.4% african american 12.6% native american 0.9% asian 4.8% pacific islander 0.2% some other race 6.2% two or more races 2.9% http://goo.gl/nbP4sethnicityeducation2 year college or more 56.6%population 25 years and older high school graduate or more 83.4% some college or more 50.1% bachelor degree or more 25.2% http://goo.gl/ajPAqincome over 20,000 70.4% http://goo.gl/U7Moe http://goo.gl/ZYnDMmarital status non single 50.6%married 54.4% widowed, divorced or separated 18.5% never married 27.1% http://goo.gl/F6Ukpcomorbiditiesno 63.3%http://goo.gl/8b5Zr

feasibility

yesnoattention question84.2%15.8%understanding100%missing dataaccepted due duration91%9%

*

*validation MturkHangoutTTOSliderTTOSliderWill to live22.67021.81818.62521.333

these values i got from analyzing the data in excel ... On Jun 15, 2012 11:19 PM, "Ricardo Pietrobon" < reply@reply.github.com> wrote:

Reply to this email directly or view it on GitHub: https://github.com/joaovissoci/Preference-Mturk/issues/5

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7109517

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114035

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114104

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7117405

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119175

tkooyen commented 12 years ago

so, I've already erased all my tests what I want to analyze in my first table is the demographic aspect of the Mturk population compared to the US census the only problem is the database for demographic variables like age, gender, etc ... includes not only the Mturk population but the hangout population as well, so before doing the analysis i've "filtered" the Mturk population the variables for this first table are age gender race education income marital status comorbidities the US census data is not in the database, if needed tell me how to include this data into the database

On Thu, Jul 19, 2012 at 9:20 PM, Ricardo Pietrobon reply@reply.github.com wrote:

perfect, in that way i can get into your code and add examples, and then create some other instructional material for you to interpret the results that will look like http://goo.gl/jKPjy

On Thu, Jul 19, 2012 at 8:18 PM, tkooyen < reply@reply.github.com

wrote:

the code is really confused because i was doing a bunch of tests to understand how the program worked i'm going to take down the codes that i did that have nothing to do with what i'm trying to analyze than i'll send you a list of the variable names

On Thu, Jul 19, 2012 at 7:35 PM, Ricardo Pietrobon reply@reply.github.com wrote:

Talitha, could you just list side by side the variable names of what you are trying to compare? i tried to go through the code but i am somewhat confused

On Thu, Jul 19, 2012 at 4:20 PM, tkooyen < reply@reply.github.com

wrote:

what are the statistical tests I have to do for these comparisons ?

On Thu, Jul 19, 2012 at 5:17 PM, Ricardo Pietrobon reply@reply.github.com wrote:

got, but what was the specific question for me?

On Thu, Jul 19, 2012 at 3:15 PM, tkooyen < reply@reply.github.com

wrote:

Hi Ricardo, I've built these tables based on the analysis we've discussed previously.

**

external validity in relation to 2010 US census tract (Table 1)

MturkUS censusageunder 35 = 53%under 18 = 24% 18 to 44 = 36.5% 45 to 64 = 26.4% 65 and over = 13% http://goo.gl/BqL7ogendermale 46 %male 49.2% http://goo.gl/slCCLracewhite = 81.8%white 72.4% african american 12.6% native american 0.9% asian 4.8% pacific islander 0.2% some other race 6.2% two or more races 2.9% http://goo.gl/nbP4sethnicityeducation2 year college or more 56.6%population 25 years and older high school graduate or more 83.4% some college or more 50.1% bachelor degree or more 25.2% http://goo.gl/ajPAqincome over 20,000 70.4% http://goo.gl/U7Moe http://goo.gl/ZYnDMmarital status non single 50.6%married 54.4% widowed, divorced or separated 18.5% never married 27.1% http://goo.gl/F6Ukpcomorbiditiesno 63.3%http://goo.gl/8b5Zr

feasibility

yesnoattention question84.2%15.8%understanding100%missing dataaccepted due duration91%9%

*

*validation MturkHangoutTTOSliderTTOSliderWill to live22.67021.81818.62521.333

these values i got from analyzing the data in excel ... On Jun 15, 2012 11:19 PM, "Ricardo Pietrobon" < reply@reply.github.com> wrote:

Reply to this email directly or view it on GitHub: https://github.com/joaovissoci/Preference-Mturk/issues/5

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7109517

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114035

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114104

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7117405

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119175

Reply to this email directly or view it on GitHub: https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119215

rpietro commented 12 years ago

ok, so just to make sure i understand the question. how do you include census info: would put the average values for each the variables that are in your data set and census. also, check how buhrmester and others have done this

Do you need the commands to get summary statistics for the variables below?

if so, did you check the section below -- it's in your script. Talitha, we probably need to set a time to go over this. please shoot me an invitation

###########################################################################################

TABLE 1: DEMOGRAPHICS

###########################################################################################

describes your entire dataset

describe(templateData)

summary(variable) qplot(variable)

t.test, where outcome is a continuous variable and predictor is a

dichotomous variable t.test(outcome~predictor)

chi square test where both outcome and predictor are categorical variables

CrossTable(outcome, predictor, chisq=TRUE, missing.include=TRUE, format="SAS", prop.r=FALSE)

On Mon, Jul 23, 2012 at 11:59 AM, tkooyen < reply@reply.github.com

wrote:

so, I've already erased all my tests what I want to analyze in my first table is the demographic aspect of the Mturk population compared to the US census the only problem is the database for demographic variables like age, gender, etc ... includes not only the Mturk population but the hangout population as well, so before doing the analysis i've "filtered" the Mturk population the variables for this first table are age gender race education income marital status comorbidities the US census data is not in the database, if needed tell me how to include this data into the database

On Thu, Jul 19, 2012 at 9:20 PM, Ricardo Pietrobon reply@reply.github.com wrote:

perfect, in that way i can get into your code and add examples, and then create some other instructional material for you to interpret the results that will look like http://goo.gl/jKPjy

On Thu, Jul 19, 2012 at 8:18 PM, tkooyen < reply@reply.github.com

wrote:

the code is really confused because i was doing a bunch of tests to understand how the program worked i'm going to take down the codes that i did that have nothing to do with what i'm trying to analyze than i'll send you a list of the variable names

On Thu, Jul 19, 2012 at 7:35 PM, Ricardo Pietrobon reply@reply.github.com wrote:

Talitha, could you just list side by side the variable names of what you are trying to compare? i tried to go through the code but i am somewhat confused

On Thu, Jul 19, 2012 at 4:20 PM, tkooyen < reply@reply.github.com

wrote:

what are the statistical tests I have to do for these comparisons ?

On Thu, Jul 19, 2012 at 5:17 PM, Ricardo Pietrobon reply@reply.github.com wrote:

got, but what was the specific question for me?

On Thu, Jul 19, 2012 at 3:15 PM, tkooyen < reply@reply.github.com

wrote:

Hi Ricardo, I've built these tables based on the analysis we've discussed previously.

**

external validity in relation to 2010 US census tract (Table 1)

MturkUS censusageunder 35 = 53%under 18 = 24% 18 to 44 = 36.5% 45 to 64 = 26.4% 65 and over = 13% http://goo.gl/BqL7ogendermale 46 %male 49.2% http://goo.gl/slCCLracewhite = 81.8%white 72.4% african american 12.6% native american 0.9% asian 4.8% pacific islander 0.2% some other race 6.2% two or more races 2.9% http://goo.gl/nbP4sethnicityeducation2 year college or more 56.6%population 25 years and older high school graduate or more 83.4% some college or more 50.1% bachelor degree or more 25.2% http://goo.gl/ajPAqincome over 20,000 70.4% http://goo.gl/U7Moe http://goo.gl/ZYnDMmarital status non single 50.6%married 54.4% widowed, divorced or separated 18.5% never married 27.1% http://goo.gl/F6Ukpcomorbiditiesno 63.3%http://goo.gl/8b5Zr

feasibility

yesnoattention question84.2%15.8%understanding100%missing dataaccepted due duration91%9%

*

*validation MturkHangoutTTOSliderTTOSliderWill to live22.67021.81818.62521.333

these values i got from analyzing the data in excel ... On Jun 15, 2012 11:19 PM, "Ricardo Pietrobon" < reply@reply.github.com> wrote:

Reply to this email directly or view it on GitHub: https://github.com/joaovissoci/Preference-Mturk/issues/5

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7109517

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114035

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114104

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7117405

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119175

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119215

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7182205

tkooyen commented 12 years ago

ok, i'll upload the census data into the data set the way you've described

i've already worked around the commands that you've sent below but just to make sure i understand, the t test and the chisquare test are ways to filter and statistically analyze the dataset, right ? i'll use one or the other dependig if my variable is continuous or categorical ...

i'm getting the hang of it, just give me some time ... the problem is I have to stop to think that I'm in excel ! if i don't get it right by the end of today, i'll send you an invite

On Mon, Jul 23, 2012 at 1:28 PM, Ricardo Pietrobon reply@reply.github.com wrote:

ok, so just to make sure i understand the question. how do you include census info: would put the average values for each the variables that are in your data set and census. also, check how buhrmester and others have done this

Do you need the commands to get summary statistics for the variables below?

if so, did you check the section below -- it's in your script. Talitha, we probably need to set a time to go over this. please shoot me an invitation

###########################################################################################

TABLE 1: DEMOGRAPHICS

###########################################################################################

describes your entire dataset

describe(templateData)

summary(variable) qplot(variable)

t.test, where outcome is a continuous variable and predictor is a

dichotomous variable t.test(outcome~predictor)

chi square test where both outcome and predictor are categorical variables

CrossTable(outcome, predictor, chisq=TRUE, missing.include=TRUE, format="SAS", prop.r=FALSE)

On Mon, Jul 23, 2012 at 11:59 AM, tkooyen < reply@reply.github.com

wrote:

so, I've already erased all my tests what I want to analyze in my first table is the demographic aspect of the Mturk population compared to the US census the only problem is the database for demographic variables like age, gender, etc ... includes not only the Mturk population but the hangout population as well, so before doing the analysis i've "filtered" the Mturk population the variables for this first table are age gender race education income marital status comorbidities the US census data is not in the database, if needed tell me how to include this data into the database

On Thu, Jul 19, 2012 at 9:20 PM, Ricardo Pietrobon reply@reply.github.com wrote:

perfect, in that way i can get into your code and add examples, and then create some other instructional material for you to interpret the results that will look like http://goo.gl/jKPjy

On Thu, Jul 19, 2012 at 8:18 PM, tkooyen < reply@reply.github.com

wrote:

the code is really confused because i was doing a bunch of tests to understand how the program worked i'm going to take down the codes that i did that have nothing to do with what i'm trying to analyze than i'll send you a list of the variable names

On Thu, Jul 19, 2012 at 7:35 PM, Ricardo Pietrobon reply@reply.github.com wrote:

Talitha, could you just list side by side the variable names of what you are trying to compare? i tried to go through the code but i am somewhat confused

On Thu, Jul 19, 2012 at 4:20 PM, tkooyen < reply@reply.github.com

wrote:

what are the statistical tests I have to do for these comparisons ?

On Thu, Jul 19, 2012 at 5:17 PM, Ricardo Pietrobon reply@reply.github.com wrote:

got, but what was the specific question for me?

On Thu, Jul 19, 2012 at 3:15 PM, tkooyen < reply@reply.github.com

wrote:

Hi Ricardo, I've built these tables based on the analysis we've discussed previously.

**

external validity in relation to 2010 US census tract (Table 1)

MturkUS censusageunder 35 = 53%under 18 = 24% 18 to 44 = 36.5% 45 to 64 = 26.4% 65 and over = 13% http://goo.gl/BqL7ogendermale 46 %male 49.2% http://goo.gl/slCCLracewhite = 81.8%white 72.4% african american 12.6% native american 0.9% asian 4.8% pacific islander 0.2% some other race 6.2% two or more races 2.9% http://goo.gl/nbP4sethnicityeducation2 year college or more 56.6%population 25 years and older high school graduate or more 83.4% some college or more 50.1% bachelor degree or more 25.2% http://goo.gl/ajPAqincome over 20,000 70.4% http://goo.gl/U7Moe http://goo.gl/ZYnDMmarital status non single 50.6%married 54.4% widowed, divorced or separated 18.5% never married 27.1% http://goo.gl/F6Ukpcomorbiditiesno 63.3%http://goo.gl/8b5Zr

feasibility

yesnoattention question84.2%15.8%understanding100%missing dataaccepted due duration91%9%

*

*validation MturkHangoutTTOSliderTTOSliderWill to live22.67021.81818.62521.333

these values i got from analyzing the data in excel ... On Jun 15, 2012 11:19 PM, "Ricardo Pietrobon" < reply@reply.github.com> wrote:

Reply to this email directly or view it on GitHub: https://github.com/joaovissoci/Preference-Mturk/issues/5

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7109517

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114035

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114104

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7117405

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119175

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119215

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7182205

Reply to this email directly or view it on GitHub: https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183045

rpietro commented 12 years ago

so, i would suggest that you post your questions using a three-point structure, just to make sure i can answer exactly what you need:

what the problem is
which code you used
what error message you got

would also be good for us to meet, can you shoot me an invitation?

On Mon, Jul 23, 2012 at 12:40 PM, tkooyen < reply@reply.github.com

wrote:

ok, i'll upload the census data into the data set the way you've described

i've already worked around the commands that you've sent below but just to make sure i understand, the t test and the chisquare test are ways to filter and statistically analyze the dataset, right ? i'll use one or the other dependig if my variable is continuous or categorical ...

i'm getting the hang of it, just give me some time ... the problem is I have to stop to think that I'm in excel ! if i don't get it right by the end of today, i'll send you an invite

On Mon, Jul 23, 2012 at 1:28 PM, Ricardo Pietrobon reply@reply.github.com wrote:

ok, so just to make sure i understand the question. how do you include census info: would put the average values for each the variables that are in your data set and census. also, check how buhrmester and others have done this

Do you need the commands to get summary statistics for the variables below?

if so, did you check the section below -- it's in your script. Talitha, we probably need to set a time to go over this. please shoot me an invitation

###########################################################################################

TABLE 1: DEMOGRAPHICS

###########################################################################################

describes your entire dataset

describe(templateData)

summary(variable) qplot(variable)

t.test, where outcome is a continuous variable and predictor is a

dichotomous variable t.test(outcome~predictor)

chi square test where both outcome and predictor are categorical

variables CrossTable(outcome, predictor, chisq=TRUE, missing.include=TRUE, format="SAS", prop.r=FALSE)

On Mon, Jul 23, 2012 at 11:59 AM, tkooyen < reply@reply.github.com

wrote:

so, I've already erased all my tests what I want to analyze in my first table is the demographic aspect of the Mturk population compared to the US census the only problem is the database for demographic variables like age, gender, etc ... includes not only the Mturk population but the hangout population as well, so before doing the analysis i've "filtered" the Mturk population the variables for this first table are age gender race education income marital status comorbidities the US census data is not in the database, if needed tell me how to include this data into the database

On Thu, Jul 19, 2012 at 9:20 PM, Ricardo Pietrobon reply@reply.github.com wrote:

perfect, in that way i can get into your code and add examples, and then create some other instructional material for you to interpret the results that will look like http://goo.gl/jKPjy

On Thu, Jul 19, 2012 at 8:18 PM, tkooyen < reply@reply.github.com

wrote:

the code is really confused because i was doing a bunch of tests to understand how the program worked i'm going to take down the codes that i did that have nothing to do with what i'm trying to analyze than i'll send you a list of the variable names

On Thu, Jul 19, 2012 at 7:35 PM, Ricardo Pietrobon reply@reply.github.com wrote:

Talitha, could you just list side by side the variable names of what you are trying to compare? i tried to go through the code but i am somewhat confused

On Thu, Jul 19, 2012 at 4:20 PM, tkooyen < reply@reply.github.com

wrote:

what are the statistical tests I have to do for these comparisons ?

On Thu, Jul 19, 2012 at 5:17 PM, Ricardo Pietrobon reply@reply.github.com wrote:

got, but what was the specific question for me?

On Thu, Jul 19, 2012 at 3:15 PM, tkooyen < reply@reply.github.com

wrote:

Hi Ricardo, I've built these tables based on the analysis we've discussed previously.

**

external validity in relation to 2010 US census tract (Table 1)

MturkUS censusageunder 35 = 53%under 18 = 24% 18 to 44 = 36.5% 45 to 64 = 26.4% 65 and over = 13% http://goo.gl/BqL7ogendermale 46 %male 49.2% http://goo.gl/slCCLracewhite = 81.8%white 72.4% african american 12.6% native american 0.9% asian 4.8% pacific islander 0.2% some other race 6.2% two or more races 2.9% http://goo.gl/nbP4sethnicityeducation2 year college or more 56.6%population 25 years and older high school graduate or more 83.4% some college or more 50.1% bachelor degree or more 25.2% http://goo.gl/ajPAqincome over 20,000 70.4% http://goo.gl/U7Moe http://goo.gl/ZYnDMmarital status non single 50.6%married 54.4% widowed, divorced or separated 18.5% never married 27.1% http://goo.gl/F6Ukpcomorbiditiesno 63.3%http://goo.gl/8b5Zr

feasibility

yesnoattention question84.2%15.8%understanding100%missing dataaccepted due duration91%9%

*

*validation MturkHangoutTTOSliderTTOSliderWill to live22.67021.81818.62521.333

these values i got from analyzing the data in excel ... On Jun 15, 2012 11:19 PM, "Ricardo Pietrobon" < reply@reply.github.com> wrote:

Reply to this email directly or view it on GitHub: https://github.com/joaovissoci/Preference-Mturk/issues/5

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7109517

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114035

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114104

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7117405

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119175

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119215

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7182205

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183045

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183349

tkooyen commented 12 years ago

ok,

problem : I want to compare the variable WillToLive between the TTO and Slider, and between the MTurk and Hangout
code I'm using :

new.Mturk <-car::recode(Mturk, "1 = 'yes'; NA = 'no'") t.test(WillToLive~new.Mturk)

Welch Two Sample t-test

data: WillToLive by new.Mturk t = -1.0914, df = 20.827, p-value = 0.2876 alternative hypothesis: true difference in means is not equal to 0 95 percent confidence interval: -5.829770 1.818218 sample estimates: mean in group no mean in group yes 20.05882 22.06460

there is no error, but this data gives me the mean of all the Turkers that answered the questionnaire regardless if they've answered the TTO or the Slider

On Mon, Jul 23, 2012 at 3:59 PM, Ricardo Pietrobon reply@reply.github.com wrote:

so, i would suggest that you post your questions using a three-point structure, just to make sure i can answer exactly what you need:

what the problem is

which code you used

what error message you got

would also be good for us to meet, can you shoot me an invitation?

On Mon, Jul 23, 2012 at 12:40 PM, tkooyen < reply@reply.github.com

wrote:

ok, i'll upload the census data into the data set the way you've described

i've already worked around the commands that you've sent below but just to make sure i understand, the t test and the chisquare test are ways to filter and statistically analyze the dataset, right ? i'll use one or the other dependig if my variable is continuous or categorical ...

i'm getting the hang of it, just give me some time ... the problem is I have to stop to think that I'm in excel ! if i don't get it right by the end of today, i'll send you an invite

On Mon, Jul 23, 2012 at 1:28 PM, Ricardo Pietrobon reply@reply.github.com wrote:

ok, so just to make sure i understand the question. how do you include census info: would put the average values for each the variables that are in your data set and census. also, check how buhrmester and others have done this

Do you need the commands to get summary statistics for the variables below?

if so, did you check the section below -- it's in your script. Talitha, we probably need to set a time to go over this. please shoot me an invitation

###########################################################################################

TABLE 1: DEMOGRAPHICS

###########################################################################################

describes your entire dataset

describe(templateData)

summary(variable) qplot(variable)

t.test, where outcome is a continuous variable and predictor is a

dichotomous variable t.test(outcome~predictor)

chi square test where both outcome and predictor are categorical

variables CrossTable(outcome, predictor, chisq=TRUE, missing.include=TRUE, format="SAS", prop.r=FALSE)

On Mon, Jul 23, 2012 at 11:59 AM, tkooyen < reply@reply.github.com

wrote:

so, I've already erased all my tests what I want to analyze in my first table is the demographic aspect of the Mturk population compared to the US census the only problem is the database for demographic variables like age, gender, etc ... includes not only the Mturk population but the hangout population as well, so before doing the analysis i've "filtered" the Mturk population the variables for this first table are age gender race education income marital status comorbidities the US census data is not in the database, if needed tell me how to include this data into the database

On Thu, Jul 19, 2012 at 9:20 PM, Ricardo Pietrobon reply@reply.github.com wrote:

perfect, in that way i can get into your code and add examples, and then create some other instructional material for you to interpret the results that will look like http://goo.gl/jKPjy

On Thu, Jul 19, 2012 at 8:18 PM, tkooyen < reply@reply.github.com

wrote:

the code is really confused because i was doing a bunch of tests to understand how the program worked i'm going to take down the codes that i did that have nothing to do with what i'm trying to analyze than i'll send you a list of the variable names

On Thu, Jul 19, 2012 at 7:35 PM, Ricardo Pietrobon reply@reply.github.com wrote:

Talitha, could you just list side by side the variable names of what you are trying to compare? i tried to go through the code but i am somewhat confused

On Thu, Jul 19, 2012 at 4:20 PM, tkooyen < reply@reply.github.com

wrote:

what are the statistical tests I have to do for these comparisons ?

On Thu, Jul 19, 2012 at 5:17 PM, Ricardo Pietrobon reply@reply.github.com wrote:

got, but what was the specific question for me?

On Thu, Jul 19, 2012 at 3:15 PM, tkooyen < reply@reply.github.com

wrote:

Hi Ricardo, I've built these tables based on the analysis we've discussed previously.

**

external validity in relation to 2010 US census tract (Table 1)

MturkUS censusageunder 35 = 53%under 18 = 24% 18 to 44 = 36.5% 45 to 64 = 26.4% 65 and over = 13% http://goo.gl/BqL7ogendermale 46 %male 49.2% http://goo.gl/slCCLracewhite = 81.8%white 72.4% african american 12.6% native american 0.9% asian 4.8% pacific islander 0.2% some other race 6.2% two or more races 2.9% http://goo.gl/nbP4sethnicityeducation2 year college or more 56.6%population 25 years and older high school graduate or more 83.4% some college or more 50.1% bachelor degree or more 25.2% http://goo.gl/ajPAqincome over 20,000 70.4% http://goo.gl/U7Moe http://goo.gl/ZYnDMmarital status non single 50.6%married 54.4% widowed, divorced or separated 18.5% never married 27.1% http://goo.gl/F6Ukpcomorbiditiesno 63.3%http://goo.gl/8b5Zr

feasibility

yesnoattention question84.2%15.8%understanding100%missing dataaccepted due duration91%9%

*

*validation MturkHangoutTTOSliderTTOSliderWill to live22.67021.81818.62521.333

these values i got from analyzing the data in excel ... On Jun 15, 2012 11:19 PM, "Ricardo Pietrobon" < reply@reply.github.com> wrote:

Reply to this email directly or view it on GitHub: https://github.com/joaovissoci/Preference-Mturk/issues/5

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7109517

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114035

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114104

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7117405

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119175

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119215

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7182205

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183045

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183349

Reply to this email directly or view it on GitHub: https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7187558

tkooyen commented 12 years ago

other doubt

problem : need help to interpret the values on the chi-square test
for example the chi-square test for age (I want to know what are the values in red)

CrossTable(Age, Mturk, chisq=TRUE, missing.include=TRUE, format="SAS", prop.r=FALSE)

Cell Contents	-------------------------		N		Chi-square contribution		N / Col Total		N / Table Total

Total Observations in Table: 404

         | Mturk
     Age |         1 |        NA | Row Total |

-------------	-----------	-----------	-----------
3	116	8	124
0.065	1.483
0.300	0.471
0.287	0.020
-------------	-----------	-----------	-----------
4	73	7	80
0.172	3.922
0.189	0.412
0.181	0.017
-------------	-----------	-----------	-----------
5	65	0	65
0.120	2.735
0.168	0.000
0.161	0.000
-------------	-----------	-----------	-----------
6	36	1	37
0.009	0.199
0.093	0.059
0.089	0.002
-------------	-----------	-----------	-----------
7	4	0	4
0.007	0.168
0.010	0.000
0.010	0.000
-------------	-----------	-----------	-----------
8	1	0	1
0.002	0.042
0.003	0.000
0.002	0.000
-------------	-----------	-----------	-----------
NA	2	0	2
0.004	0.084
0.005	0.000
0.005	0.000
-------------	-----------	-----------	-----------
Column Total	387	17	404
0.958	0.042
-------------	-----------	-----------	-----------

Statistics for All Table Factors

Pearson's Chi-squared test

Chi^2 = 11.1961 d.f. = 7 p = 0.130291

On Mon, Jul 30, 2012 at 3:34 PM, Talitha Yen tkooyen@gmail.com wrote:

ok,

problem : I want to compare the variable WillToLive between the TTO and Slider, and between the MTurk and Hangout

code I'm using :

new.Mturk <-car::recode(Mturk, "1 = 'yes'; NA = 'no'") t.test(WillToLive~new.Mturk)
    Welch Two Sample t-test
data: WillToLive by new.Mturk t = -1.0914, df = 20.827, p-value = 0.2876 alternative hypothesis: true difference in means is not equal to 0 95 percent confidence interval: -5.829770 1.818218 sample estimates: mean in group no mean in group yes 20.05882 22.06460

there is no error, but this data gives me the mean of all the Turkers that answered the questionnaire regardless if they've answered the TTO or the Slider

On Mon, Jul 23, 2012 at 3:59 PM, Ricardo Pietrobon < reply@reply.github.com

wrote:

so, i would suggest that you post your questions using a three-point structure, just to make sure i can answer exactly what you need:

what the problem is

which code you used

what error message you got

would also be good for us to meet, can you shoot me an invitation?

On Mon, Jul 23, 2012 at 12:40 PM, tkooyen < reply@reply.github.com

wrote:

ok, i'll upload the census data into the data set the way you've described

i've already worked around the commands that you've sent below but just to make sure i understand, the t test and the chisquare test are ways to filter and statistically analyze the dataset, right ? i'll use one or the other dependig if my variable is continuous or categorical ...

i'm getting the hang of it, just give me some time ... the problem is I have to stop to think that I'm in excel ! if i don't get it right by the end of today, i'll send you an invite

On Mon, Jul 23, 2012 at 1:28 PM, Ricardo Pietrobon reply@reply.github.com wrote:

ok, so just to make sure i understand the question. how do you include census info: would put the average values for each the variables that are in your data set and census. also, check how buhrmester and others have done this

Do you need the commands to get summary statistics for the variables below?

if so, did you check the section below -- it's in your script. Talitha, we probably need to set a time to go over this. please shoot me an invitation

###########################################################################################

TABLE 1: DEMOGRAPHICS

###########################################################################################

describes your entire dataset

describe(templateData)

summary(variable) qplot(variable)

t.test, where outcome is a continuous variable and predictor is a

dichotomous variable t.test(outcome~predictor)

chi square test where both outcome and predictor are categorical

variables CrossTable(outcome, predictor, chisq=TRUE, missing.include=TRUE, format="SAS", prop.r=FALSE)

On Mon, Jul 23, 2012 at 11:59 AM, tkooyen < reply@reply.github.com

wrote:

so, I've already erased all my tests what I want to analyze in my first table is the demographic aspect of the Mturk population compared to the US census the only problem is the database for demographic variables like age, gender, etc ... includes not only the Mturk population but the hangout population as well, so before doing the analysis i've "filtered" the Mturk population the variables for this first table are age gender race education income marital status comorbidities the US census data is not in the database, if needed tell me how to include this data into the database

On Thu, Jul 19, 2012 at 9:20 PM, Ricardo Pietrobon reply@reply.github.com wrote:

perfect, in that way i can get into your code and add examples, and then create some other instructional material for you to interpret the results that will look like http://goo.gl/jKPjy

On Thu, Jul 19, 2012 at 8:18 PM, tkooyen < reply@reply.github.com

wrote:

the code is really confused because i was doing a bunch of tests to understand how the program worked i'm going to take down the codes that i did that have nothing to do with what i'm trying to analyze than i'll send you a list of the variable names

On Thu, Jul 19, 2012 at 7:35 PM, Ricardo Pietrobon reply@reply.github.com wrote:

Talitha, could you just list side by side the variable names of what you are trying to compare? i tried to go through the code but i am somewhat confused

On Thu, Jul 19, 2012 at 4:20 PM, tkooyen < reply@reply.github.com

wrote:

what are the statistical tests I have to do for these comparisons ?

On Thu, Jul 19, 2012 at 5:17 PM, Ricardo Pietrobon reply@reply.github.com wrote:

got, but what was the specific question for me?

On Thu, Jul 19, 2012 at 3:15 PM, tkooyen < reply@reply.github.com

wrote:

Hi Ricardo, I've built these tables based on the analysis we've discussed previously.

**

external validity in relation to 2010 US census tract (Table 1)

MturkUS censusageunder 35 = 53%under 18 = 24% 18 to 44 = 36.5% 45 to 64 = 26.4% 65 and over = 13% http://goo.gl/BqL7ogendermale 46 %male 49.2% http://goo.gl/slCCLracewhite = 81.8%white 72.4% african american 12.6% native american 0.9% asian 4.8% pacific islander 0.2% some other race 6.2% two or more races 2.9% http://goo.gl/nbP4sethnicityeducation2 year college or more 56.6%population 25 years and older high school graduate or more 83.4% some college or more 50.1% bachelor degree or more 25.2% http://goo.gl/ajPAqincome over 20,000 70.4% http://goo.gl/U7Moe http://goo.gl/ZYnDMmarital status non single 50.6%married 54.4% widowed, divorced or separated 18.5% never married 27.1% http://goo.gl/F6Ukpcomorbiditiesno 63.3%http://goo.gl/8b5Zr

feasibility

yesnoattention question84.2%15.8%understanding100%missing dataaccepted due duration91%9%

*

*validation MturkHangoutTTOSliderTTOSliderWill to live22.67021.81818.62521.333

these values i got from analyzing the data in excel ... On Jun 15, 2012 11:19 PM, "Ricardo Pietrobon" < reply@reply.github.com> wrote:

Reply to this email directly or view it on GitHub: https://github.com/joaovissoci/Preference-Mturk/issues/5

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7109517

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114035

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114104

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7117405

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119175

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119215

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7182205

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183045

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183349

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7187558

rpietro commented 12 years ago

I am not entirely sure i understood the question, but:

you compared willtolive by mturk (yes/no) - this is part of the anwer
if you want to compare willtolive by TTO and Slider then you have to create a dichotomous variable that is either TTO or slider. I don't know which variables currently represent TTO and slider and so i can't really give you the code corresponding to that.

On Mon, Jul 30, 2012 at 2:34 PM, tkooyen < reply@reply.github.com

wrote:

ok,

problem : I want to compare the variable WillToLive between the TTO and Slider, and between the MTurk and Hangout

code I'm using :

new.Mturk <-car::recode(Mturk, "1 = 'yes'; NA = 'no'") t.test(WillToLive~new.Mturk)
    Welch Two Sample t-test
data: WillToLive by new.Mturk t = -1.0914, df = 20.827, p-value = 0.2876 alternative hypothesis: true difference in means is not equal to 0 95 percent confidence interval: -5.829770 1.818218 sample estimates: mean in group no mean in group yes 20.05882 22.06460

there is no error, but this data gives me the mean of all the Turkers that answered the questionnaire regardless if they've answered the TTO or the Slider

On Mon, Jul 23, 2012 at 3:59 PM, Ricardo Pietrobon reply@reply.github.com wrote:

so, i would suggest that you post your questions using a three-point structure, just to make sure i can answer exactly what you need:

what the problem is

which code you used

what error message you got

would also be good for us to meet, can you shoot me an invitation?

On Mon, Jul 23, 2012 at 12:40 PM, tkooyen < reply@reply.github.com

wrote:

ok, i'll upload the census data into the data set the way you've described

i've already worked around the commands that you've sent below but just to make sure i understand, the t test and the chisquare test are ways to filter and statistically analyze the dataset, right ? i'll use one or the other dependig if my variable is continuous or categorical ...

i'm getting the hang of it, just give me some time ... the problem is I have to stop to think that I'm in excel ! if i don't get it right by the end of today, i'll send you an invite

On Mon, Jul 23, 2012 at 1:28 PM, Ricardo Pietrobon reply@reply.github.com wrote:

ok, so just to make sure i understand the question. how do you include census info: would put the average values for each the variables that are in your data set and census. also, check how buhrmester and others have done this

Do you need the commands to get summary statistics for the variables below?

if so, did you check the section below -- it's in your script. Talitha, we probably need to set a time to go over this. please shoot me an invitation

###########################################################################################

TABLE 1: DEMOGRAPHICS

###########################################################################################

describes your entire dataset

describe(templateData)

summary(variable) qplot(variable)

t.test, where outcome is a continuous variable and predictor is a

dichotomous variable t.test(outcome~predictor)

chi square test where both outcome and predictor are categorical

variables CrossTable(outcome, predictor, chisq=TRUE, missing.include=TRUE, format="SAS", prop.r=FALSE)

On Mon, Jul 23, 2012 at 11:59 AM, tkooyen < reply@reply.github.com

wrote:

so, I've already erased all my tests what I want to analyze in my first table is the demographic aspect of the Mturk population compared to the US census the only problem is the database for demographic variables like age, gender, etc ... includes not only the Mturk population but the hangout population as well, so before doing the analysis i've "filtered" the Mturk population the variables for this first table are age gender race education income marital status comorbidities the US census data is not in the database, if needed tell me how to include this data into the database

On Thu, Jul 19, 2012 at 9:20 PM, Ricardo Pietrobon reply@reply.github.com wrote:

perfect, in that way i can get into your code and add examples, and then create some other instructional material for you to interpret the results that will look like http://goo.gl/jKPjy

On Thu, Jul 19, 2012 at 8:18 PM, tkooyen < reply@reply.github.com

wrote:

the code is really confused because i was doing a bunch of tests to understand how the program worked i'm going to take down the codes that i did that have nothing to do with what i'm trying to analyze than i'll send you a list of the variable names

On Thu, Jul 19, 2012 at 7:35 PM, Ricardo Pietrobon reply@reply.github.com wrote:

Talitha, could you just list side by side the variable names of what you are trying to compare? i tried to go through the code but i am somewhat confused

On Thu, Jul 19, 2012 at 4:20 PM, tkooyen < reply@reply.github.com

wrote:

what are the statistical tests I have to do for these comparisons ?

On Thu, Jul 19, 2012 at 5:17 PM, Ricardo Pietrobon reply@reply.github.com wrote:

got, but what was the specific question for me?

On Thu, Jul 19, 2012 at 3:15 PM, tkooyen < reply@reply.github.com

wrote:

Hi Ricardo, I've built these tables based on the analysis we've discussed previously.

**

external validity in relation to 2010 US census tract (Table 1)

MturkUS censusageunder 35 = 53%under 18 = 24% 18 to 44 = 36.5% 45 to 64 = 26.4% 65 and over = 13% http://goo.gl/BqL7ogendermale 46 %male 49.2% http://goo.gl/slCCLracewhite = 81.8%white 72.4% african american 12.6% native american 0.9% asian 4.8% pacific islander 0.2% some other race 6.2% two or more races 2.9% http://goo.gl/nbP4sethnicityeducation2 year college or more 56.6%population 25 years and older high school graduate or more 83.4% some college or more 50.1% bachelor degree or more 25.2% http://goo.gl/ajPAqincome over 20,000 70.4% http://goo.gl/U7Moe http://goo.gl/ZYnDMmarital status non single 50.6%married 54.4% widowed, divorced or separated 18.5% never married 27.1% http://goo.gl/F6Ukpcomorbiditiesno 63.3%http://goo.gl/8b5Zr

feasibility

yesnoattention question84.2%15.8%understanding100%missing dataaccepted due duration91%9%

*

*validation MturkHangoutTTOSliderTTOSliderWill to live22.67021.81818.62521.333

these values i got from analyzing the data in excel ... On Jun 15, 2012 11:19 PM, "Ricardo Pietrobon" < reply@reply.github.com> wrote:

Reply to this email directly or view it on GitHub: https://github.com/joaovissoci/Preference-Mturk/issues/5

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7109517

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114035

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114104

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7117405

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119175

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119215

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7182205

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183045

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183349

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7187558

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7377314

rpietro commented 12 years ago

sorry, the color doesn't come across github, you will have to tell me what specifically would like to interpret

On Mon, Jul 30, 2012 at 3:01 PM, tkooyen < reply@reply.github.com

wrote:

other doubt

problem : need help to interpret the values on the chi-square test

for example the chi-square test for age (I want to know what are the values in red)

CrossTable(Age, Mturk, chisq=TRUE, missing.include=TRUE, format="SAS", prop.r=FALSE)

Cell Contents ------------------------- N Chi-square contribution N / Col Total N / Table Total

Total Observations in Table: 404
         | Mturk
     Age |         1 |        NA | Row Total |
------------- ----------- ----------- ----------- 2 90 1 91 0.092 2.090 0.233 0.059 0.223 0.002

3 116 8 124

0.065 1.483

0.300 0.471

0.287 0.020

------------- ----------- ----------- -----------

4 73 7 80

0.172 3.922

0.189 0.412

0.181 0.017

------------- ----------- ----------- -----------

5 65 0 65

0.120 2.735

0.168 0.000

0.161 0.000

------------- ----------- ----------- -----------

6 36 1 37

0.009 0.199

0.093 0.059

0.089 0.002

------------- ----------- ----------- -----------

7 4 0 4

0.007 0.168

0.010 0.000

0.010 0.000

------------- ----------- ----------- -----------

8 1 0 1

0.002 0.042

0.003 0.000

0.002 0.000

------------- ----------- ----------- -----------

NA 2 0 2

0.004 0.084

0.005 0.000

0.005 0.000

------------- ----------- ----------- -----------

Column Total 387 17 404

0.958 0.042

------------- ----------- ----------- -----------

Statistics for All Table Factors

Pearson's Chi-squared test

Chi^2 = 11.1961 d.f. = 7 p = 0.130291

On Mon, Jul 30, 2012 at 3:34 PM, Talitha Yen tkooyen@gmail.com wrote:
ok,

problem : I want to compare the variable WillToLive between the TTO and Slider, and between the MTurk and Hangout

code I'm using :

new.Mturk <-car::recode(Mturk, "1 = 'yes'; NA = 'no'") t.test(WillToLive~new.Mturk)
    Welch Two Sample t-test
data: WillToLive by new.Mturk t = -1.0914, df = 20.827, p-value = 0.2876 alternative hypothesis: true difference in means is not equal to 0 95 percent confidence interval: -5.829770 1.818218 sample estimates: mean in group no mean in group yes 20.05882 22.06460

there is no error, but this data gives me the mean of all the Turkers that answered the questionnaire regardless if they've answered the TTO or the Slider

On Mon, Jul 23, 2012 at 3:59 PM, Ricardo Pietrobon < reply@reply.github.com

wrote:

so, i would suggest that you post your questions using a three-point structure, just to make sure i can answer exactly what you need:

what the problem is

which code you used

what error message you got

would also be good for us to meet, can you shoot me an invitation?

On Mon, Jul 23, 2012 at 12:40 PM, tkooyen < reply@reply.github.com

wrote:

ok, i'll upload the census data into the data set the way you've described

i've already worked around the commands that you've sent below but just to make sure i understand, the t test and the chisquare test are ways to filter and statistically analyze the dataset, right ? i'll use one or the other dependig if my variable is continuous or categorical ...

i'm getting the hang of it, just give me some time ... the problem is I have to stop to think that I'm in excel ! if i don't get it right by the end of today, i'll send you an invite

On Mon, Jul 23, 2012 at 1:28 PM, Ricardo Pietrobon reply@reply.github.com wrote:

ok, so just to make sure i understand the question. how do you include census info: would put the average values for each the variables that are in your data set and census. also, check how buhrmester and others have done this

Do you need the commands to get summary statistics for the variables below?

if so, did you check the section below -- it's in your script. Talitha, we probably need to set a time to go over this. please shoot me an invitation
###########################################################################################

TABLE 1: DEMOGRAPHICS

###########################################################################################

describes your entire dataset

describe(templateData)

summary(variable) qplot(variable)

t.test, where outcome is a continuous variable and predictor is a

dichotomous variable t.test(outcome~predictor)

chi square test where both outcome and predictor are categorical

variables CrossTable(outcome, predictor, chisq=TRUE, missing.include=TRUE, format="SAS", prop.r=FALSE)

On Mon, Jul 23, 2012 at 11:59 AM, tkooyen < reply@reply.github.com

wrote:

so, I've already erased all my tests what I want to analyze in my first table is the demographic aspect of the Mturk population compared to the US census the only problem is the database for demographic variables like age, gender, etc ... includes not only the Mturk population but the hangout population as well, so before doing the analysis i've "filtered" the Mturk population the variables for this first table are age gender race education income marital status comorbidities the US census data is not in the database, if needed tell me how to include this data into the database

On Thu, Jul 19, 2012 at 9:20 PM, Ricardo Pietrobon reply@reply.github.com wrote:

perfect, in that way i can get into your code and add examples, and then create some other instructional material for you to interpret the results that will look like http://goo.gl/jKPjy

On Thu, Jul 19, 2012 at 8:18 PM, tkooyen < reply@reply.github.com

wrote:

the code is really confused because i was doing a bunch of tests to understand how the program worked i'm going to take down the codes that i did that have nothing to do with what i'm trying to analyze than i'll send you a list of the variable names

On Thu, Jul 19, 2012 at 7:35 PM, Ricardo Pietrobon reply@reply.github.com wrote:

Talitha, could you just list side by side the variable names of what you are trying to compare? i tried to go through the code but i am somewhat confused

On Thu, Jul 19, 2012 at 4:20 PM, tkooyen < reply@reply.github.com

wrote:

what are the statistical tests I have to do for these comparisons ?

On Thu, Jul 19, 2012 at 5:17 PM, Ricardo Pietrobon reply@reply.github.com wrote:

got, but what was the specific question for me?

On Thu, Jul 19, 2012 at 3:15 PM, tkooyen < reply@reply.github.com

wrote:

Hi Ricardo, I've built these tables based on the analysis we've discussed previously.

**

external validity in relation to 2010 US census tract (Table 1)

MturkUS censusageunder 35 = 53%under 18 = 24% 18 to 44 = 36.5% 45 to 64 = 26.4% 65 and over = 13% http://goo.gl/BqL7ogendermale 46 %male 49.2% http://goo.gl/slCCLracewhite = 81.8%white 72.4% african american 12.6% native american 0.9% asian 4.8% pacific islander 0.2% some other race 6.2% two or more races 2.9% http://goo.gl/nbP4sethnicityeducation2 year college or more 56.6%population 25 years and older high school graduate or more 83.4% some college or more 50.1% bachelor degree or more 25.2% http://goo.gl/ajPAqincome over 20,000 70.4% http://goo.gl/U7Moe http://goo.gl/ZYnDMmarital status non single 50.6%married 54.4% widowed, divorced or separated 18.5% never married 27.1% http://goo.gl/F6Ukpcomorbiditiesno 63.3% http://goo.gl/8b5Zr

feasibility

yesnoattention question84.2%15.8%understanding100%missing dataaccepted due duration91%9%

*

*validation MturkHangoutTTOSliderTTOSliderWill to live22.67021.81818.62521.333

these values i got from analyzing the data in excel ... On Jun 15, 2012 11:19 PM, "Ricardo Pietrobon" < reply@reply.github.com> wrote:

Reply to this email directly or view it on GitHub: https://github.com/joaovissoci/Preference-Mturk/issues/5

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7109517

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114035

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114104

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7117405

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119175

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119215

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7182205

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183045

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183349

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7187558

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7378131

-------------	-----------	-----------	-----------	2	90	1	91		0.092	2.090			0.233	0.059			0.223	0.002
3	116	8	124
0.065	1.483
0.300	0.471
0.287	0.020
-------------	-----------	-----------	-----------
4	73	7	80
0.172	3.922
0.189	0.412
0.181	0.017
-------------	-----------	-----------	-----------
5	65	0	65
0.120	2.735
0.168	0.000
0.161	0.000
-------------	-----------	-----------	-----------
6	36	1	37
0.009	0.199
0.093	0.059
0.089	0.002
-------------	-----------	-----------	-----------
7	4	0	4
0.007	0.168
0.010	0.000
0.010	0.000
-------------	-----------	-----------	-----------
8	1	0	1
0.002	0.042
0.003	0.000
0.002	0.000
-------------	-----------	-----------	-----------
NA	2	0	2
0.004	0.084
0.005	0.000
0.005	0.000
-------------	-----------	-----------	-----------
Column Total	387	17	404
0.958	0.042
-------------	-----------	-----------	-----------

rpietro commented 12 years ago

also, please check http://goo.gl/iVIqw

On Mon, Jul 30, 2012 at 9:48 PM, Ricardo Pietrobon pietr007@gmail.comwrote:

sorry, the color doesn't come across github, you will have to tell me what specifically would like to interpret

On Mon, Jul 30, 2012 at 3:01 PM, tkooyen < reply@reply.github.com
wrote:

other doubt

problem : need help to interpret the values on the chi-square test

for example the chi-square test for age (I want to know what are the values in red)

CrossTable(Age, Mturk, chisq=TRUE, missing.include=TRUE, format="SAS", prop.r=FALSE)

Cell Contents ------------------------- N Chi-square contribution N / Col Total N / Table Total

Total Observations in Table: 404
         | Mturk
     Age |         1 |        NA | Row Total |
------------- ----------- ----------- ----------- 2 90 1 91 0.092 2.090 0.233 0.059 0.223 0.002

3 116 8 124

0.065 1.483

0.300 0.471

0.287 0.020

------------- ----------- ----------- -----------

4 73 7 80

0.172 3.922

0.189 0.412

0.181 0.017

------------- ----------- ----------- -----------

5 65 0 65

0.120 2.735

0.168 0.000

0.161 0.000

------------- ----------- ----------- -----------

6 36 1 37

0.009 0.199

0.093 0.059

0.089 0.002

------------- ----------- ----------- -----------

7 4 0 4

0.007 0.168

0.010 0.000

0.010 0.000

------------- ----------- ----------- -----------

8 1 0 1

0.002 0.042

0.003 0.000

0.002 0.000

------------- ----------- ----------- -----------

NA 2 0 2

0.004 0.084

0.005 0.000

0.005 0.000

------------- ----------- ----------- -----------

Column Total 387 17 404

0.958 0.042

------------- ----------- ----------- -----------

Statistics for All Table Factors

Pearson's Chi-squared test

Chi^2 = 11.1961 d.f. = 7 p = 0.130291

On Mon, Jul 30, 2012 at 3:34 PM, Talitha Yen tkooyen@gmail.com wrote:
ok,

problem : I want to compare the variable WillToLive between the TTO and Slider, and between the MTurk and Hangout

code I'm using :

new.Mturk <-car::recode(Mturk, "1 = 'yes'; NA = 'no'") t.test(WillToLive~new.Mturk)
    Welch Two Sample t-test
data: WillToLive by new.Mturk t = -1.0914, df = 20.827, p-value = 0.2876 alternative hypothesis: true difference in means is not equal to 0 95 percent confidence interval: -5.829770 1.818218 sample estimates: mean in group no mean in group yes 20.05882 22.06460

there is no error, but this data gives me the mean of all the Turkers that answered the questionnaire regardless if they've answered the TTO or the Slider

On Mon, Jul 23, 2012 at 3:59 PM, Ricardo Pietrobon < reply@reply.github.com

wrote:

so, i would suggest that you post your questions using a three-point structure, just to make sure i can answer exactly what you need:

what the problem is

which code you used

what error message you got

would also be good for us to meet, can you shoot me an invitation?

On Mon, Jul 23, 2012 at 12:40 PM, tkooyen < reply@reply.github.com

wrote:

ok, i'll upload the census data into the data set the way you've described

i've already worked around the commands that you've sent below but just to make sure i understand, the t test and the chisquare test are ways to filter and statistically analyze the dataset, right ? i'll use one or the other dependig if my variable is continuous or categorical ...

i'm getting the hang of it, just give me some time ... the problem is I have to stop to think that I'm in excel ! if i don't get it right by the end of today, i'll send you an invite

On Mon, Jul 23, 2012 at 1:28 PM, Ricardo Pietrobon reply@reply.github.com wrote:

ok, so just to make sure i understand the question. how do you include census info: would put the average values for each the variables that are in your data set and census. also, check how buhrmester and others have done this

Do you need the commands to get summary statistics for the variables below?

if so, did you check the section below -- it's in your script. Talitha, we probably need to set a time to go over this. please shoot me an invitation
###########################################################################################

TABLE 1: DEMOGRAPHICS

###########################################################################################

describes your entire dataset

describe(templateData)

summary(variable) qplot(variable)

t.test, where outcome is a continuous variable and predictor is a

dichotomous variable t.test(outcome~predictor)

chi square test where both outcome and predictor are categorical

variables CrossTable(outcome, predictor, chisq=TRUE, missing.include=TRUE, format="SAS", prop.r=FALSE)

On Mon, Jul 23, 2012 at 11:59 AM, tkooyen < reply@reply.github.com

wrote:

so, I've already erased all my tests what I want to analyze in my first table is the demographic aspect of the Mturk population compared to the US census the only problem is the database for demographic variables like age, gender, etc ... includes not only the Mturk population but the hangout population as well, so before doing the analysis i've "filtered" the Mturk population the variables for this first table are age gender race education income marital status comorbidities the US census data is not in the database, if needed tell me how to include this data into the database

On Thu, Jul 19, 2012 at 9:20 PM, Ricardo Pietrobon reply@reply.github.com wrote:

perfect, in that way i can get into your code and add examples, and then create some other instructional material for you to interpret the results that will look like http://goo.gl/jKPjy

On Thu, Jul 19, 2012 at 8:18 PM, tkooyen < reply@reply.github.com

wrote:

the code is really confused because i was doing a bunch of tests to understand how the program worked i'm going to take down the codes that i did that have nothing to do with what i'm trying to analyze than i'll send you a list of the variable names

On Thu, Jul 19, 2012 at 7:35 PM, Ricardo Pietrobon reply@reply.github.com wrote:

Talitha, could you just list side by side the variable names of what you are trying to compare? i tried to go through the code but i am somewhat confused

On Thu, Jul 19, 2012 at 4:20 PM, tkooyen < reply@reply.github.com

wrote:

what are the statistical tests I have to do for these comparisons ?

On Thu, Jul 19, 2012 at 5:17 PM, Ricardo Pietrobon reply@reply.github.com wrote:

got, but what was the specific question for me?

On Thu, Jul 19, 2012 at 3:15 PM, tkooyen < reply@reply.github.com

wrote:

Hi Ricardo, I've built these tables based on the analysis we've discussed previously.

**

external validity in relation to 2010 US census tract (Table 1)

MturkUS censusageunder 35 = 53%under 18 = 24% 18 to 44 = 36.5% 45 to 64 = 26.4% 65 and over = 13% http://goo.gl/BqL7ogendermale 46 %male 49.2% http://goo.gl/slCCLracewhite = 81.8%white 72.4% african american 12.6% native american 0.9% asian 4.8% pacific islander 0.2% some other race 6.2% two or more races 2.9% http://goo.gl/nbP4sethnicityeducation2 year college or more 56.6%population 25 years and older high school graduate or more 83.4% some college or more 50.1% bachelor degree or more 25.2% http://goo.gl/ajPAqincome over 20,000 70.4% http://goo.gl/U7Moe http://goo.gl/ZYnDMmarital status non single 50.6%married 54.4% widowed, divorced or separated 18.5% never married 27.1% http://goo.gl/F6Ukpcomorbiditiesno 63.3% http://goo.gl/8b5Zr

feasibility

yesnoattention question84.2%15.8%understanding100%missing dataaccepted due duration91%9%

*

*validation MturkHangoutTTOSliderTTOSliderWill to live22.67021.81818.62521.333

these values i got from analyzing the data in excel ... On Jun 15, 2012 11:19 PM, "Ricardo Pietrobon" < reply@reply.github.com> wrote:

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7109517

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114035

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114104

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7117405

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119175

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119215

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7182205

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183045

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183349

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7187558

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7378131

-------------	-----------	-----------	-----------	2	90	1	91		0.092	2.090			0.233	0.059			0.223	0.002
3	116	8	124
0.065	1.483
0.300	0.471
0.287	0.020
-------------	-----------	-----------	-----------
4	73	7	80
0.172	3.922
0.189	0.412
0.181	0.017
-------------	-----------	-----------	-----------
5	65	0	65
0.120	2.735
0.168	0.000
0.161	0.000
-------------	-----------	-----------	-----------
6	36	1	37
0.009	0.199
0.093	0.059
0.089	0.002
-------------	-----------	-----------	-----------
7	4	0	4
0.007	0.168
0.010	0.000
0.010	0.000
-------------	-----------	-----------	-----------
8	1	0	1
0.002	0.042
0.003	0.000
0.002	0.000
-------------	-----------	-----------	-----------
NA	2	0	2
0.004	0.084
0.005	0.000
0.005	0.000
-------------	-----------	-----------	-----------
Column Total	387	17	404
0.958	0.042
-------------	-----------	-----------	-----------

tkooyen commented 12 years ago

posting the issue as we discussed earlier

Problem: Ricardo, I don't know how to compare the mturk sample against the census population. all my comparison are for categorical variables Code and error: i don't what the test is, therefore i couldn't create the code

On Mon, Jul 30, 2012 at 10:48 PM, Ricardo Pietrobon < reply@reply.github.com

wrote:

sorry, the color doesn't come across github, you will have to tell me what specifically would like to interpret

On Mon, Jul 30, 2012 at 3:01 PM, tkooyen < reply@reply.github.com
wrote:

other doubt

problem : need help to interpret the values on the chi-square test

for example the chi-square test for age (I want to know what are the values in red)

CrossTable(Age, Mturk, chisq=TRUE, missing.include=TRUE, format="SAS", prop.r=FALSE)

Cell Contents ------------------------- N Chi-square contribution N / Col Total N / Table Total

Total Observations in Table: 404
         | Mturk
     Age |         1 |        NA | Row Total |
------------- ----------- ----------- ----------- 2 90 1 91 0.092 2.090 0.233 0.059 0.223 0.002

3 116 8 124

0.065 1.483

0.300 0.471

0.287 0.020

------------- ----------- ----------- -----------

4 73 7 80

0.172 3.922

0.189 0.412

0.181 0.017

------------- ----------- ----------- -----------

5 65 0 65

0.120 2.735

0.168 0.000

0.161 0.000

------------- ----------- ----------- -----------

6 36 1 37

0.009 0.199

0.093 0.059

0.089 0.002

------------- ----------- ----------- -----------

7 4 0 4

0.007 0.168

0.010 0.000

0.010 0.000

------------- ----------- ----------- -----------

8 1 0 1

0.002 0.042

0.003 0.000

0.002 0.000

------------- ----------- ----------- -----------

NA 2 0 2

0.004 0.084

0.005 0.000

0.005 0.000

------------- ----------- ----------- -----------

Column Total 387 17 404

0.958 0.042

------------- ----------- ----------- -----------

Statistics for All Table Factors

Pearson's Chi-squared test

Chi^2 = 11.1961 d.f. = 7 p = 0.130291

On Mon, Jul 30, 2012 at 3:34 PM, Talitha Yen tkooyen@gmail.com wrote:
ok,

problem : I want to compare the variable WillToLive between the TTO and Slider, and between the MTurk and Hangout

code I'm using :

new.Mturk <-car::recode(Mturk, "1 = 'yes'; NA = 'no'") t.test(WillToLive~new.Mturk)
    Welch Two Sample t-test
data: WillToLive by new.Mturk t = -1.0914, df = 20.827, p-value = 0.2876 alternative hypothesis: true difference in means is not equal to 0 95 percent confidence interval: -5.829770 1.818218 sample estimates: mean in group no mean in group yes 20.05882 22.06460

there is no error, but this data gives me the mean of all the Turkers that answered the questionnaire regardless if they've answered the TTO or the Slider

On Mon, Jul 23, 2012 at 3:59 PM, Ricardo Pietrobon < reply@reply.github.com

wrote:

so, i would suggest that you post your questions using a three-point structure, just to make sure i can answer exactly what you need:

what the problem is

which code you used

what error message you got

would also be good for us to meet, can you shoot me an invitation?

On Mon, Jul 23, 2012 at 12:40 PM, tkooyen < reply@reply.github.com

wrote:

ok, i'll upload the census data into the data set the way you've described

i've already worked around the commands that you've sent below but just to make sure i understand, the t test and the chisquare test are ways to filter and statistically analyze the dataset, right ? i'll use one or the other dependig if my variable is continuous or categorical ...

i'm getting the hang of it, just give me some time ... the problem is I have to stop to think that I'm in excel ! if i don't get it right by the end of today, i'll send you an invite

On Mon, Jul 23, 2012 at 1:28 PM, Ricardo Pietrobon reply@reply.github.com wrote:

ok, so just to make sure i understand the question. how do you include census info: would put the average values for each the variables that are in your data set and census. also, check how buhrmester and others have done this

Do you need the commands to get summary statistics for the variables below?

if so, did you check the section below -- it's in your script. Talitha, we probably need to set a time to go over this. please shoot me an invitation
###########################################################################################

TABLE 1: DEMOGRAPHICS

###########################################################################################

describes your entire dataset

describe(templateData)

summary(variable) qplot(variable)

t.test, where outcome is a continuous variable and predictor is a

dichotomous variable t.test(outcome~predictor)

chi square test where both outcome and predictor are categorical

variables CrossTable(outcome, predictor, chisq=TRUE, missing.include=TRUE, format="SAS", prop.r=FALSE)

On Mon, Jul 23, 2012 at 11:59 AM, tkooyen < reply@reply.github.com

wrote:

so, I've already erased all my tests what I want to analyze in my first table is the demographic aspect of the Mturk population compared to the US census the only problem is the database for demographic variables like age, gender, etc ... includes not only the Mturk population but the hangout population as well, so before doing the analysis i've "filtered" the Mturk population the variables for this first table are age gender race education income marital status comorbidities the US census data is not in the database, if needed tell me how to include this data into the database

On Thu, Jul 19, 2012 at 9:20 PM, Ricardo Pietrobon reply@reply.github.com wrote:

perfect, in that way i can get into your code and add examples, and then create some other instructional material for you to interpret the results that will look like http://goo.gl/jKPjy

On Thu, Jul 19, 2012 at 8:18 PM, tkooyen < reply@reply.github.com

wrote:

the code is really confused because i was doing a bunch of tests to understand how the program worked i'm going to take down the codes that i did that have nothing to do with what i'm trying to analyze than i'll send you a list of the variable names

On Thu, Jul 19, 2012 at 7:35 PM, Ricardo Pietrobon reply@reply.github.com wrote:

Talitha, could you just list side by side the variable names of what you are trying to compare? i tried to go through the code but i am somewhat confused

On Thu, Jul 19, 2012 at 4:20 PM, tkooyen < reply@reply.github.com

wrote:

what are the statistical tests I have to do for these comparisons ?

On Thu, Jul 19, 2012 at 5:17 PM, Ricardo Pietrobon reply@reply.github.com wrote:

got, but what was the specific question for me?

On Thu, Jul 19, 2012 at 3:15 PM, tkooyen < reply@reply.github.com

wrote:

Hi Ricardo, I've built these tables based on the analysis we've discussed previously.

**

external validity in relation to 2010 US census tract (Table 1)

MturkUS censusageunder 35 = 53%under 18 = 24% 18 to 44 = 36.5% 45 to 64 = 26.4% 65 and over = 13% http://goo.gl/BqL7ogendermale 46 %male 49.2% http://goo.gl/slCCLracewhite = 81.8%white 72.4% african american 12.6% native american 0.9% asian 4.8% pacific islander 0.2% some other race 6.2% two or more races 2.9% http://goo.gl/nbP4sethnicityeducation2 year college or more 56.6%population 25 years and older high school graduate or more 83.4% some college or more 50.1% bachelor degree or more 25.2% http://goo.gl/ajPAqincome over 20,000 70.4% http://goo.gl/U7Moe http://goo.gl/ZYnDMmarital status non single 50.6%married 54.4% widowed, divorced or separated 18.5% never married 27.1% http://goo.gl/F6Ukpcomorbiditiesno 63.3% http://goo.gl/8b5Zr

feasibility

yesnoattention question84.2%15.8%understanding100%missing dataaccepted due duration91%9%

*

*validation MturkHangoutTTOSliderTTOSliderWill to live22.67021.81818.62521.333

these values i got from analyzing the data in excel ... On Jun 15, 2012 11:19 PM, "Ricardo Pietrobon" < reply@reply.github.com> wrote:

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7109517

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114035

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114104

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7117405

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119175

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119215

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7182205

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183045

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183349

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7187558

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7378131
Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7387672

-------------	-----------	-----------	-----------	2	90	1	91		0.092	2.090			0.233	0.059			0.223	0.002
3	116	8	124
0.065	1.483
0.300	0.471
0.287	0.020
-------------	-----------	-----------	-----------
4	73	7	80
0.172	3.922
0.189	0.412
0.181	0.017
-------------	-----------	-----------	-----------
5	65	0	65
0.120	2.735
0.168	0.000
0.161	0.000
-------------	-----------	-----------	-----------
6	36	1	37
0.009	0.199
0.093	0.059
0.089	0.002
-------------	-----------	-----------	-----------
7	4	0	4
0.007	0.168
0.010	0.000
0.010	0.000
-------------	-----------	-----------	-----------
8	1	0	1
0.002	0.042
0.003	0.000
0.002	0.000
-------------	-----------	-----------	-----------
NA	2	0	2
0.004	0.084
0.005	0.000
0.005	0.000
-------------	-----------	-----------	-----------
Column Total	387	17	404
0.958	0.042
-------------	-----------	-----------	-----------

tkooyen commented 12 years ago

Problem : I don't know what the values of the chi-square test are Code and error : please send me the chi-square toolbox

On Sat, Aug 4, 2012 at 7:47 PM, Talitha Yen tkooyen@gmail.com wrote:

posting the issue as we discussed earlier

Problem: Ricardo, I don't know how to compare the mturk sample against the census population. all my comparison are for categorical variables Code and error: i don't what the test is, therefore i couldn't create the code

On Mon, Jul 30, 2012 at 10:48 PM, Ricardo Pietrobon < reply@reply.github.com
wrote:

sorry, the color doesn't come across github, you will have to tell me what specifically would like to interpret

On Mon, Jul 30, 2012 at 3:01 PM, tkooyen < reply@reply.github.com
wrote:

other doubt

problem : need help to interpret the values on the chi-square test

for example the chi-square test for age (I want to know what are the values in red)

CrossTable(Age, Mturk, chisq=TRUE, missing.include=TRUE, format="SAS", prop.r=FALSE)

Cell Contents ------------------------- N Chi-square contribution N / Col Total N / Table Total

Total Observations in Table: 404
         | Mturk
     Age |         1 |        NA | Row Total |
------------- ----------- ----------- ----------- 2 90 1 91 0.092 2.090 0.233 0.059 0.223 0.002

3 116 8 124

0.065 1.483

0.300 0.471

0.287 0.020

------------- ----------- ----------- -----------

4 73 7 80

0.172 3.922

0.189 0.412

0.181 0.017

------------- ----------- ----------- -----------

5 65 0 65

0.120 2.735

0.168 0.000

0.161 0.000

------------- ----------- ----------- -----------

6 36 1 37

0.009 0.199

0.093 0.059

0.089 0.002

------------- ----------- ----------- -----------

7 4 0 4

0.007 0.168

0.010 0.000

0.010 0.000

------------- ----------- ----------- -----------

8 1 0 1

0.002 0.042

0.003 0.000

0.002 0.000

------------- ----------- ----------- -----------

NA 2 0 2

0.004 0.084

0.005 0.000

0.005 0.000

------------- ----------- ----------- -----------

Column Total 387 17 404

0.958 0.042

------------- ----------- ----------- -----------

Statistics for All Table Factors

Pearson's Chi-squared test

Chi^2 = 11.1961 d.f. = 7 p = 0.130291

On Mon, Jul 30, 2012 at 3:34 PM, Talitha Yen tkooyen@gmail.com wrote:
ok,

problem : I want to compare the variable WillToLive between the TTO and Slider, and between the MTurk and Hangout

code I'm using :

new.Mturk <-car::recode(Mturk, "1 = 'yes'; NA = 'no'") t.test(WillToLive~new.Mturk)
    Welch Two Sample t-test
data: WillToLive by new.Mturk t = -1.0914, df = 20.827, p-value = 0.2876 alternative hypothesis: true difference in means is not equal to 0 95 percent confidence interval: -5.829770 1.818218 sample estimates: mean in group no mean in group yes 20.05882 22.06460

there is no error, but this data gives me the mean of all the Turkers that answered the questionnaire regardless if they've answered the TTO or the Slider

On Mon, Jul 23, 2012 at 3:59 PM, Ricardo Pietrobon < reply@reply.github.com

wrote:

so, i would suggest that you post your questions using a three-point structure, just to make sure i can answer exactly what you need:

what the problem is

which code you used

what error message you got

would also be good for us to meet, can you shoot me an invitation?

On Mon, Jul 23, 2012 at 12:40 PM, tkooyen < reply@reply.github.com

wrote:

ok, i'll upload the census data into the data set the way you've described

i've already worked around the commands that you've sent below but just to make sure i understand, the t test and the chisquare test are ways to filter and statistically analyze the dataset, right ? i'll use one or the other dependig if my variable is continuous or categorical ...

i'm getting the hang of it, just give me some time ... the problem is I have to stop to think that I'm in excel ! if i don't get it right by the end of today, i'll send you an invite

On Mon, Jul 23, 2012 at 1:28 PM, Ricardo Pietrobon reply@reply.github.com wrote:

ok, so just to make sure i understand the question. how do you include census info: would put the average values for each the variables that are in your data set and census. also, check how buhrmester and others have done this

Do you need the commands to get summary statistics for the variables below?

if so, did you check the section below -- it's in your script. Talitha, we probably need to set a time to go over this. please shoot me an invitation
###########################################################################################

TABLE 1: DEMOGRAPHICS

###########################################################################################

describes your entire dataset

describe(templateData)

summary(variable) qplot(variable)

t.test, where outcome is a continuous variable and predictor is a

dichotomous variable t.test(outcome~predictor)

chi square test where both outcome and predictor are categorical

variables CrossTable(outcome, predictor, chisq=TRUE, missing.include=TRUE, format="SAS", prop.r=FALSE)

On Mon, Jul 23, 2012 at 11:59 AM, tkooyen < reply@reply.github.com

wrote:

so, I've already erased all my tests what I want to analyze in my first table is the demographic aspect of the Mturk population compared to the US census the only problem is the database for demographic variables like age, gender, etc ... includes not only the Mturk population but the hangout population as well, so before doing the analysis i've "filtered" the Mturk population the variables for this first table are age gender race education income marital status comorbidities the US census data is not in the database, if needed tell me how to include this data into the database

On Thu, Jul 19, 2012 at 9:20 PM, Ricardo Pietrobon reply@reply.github.com wrote:

perfect, in that way i can get into your code and add examples, and then create some other instructional material for you to interpret the results that will look like http://goo.gl/jKPjy

On Thu, Jul 19, 2012 at 8:18 PM, tkooyen < reply@reply.github.com

wrote:

the code is really confused because i was doing a bunch of tests to understand how the program worked i'm going to take down the codes that i did that have nothing to do with what i'm trying to analyze than i'll send you a list of the variable names

On Thu, Jul 19, 2012 at 7:35 PM, Ricardo Pietrobon reply@reply.github.com wrote:

Talitha, could you just list side by side the variable names of what you are trying to compare? i tried to go through the code but i am somewhat confused

On Thu, Jul 19, 2012 at 4:20 PM, tkooyen < reply@reply.github.com

wrote:

what are the statistical tests I have to do for these comparisons ?

On Thu, Jul 19, 2012 at 5:17 PM, Ricardo Pietrobon reply@reply.github.com wrote:

got, but what was the specific question for me?

On Thu, Jul 19, 2012 at 3:15 PM, tkooyen < reply@reply.github.com

wrote:

Hi Ricardo, I've built these tables based on the analysis we've discussed previously.

**

external validity in relation to 2010 US census tract (Table 1)

MturkUS censusageunder 35 = 53%under 18 = 24% 18 to 44 = 36.5% 45 to 64 = 26.4% 65 and over = 13% http://goo.gl/BqL7ogendermale 46 %male 49.2% http://goo.gl/slCCLracewhite = 81.8%white 72.4% african american 12.6% native american 0.9% asian 4.8% pacific islander 0.2% some other race 6.2% two or more races 2.9% http://goo.gl/nbP4sethnicityeducation2 year college or more 56.6%population 25 years and older high school graduate or more 83.4% some college or more 50.1% bachelor degree or more 25.2% http://goo.gl/ajPAqincome over 20,000 70.4% http://goo.gl/U7Moe http://goo.gl/ZYnDMmarital status non single 50.6%married 54.4% widowed, divorced or separated 18.5% never married 27.1% http://goo.gl/F6Ukpcomorbiditiesno 63.3% http://goo.gl/8b5Zr

feasibility

yesnoattention question84.2%15.8%understanding100%missing dataaccepted due duration91%9%

*

*validation MturkHangoutTTOSliderTTOSliderWill to live22.67021.81818.62521.333

these values i got from analyzing the data in excel ... On Jun 15, 2012 11:19 PM, "Ricardo Pietrobon" < reply@reply.github.com> wrote:

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7109517

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114035

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114104

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7117405

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119175

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119215

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7182205

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183045

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183349

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7187558

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7378131
Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7387672

-------------	-----------	-----------	-----------	2	90	1	91		0.092	2.090			0.233	0.059			0.223	0.002
3	116	8	124
0.065	1.483
0.300	0.471
0.287	0.020
-------------	-----------	-----------	-----------
4	73	7	80
0.172	3.922
0.189	0.412
0.181	0.017
-------------	-----------	-----------	-----------
5	65	0	65
0.120	2.735
0.168	0.000
0.161	0.000
-------------	-----------	-----------	-----------
6	36	1	37
0.009	0.199
0.093	0.059
0.089	0.002
-------------	-----------	-----------	-----------
7	4	0	4
0.007	0.168
0.010	0.000
0.010	0.000
-------------	-----------	-----------	-----------
8	1	0	1
0.002	0.042
0.003	0.000
0.002	0.000
-------------	-----------	-----------	-----------
NA	2	0	2
0.004	0.084
0.005	0.000
0.005	0.000
-------------	-----------	-----------	-----------
Column Total	387	17	404
0.958	0.042
-------------	-----------	-----------	-----------

tkooyen commented 12 years ago

Problem : I don't know how to do a test that compares one categorical variable with other two variables (categorical) at the same time (without having to create a new variable that combines the previous two or three variables) Code and error : i don't have this code

On Sat, Aug 4, 2012 at 8:17 PM, Talitha Yen tkooyen@gmail.com wrote:

Problem : I don't know what the values of the chi-square test are Code and error : please send me the chi-square toolbox

On Sat, Aug 4, 2012 at 7:47 PM, Talitha Yen tkooyen@gmail.com wrote:
posting the issue as we discussed earlier

Problem: Ricardo, I don't know how to compare the mturk sample against the census population. all my comparison are for categorical variables Code and error: i don't what the test is, therefore i couldn't create the code

On Mon, Jul 30, 2012 at 10:48 PM, Ricardo Pietrobon < reply@reply.github.com
wrote:

sorry, the color doesn't come across github, you will have to tell me what specifically would like to interpret

On Mon, Jul 30, 2012 at 3:01 PM, tkooyen < reply@reply.github.com
wrote:

other doubt

problem : need help to interpret the values on the chi-square test

for example the chi-square test for age (I want to know what are the values in red)

CrossTable(Age, Mturk, chisq=TRUE, missing.include=TRUE, format="SAS", prop.r=FALSE)

Cell Contents ------------------------- N Chi-square contribution N / Col Total N / Table Total

Total Observations in Table: 404
         | Mturk
     Age |         1 |        NA | Row Total |
------------- ----------- ----------- ----------- 2 90 1 91 0.092 2.090 0.233 0.059 0.223 0.002

3 116 8 124

0.065 1.483

0.300 0.471

0.287 0.020

------------- ----------- ----------- -----------

4 73 7 80

0.172 3.922

0.189 0.412

0.181 0.017

------------- ----------- ----------- -----------

5 65 0 65

0.120 2.735

0.168 0.000

0.161 0.000

------------- ----------- ----------- -----------

6 36 1 37

0.009 0.199

0.093 0.059

0.089 0.002

------------- ----------- ----------- -----------

7 4 0 4

0.007 0.168

0.010 0.000

0.010 0.000

------------- ----------- ----------- -----------

8 1 0 1

0.002 0.042

0.003 0.000

0.002 0.000

------------- ----------- ----------- -----------

NA 2 0 2

0.004 0.084

0.005 0.000

0.005 0.000

------------- ----------- ----------- -----------

Column Total 387 17 404

0.958 0.042

------------- ----------- ----------- -----------

Statistics for All Table Factors

Pearson's Chi-squared test

Chi^2 = 11.1961 d.f. = 7 p = 0.130291

On Mon, Jul 30, 2012 at 3:34 PM, Talitha Yen tkooyen@gmail.com wrote:
ok,

problem : I want to compare the variable WillToLive between the TTO and Slider, and between the MTurk and Hangout

code I'm using :

new.Mturk <-car::recode(Mturk, "1 = 'yes'; NA = 'no'") t.test(WillToLive~new.Mturk)
    Welch Two Sample t-test
data: WillToLive by new.Mturk t = -1.0914, df = 20.827, p-value = 0.2876 alternative hypothesis: true difference in means is not equal to 0 95 percent confidence interval: -5.829770 1.818218 sample estimates: mean in group no mean in group yes 20.05882 22.06460

there is no error, but this data gives me the mean of all the Turkers that answered the questionnaire regardless if they've answered the TTO or the Slider

On Mon, Jul 23, 2012 at 3:59 PM, Ricardo Pietrobon < reply@reply.github.com

wrote:

so, i would suggest that you post your questions using a three-point structure, just to make sure i can answer exactly what you need:

what the problem is

which code you used

what error message you got

would also be good for us to meet, can you shoot me an invitation?

On Mon, Jul 23, 2012 at 12:40 PM, tkooyen < reply@reply.github.com

wrote:

ok, i'll upload the census data into the data set the way you've described

i've already worked around the commands that you've sent below but just to make sure i understand, the t test and the chisquare test are ways to filter and statistically analyze the dataset, right ? i'll use one or the other dependig if my variable is continuous or categorical ...

i'm getting the hang of it, just give me some time ... the problem is I have to stop to think that I'm in excel ! if i don't get it right by the end of today, i'll send you an invite

On Mon, Jul 23, 2012 at 1:28 PM, Ricardo Pietrobon reply@reply.github.com wrote:

ok, so just to make sure i understand the question. how do you include census info: would put the average values for each the variables that are in your data set and census. also, check how buhrmester and others have done this

Do you need the commands to get summary statistics for the variables below?

if so, did you check the section below -- it's in your script. Talitha, we probably need to set a time to go over this. please shoot me an invitation
###########################################################################################

TABLE 1: DEMOGRAPHICS

###########################################################################################

describes your entire dataset

describe(templateData)

summary(variable) qplot(variable)

t.test, where outcome is a continuous variable and predictor is

a dichotomous variable t.test(outcome~predictor)

chi square test where both outcome and predictor are categorical

variables CrossTable(outcome, predictor, chisq=TRUE, missing.include=TRUE, format="SAS", prop.r=FALSE)

On Mon, Jul 23, 2012 at 11:59 AM, tkooyen < reply@reply.github.com

wrote:

so, I've already erased all my tests what I want to analyze in my first table is the demographic aspect of the Mturk population compared to the US census the only problem is the database for demographic variables like age, gender, etc ... includes not only the Mturk population but the hangout population as well, so before doing the analysis i've "filtered" the Mturk population the variables for this first table are age gender race education income marital status comorbidities the US census data is not in the database, if needed tell me how to include this data into the database

On Thu, Jul 19, 2012 at 9:20 PM, Ricardo Pietrobon reply@reply.github.com wrote:

perfect, in that way i can get into your code and add examples, and then create some other instructional material for you to interpret the results that will look like http://goo.gl/jKPjy

On Thu, Jul 19, 2012 at 8:18 PM, tkooyen < reply@reply.github.com

wrote:

the code is really confused because i was doing a bunch of tests to understand how the program worked i'm going to take down the codes that i did that have nothing to do with what i'm trying to analyze than i'll send you a list of the variable names

On Thu, Jul 19, 2012 at 7:35 PM, Ricardo Pietrobon reply@reply.github.com wrote:

Talitha, could you just list side by side the variable names of what you are trying to compare? i tried to go through the code but i am somewhat confused

On Thu, Jul 19, 2012 at 4:20 PM, tkooyen < reply@reply.github.com

wrote:

what are the statistical tests I have to do for these comparisons ?

On Thu, Jul 19, 2012 at 5:17 PM, Ricardo Pietrobon reply@reply.github.com wrote:

got, but what was the specific question for me?

On Thu, Jul 19, 2012 at 3:15 PM, tkooyen < reply@reply.github.com

wrote:

Hi Ricardo, I've built these tables based on the analysis we've discussed previously.

**

external validity in relation to 2010 US census tract (Table 1)

MturkUS censusageunder 35 = 53%under 18 = 24% 18 to 44 = 36.5% 45 to 64 = 26.4% 65 and over = 13% http://goo.gl/BqL7ogendermale 46 %male 49.2% http://goo.gl/slCCLracewhite = 81.8%white 72.4% african american 12.6% native american 0.9% asian 4.8% pacific islander 0.2% some other race 6.2% two or more races 2.9% http://goo.gl/nbP4sethnicityeducation2 year college or more 56.6%population 25 years and older high school graduate or more 83.4% some college or more 50.1% bachelor degree or more 25.2% http://goo.gl/ajPAqincome over 20,000 70.4% http://goo.gl/U7Moe http://goo.gl/ZYnDMmarital status non single 50.6%married 54.4% widowed, divorced or separated 18.5% never married 27.1% http://goo.gl/F6Ukpcomorbiditiesno 63.3% http://goo.gl/8b5Zr

feasibility

yesnoattention question84.2%15.8%understanding100%missing dataaccepted due duration91%9%

*

*validation MturkHangoutTTOSliderTTOSliderWill to live22.67021.81818.62521.333

these values i got from analyzing the data in excel ... On Jun 15, 2012 11:19 PM, "Ricardo Pietrobon" < reply@reply.github.com> wrote:

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7109517

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114035

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114104

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7117405

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119175

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119215

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7182205

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183045

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183349

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7187558

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7378131
Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7387672

-------------	-----------	-----------	-----------	2	90	1	91		0.092	2.090			0.233	0.059			0.223	0.002
3	116	8	124
0.065	1.483
0.300	0.471
0.287	0.020
-------------	-----------	-----------	-----------
4	73	7	80
0.172	3.922
0.189	0.412
0.181	0.017
-------------	-----------	-----------	-----------
5	65	0	65
0.120	2.735
0.168	0.000
0.161	0.000
-------------	-----------	-----------	-----------
6	36	1	37
0.009	0.199
0.093	0.059
0.089	0.002
-------------	-----------	-----------	-----------
7	4	0	4
0.007	0.168
0.010	0.000
0.010	0.000
-------------	-----------	-----------	-----------
8	1	0	1
0.002	0.042
0.003	0.000
0.002	0.000
-------------	-----------	-----------	-----------
NA	2	0	2
0.004	0.084
0.005	0.000
0.005	0.000
-------------	-----------	-----------	-----------
Column Total	387	17	404
0.958	0.042
-------------	-----------	-----------	-----------

rpietro commented 12 years ago

for continuous variables, i would use the one-sample t-test (see http://goo.gl/vptEf , where mu is the mean value obtained from your census estimate). for categorical variables i would use a regular chi-square test, but then comparing your study sample against the numerator and denominator for the us census -- the latter should be easy to calculate based on the numbers they provide. see http://goo.gl/qkYOs . in your case:

numerators = c(A, B) # A is your sample numerator and B is you census numerator (total number of people in the US with a positive variable, like being female or of a certain age group, etc) denominators = (C, D) # C is the people in your sample who don't have the characteristic, D is the same in your US sample, i.e. total US sample minus those who are not female or minus those who do not have a certain age table1 <- rbind(numerators, denominators) chisq.test(table1)

               # or col1 = c(91,150,109)

row2 = c(150,200,155) # and col2 = c(90,200,198) row3 = c(109,198,172) # and col3 = c(51,155,172) data.table = rbind(row1,row2,row3) # and data.table = cbind(col1,col2,col3) data.table [,1] [,2] [,3] row1 91 90 51 row2 150 200 155 row3 109 198 172 chisq.test(data.table)

On Sat, Aug 4, 2012 at 6:47 PM, tkooyen < reply@reply.github.com

wrote:

posting the issue as we discussed earlier

Problem: Ricardo, I don't know how to compare the mturk sample against the census population. all my comparison are for categorical variables Code and error: i don't what the test is, therefore i couldn't create the code

On Mon, Jul 30, 2012 at 10:48 PM, Ricardo Pietrobon < reply@reply.github.com
wrote:

sorry, the color doesn't come across github, you will have to tell me what specifically would like to interpret

On Mon, Jul 30, 2012 at 3:01 PM, tkooyen < reply@reply.github.com
wrote:

other doubt

problem : need help to interpret the values on the chi-square test

for example the chi-square test for age (I want to know what are the values in red)

CrossTable(Age, Mturk, chisq=TRUE, missing.include=TRUE, format="SAS", prop.r=FALSE)

Cell Contents ------------------------- N Chi-square contribution N / Col Total N / Table Total

Total Observations in Table: 404
         | Mturk
     Age |         1 |        NA | Row Total |
------------- ----------- ----------- ----------- 2 90 1 91 0.092 2.090 0.233 0.059 0.223 0.002

3 116 8 124

0.065 1.483

0.300 0.471

0.287 0.020

------------- ----------- ----------- -----------

4 73 7 80

0.172 3.922

0.189 0.412

0.181 0.017

------------- ----------- ----------- -----------

5 65 0 65

0.120 2.735

0.168 0.000

0.161 0.000

------------- ----------- ----------- -----------

6 36 1 37

0.009 0.199

0.093 0.059

0.089 0.002

------------- ----------- ----------- -----------

7 4 0 4

0.007 0.168

0.010 0.000

0.010 0.000

------------- ----------- ----------- -----------

8 1 0 1

0.002 0.042

0.003 0.000

0.002 0.000

------------- ----------- ----------- -----------

NA 2 0 2

0.004 0.084

0.005 0.000

0.005 0.000

------------- ----------- ----------- -----------

Column Total 387 17 404

0.958 0.042

------------- ----------- ----------- -----------

Statistics for All Table Factors

Pearson's Chi-squared test

Chi^2 = 11.1961 d.f. = 7 p = 0.130291

On Mon, Jul 30, 2012 at 3:34 PM, Talitha Yen tkooyen@gmail.com wrote:
ok,

problem : I want to compare the variable WillToLive between the TTO and Slider, and between the MTurk and Hangout

code I'm using :

new.Mturk <-car::recode(Mturk, "1 = 'yes'; NA = 'no'") t.test(WillToLive~new.Mturk)
    Welch Two Sample t-test
data: WillToLive by new.Mturk t = -1.0914, df = 20.827, p-value = 0.2876 alternative hypothesis: true difference in means is not equal to 0 95 percent confidence interval: -5.829770 1.818218 sample estimates: mean in group no mean in group yes 20.05882 22.06460

there is no error, but this data gives me the mean of all the Turkers that answered the questionnaire regardless if they've answered the TTO or the Slider

On Mon, Jul 23, 2012 at 3:59 PM, Ricardo Pietrobon < reply@reply.github.com

wrote:

so, i would suggest that you post your questions using a three-point structure, just to make sure i can answer exactly what you need:

what the problem is

which code you used

what error message you got

would also be good for us to meet, can you shoot me an invitation?

On Mon, Jul 23, 2012 at 12:40 PM, tkooyen < reply@reply.github.com

wrote:

ok, i'll upload the census data into the data set the way you've described

i've already worked around the commands that you've sent below but just to make sure i understand, the t test and the chisquare test are ways to filter and statistically analyze the dataset, right ? i'll use one or the other dependig if my variable is continuous or categorical ...

i'm getting the hang of it, just give me some time ... the problem is I have to stop to think that I'm in excel ! if i don't get it right by the end of today, i'll send you an invite

On Mon, Jul 23, 2012 at 1:28 PM, Ricardo Pietrobon reply@reply.github.com wrote:

ok, so just to make sure i understand the question. how do you include census info: would put the average values for each the variables that are in your data set and census. also, check how buhrmester and others have done this

Do you need the commands to get summary statistics for the variables below?

if so, did you check the section below -- it's in your script. Talitha, we probably need to set a time to go over this. please shoot me an invitation
###########################################################################################

TABLE 1: DEMOGRAPHICS

###########################################################################################

describes your entire dataset

describe(templateData)

summary(variable) qplot(variable)

t.test, where outcome is a continuous variable and predictor is

a dichotomous variable t.test(outcome~predictor)

chi square test where both outcome and predictor are categorical

variables CrossTable(outcome, predictor, chisq=TRUE, missing.include=TRUE, format="SAS", prop.r=FALSE)

On Mon, Jul 23, 2012 at 11:59 AM, tkooyen < reply@reply.github.com

wrote:

so, I've already erased all my tests what I want to analyze in my first table is the demographic aspect of the Mturk population compared to the US census the only problem is the database for demographic variables like age, gender, etc ... includes not only the Mturk population but the hangout population as well, so before doing the analysis i've "filtered" the Mturk population the variables for this first table are age gender race education income marital status comorbidities the US census data is not in the database, if needed tell me how to include this data into the database

On Thu, Jul 19, 2012 at 9:20 PM, Ricardo Pietrobon reply@reply.github.com wrote:

perfect, in that way i can get into your code and add examples, and then create some other instructional material for you to interpret the results that will look like http://goo.gl/jKPjy

On Thu, Jul 19, 2012 at 8:18 PM, tkooyen < reply@reply.github.com

wrote:

the code is really confused because i was doing a bunch of tests to understand how the program worked i'm going to take down the codes that i did that have nothing to do with what i'm trying to analyze than i'll send you a list of the variable names

On Thu, Jul 19, 2012 at 7:35 PM, Ricardo Pietrobon reply@reply.github.com wrote:

Talitha, could you just list side by side the variable names of what you are trying to compare? i tried to go through the code but i am somewhat confused

On Thu, Jul 19, 2012 at 4:20 PM, tkooyen < reply@reply.github.com

wrote:

what are the statistical tests I have to do for these comparisons ?

On Thu, Jul 19, 2012 at 5:17 PM, Ricardo Pietrobon reply@reply.github.com wrote:

got, but what was the specific question for me?

On Thu, Jul 19, 2012 at 3:15 PM, tkooyen < reply@reply.github.com

wrote:

Hi Ricardo, I've built these tables based on the analysis we've discussed previously.

**

external validity in relation to 2010 US census tract (Table 1)

MturkUS censusageunder 35 = 53%under 18 = 24% 18 to 44 = 36.5% 45 to 64 = 26.4% 65 and over = 13% http://goo.gl/BqL7ogendermale 46 %male 49.2% http://goo.gl/slCCLracewhite = 81.8%white 72.4% african american 12.6% native american 0.9% asian 4.8% pacific islander 0.2% some other race 6.2% two or more races 2.9% http://goo.gl/nbP4sethnicityeducation2 year college or more 56.6%population 25 years and older high school graduate or more 83.4% some college or more 50.1% bachelor degree or more 25.2% http://goo.gl/ajPAqincome over 20,000 70.4% http://goo.gl/U7Moe http://goo.gl/ZYnDMmarital status non single 50.6%married 54.4% widowed, divorced or separated 18.5% never married 27.1% http://goo.gl/F6Ukpcomorbiditiesno 63.3% http://goo.gl/8b5Zr

feasibility

yesnoattention question84.2%15.8%understanding100%missing dataaccepted due duration91%9%

*

*validation MturkHangoutTTOSliderTTOSliderWill to live22.67021.81818.62521.333

these values i got from analyzing the data in excel ... On Jun 15, 2012 11:19 PM, "Ricardo Pietrobon" < reply@reply.github.com> wrote:

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7109517

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114035

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114104

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7117405

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119175

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119215

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7182205

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183045

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183349

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7187558

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7378131

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7387672
Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7506127

-------------	-----------	-----------	-----------	2	90	1	91		0.092	2.090			0.233	0.059			0.223	0.002
3	116	8	124
0.065	1.483
0.300	0.471
0.287	0.020
-------------	-----------	-----------	-----------
4	73	7	80
0.172	3.922
0.189	0.412
0.181	0.017
-------------	-----------	-----------	-----------
5	65	0	65
0.120	2.735
0.168	0.000
0.161	0.000
-------------	-----------	-----------	-----------
6	36	1	37
0.009	0.199
0.093	0.059
0.089	0.002
-------------	-----------	-----------	-----------
7	4	0	4
0.007	0.168
0.010	0.000
0.010	0.000
-------------	-----------	-----------	-----------
8	1	0	1
0.002	0.042
0.003	0.000
0.002	0.000
-------------	-----------	-----------	-----------
NA	2	0	2
0.004	0.084
0.005	0.000
0.005	0.000
-------------	-----------	-----------	-----------
Column Total	387	17	404
0.958	0.042
-------------	-----------	-----------	-----------

rpietro commented 12 years ago

here you go: http://goo.gl/T1mEa

On Sat, Aug 4, 2012 at 7:17 PM, tkooyen < reply@reply.github.com

wrote:

Problem : I don't know what the values of the chi-square test are Code and error : please send me the chi-square toolbox

On Sat, Aug 4, 2012 at 7:47 PM, Talitha Yen tkooyen@gmail.com wrote:
posting the issue as we discussed earlier

Problem: Ricardo, I don't know how to compare the mturk sample against the census population. all my comparison are for categorical variables Code and error: i don't what the test is, therefore i couldn't create the code

On Mon, Jul 30, 2012 at 10:48 PM, Ricardo Pietrobon < reply@reply.github.com
wrote:

sorry, the color doesn't come across github, you will have to tell me what specifically would like to interpret

On Mon, Jul 30, 2012 at 3:01 PM, tkooyen < reply@reply.github.com
wrote:

other doubt

problem : need help to interpret the values on the chi-square test

for example the chi-square test for age (I want to know what are the values in red)

CrossTable(Age, Mturk, chisq=TRUE, missing.include=TRUE, format="SAS", prop.r=FALSE)

Cell Contents ------------------------- N Chi-square contribution N / Col Total N / Table Total

Total Observations in Table: 404
         | Mturk
     Age |         1 |        NA | Row Total |
------------- ----------- ----------- ----------- 2 90 1 91 0.092 2.090 0.233 0.059 0.223 0.002

3 116 8 124

0.065 1.483

0.300 0.471

0.287 0.020

------------- ----------- ----------- -----------

4 73 7 80

0.172 3.922

0.189 0.412

0.181 0.017

------------- ----------- ----------- -----------

5 65 0 65

0.120 2.735

0.168 0.000

0.161 0.000

------------- ----------- ----------- -----------

6 36 1 37

0.009 0.199

0.093 0.059

0.089 0.002

------------- ----------- ----------- -----------

7 4 0 4

0.007 0.168

0.010 0.000

0.010 0.000

------------- ----------- ----------- -----------

8 1 0 1

0.002 0.042

0.003 0.000

0.002 0.000

------------- ----------- ----------- -----------

NA 2 0 2

0.004 0.084

0.005 0.000

0.005 0.000

------------- ----------- ----------- -----------

Column Total 387 17 404

0.958 0.042

------------- ----------- ----------- -----------

Statistics for All Table Factors

Pearson's Chi-squared test

Chi^2 = 11.1961 d.f. = 7 p = 0.130291

On Mon, Jul 30, 2012 at 3:34 PM, Talitha Yen tkooyen@gmail.com wrote:
ok,

problem : I want to compare the variable WillToLive between the TTO and Slider, and between the MTurk and Hangout

code I'm using :

new.Mturk <-car::recode(Mturk, "1 = 'yes'; NA = 'no'") t.test(WillToLive~new.Mturk)
    Welch Two Sample t-test
data: WillToLive by new.Mturk t = -1.0914, df = 20.827, p-value = 0.2876 alternative hypothesis: true difference in means is not equal to 0 95 percent confidence interval: -5.829770 1.818218 sample estimates: mean in group no mean in group yes 20.05882 22.06460

there is no error, but this data gives me the mean of all the Turkers that answered the questionnaire regardless if they've answered the TTO or the Slider

On Mon, Jul 23, 2012 at 3:59 PM, Ricardo Pietrobon < reply@reply.github.com

wrote:

so, i would suggest that you post your questions using a three-point structure, just to make sure i can answer exactly what you need:

what the problem is

which code you used

what error message you got

would also be good for us to meet, can you shoot me an invitation?

On Mon, Jul 23, 2012 at 12:40 PM, tkooyen < reply@reply.github.com

wrote:

ok, i'll upload the census data into the data set the way you've described

i've already worked around the commands that you've sent below but just to make sure i understand, the t test and the chisquare test are ways to filter and statistically analyze the dataset, right ? i'll use one or the other dependig if my variable is continuous or categorical ...

i'm getting the hang of it, just give me some time ... the problem is I have to stop to think that I'm in excel ! if i don't get it right by the end of today, i'll send you an invite

On Mon, Jul 23, 2012 at 1:28 PM, Ricardo Pietrobon reply@reply.github.com wrote:

ok, so just to make sure i understand the question. how do you include census info: would put the average values for each the variables that are in your data set and census. also, check how buhrmester and others have done this

Do you need the commands to get summary statistics for the variables below?

if so, did you check the section below -- it's in your script. Talitha, we probably need to set a time to go over this. please shoot me an invitation
###########################################################################################

TABLE 1: DEMOGRAPHICS

###########################################################################################

describes your entire dataset

describe(templateData)

summary(variable) qplot(variable)

t.test, where outcome is a continuous variable and predictor

is a dichotomous variable t.test(outcome~predictor)

chi square test where both outcome and predictor are

categorical variables CrossTable(outcome, predictor, chisq=TRUE, missing.include=TRUE, format="SAS", prop.r=FALSE)

On Mon, Jul 23, 2012 at 11:59 AM, tkooyen < reply@reply.github.com

wrote:

so, I've already erased all my tests what I want to analyze in my first table is the demographic aspect of the Mturk population compared to the US census the only problem is the database for demographic variables like age, gender, etc ... includes not only the Mturk population but the hangout population as well, so before doing the analysis i've "filtered" the Mturk population the variables for this first table are age gender race education income marital status comorbidities the US census data is not in the database, if needed tell me how to include this data into the database

On Thu, Jul 19, 2012 at 9:20 PM, Ricardo Pietrobon reply@reply.github.com wrote:

perfect, in that way i can get into your code and add examples, and then create some other instructional material for you to interpret the results that will look like http://goo.gl/jKPjy

On Thu, Jul 19, 2012 at 8:18 PM, tkooyen < reply@reply.github.com

wrote:

the code is really confused because i was doing a bunch of tests to understand how the program worked i'm going to take down the codes that i did that have nothing to do with what i'm trying to analyze than i'll send you a list of the variable names

On Thu, Jul 19, 2012 at 7:35 PM, Ricardo Pietrobon reply@reply.github.com wrote:

Talitha, could you just list side by side the variable names of what you are trying to compare? i tried to go through the code but i am somewhat confused

On Thu, Jul 19, 2012 at 4:20 PM, tkooyen < reply@reply.github.com

wrote:

what are the statistical tests I have to do for these comparisons ?

On Thu, Jul 19, 2012 at 5:17 PM, Ricardo Pietrobon reply@reply.github.com wrote:

got, but what was the specific question for me?

On Thu, Jul 19, 2012 at 3:15 PM, tkooyen < reply@reply.github.com

wrote:

Hi Ricardo, I've built these tables based on the analysis we've discussed previously.

**

external validity in relation to 2010 US census tract (Table 1)

MturkUS censusageunder 35 = 53%under 18 = 24% 18 to 44 = 36.5% 45 to 64 = 26.4% 65 and over = 13% http://goo.gl/BqL7ogendermale 46 %male 49.2% http://goo.gl/slCCLracewhite = 81.8%white 72.4% african american 12.6% native american 0.9% asian 4.8% pacific islander 0.2% some other race 6.2% two or more races 2.9% http://goo.gl/nbP4sethnicityeducation2 year college or more 56.6%population 25 years and older high school graduate or more 83.4% some college or more 50.1% bachelor degree or more 25.2% http://goo.gl/ajPAqincome over 20,000 70.4% http://goo.gl/U7Moe http://goo.gl/ZYnDMmarital status non single 50.6%married 54.4% widowed, divorced or separated 18.5% never married 27.1% http://goo.gl/F6Ukpcomorbiditiesno 63.3% http://goo.gl/8b5Zr

feasibility

yesnoattention question84.2%15.8%understanding100%missing dataaccepted due duration91%9%

*

*validation MturkHangoutTTOSliderTTOSliderWill to live22.67021.81818.62521.333

these values i got from analyzing the data in excel ... On Jun 15, 2012 11:19 PM, "Ricardo Pietrobon" < reply@reply.github.com> wrote:

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7109517

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114035

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114104

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7117405

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119175

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119215

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7182205

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183045

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183349

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7187558

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7378131

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7387672
Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7506285

-------------	-----------	-----------	-----------	2	90	1	91		0.092	2.090			0.233	0.059			0.223	0.002
3	116	8	124
0.065	1.483
0.300	0.471
0.287	0.020
-------------	-----------	-----------	-----------
4	73	7	80
0.172	3.922
0.189	0.412
0.181	0.017
-------------	-----------	-----------	-----------
5	65	0	65
0.120	2.735
0.168	0.000
0.161	0.000
-------------	-----------	-----------	-----------
6	36	1	37
0.009	0.199
0.093	0.059
0.089	0.002
-------------	-----------	-----------	-----------
7	4	0	4
0.007	0.168
0.010	0.000
0.010	0.000
-------------	-----------	-----------	-----------
8	1	0	1
0.002	0.042
0.003	0.000
0.002	0.000
-------------	-----------	-----------	-----------
NA	2	0	2
0.004	0.084
0.005	0.000
0.005	0.000
-------------	-----------	-----------	-----------
Column Total	387	17	404
0.958	0.042
-------------	-----------	-----------	-----------

rpietro commented 12 years ago

see http://goo.gl/T1mEa

the only difference is that your numerator and denominator vector will have three numbers each. alternatively, you can compare three variables using the second format:

chisquare.test (var1, var2, var3)

On Sat, Aug 4, 2012 at 7:25 PM, tkooyen < reply@reply.github.com

wrote:

Problem : I don't know how to do a test that compares one categorical variable with other two variables (categorical) at the same time (without having to create a new variable that combines the previous two or three variables) Code and error : i don't have this code

On Sat, Aug 4, 2012 at 8:17 PM, Talitha Yen tkooyen@gmail.com wrote:
Problem : I don't know what the values of the chi-square test are Code and error : please send me the chi-square toolbox

On Sat, Aug 4, 2012 at 7:47 PM, Talitha Yen tkooyen@gmail.com wrote:
posting the issue as we discussed earlier

Problem: Ricardo, I don't know how to compare the mturk sample against the census population. all my comparison are for categorical variables Code and error: i don't what the test is, therefore i couldn't create the code

On Mon, Jul 30, 2012 at 10:48 PM, Ricardo Pietrobon < reply@reply.github.com
wrote:

sorry, the color doesn't come across github, you will have to tell me what specifically would like to interpret

On Mon, Jul 30, 2012 at 3:01 PM, tkooyen < reply@reply.github.com
wrote:

other doubt

problem : need help to interpret the values on the chi-square test

for example the chi-square test for age (I want to know what are the values in red)

CrossTable(Age, Mturk, chisq=TRUE, missing.include=TRUE, format="SAS", prop.r=FALSE)

Cell Contents ------------------------- N Chi-square contribution N / Col Total N / Table Total

Total Observations in Table: 404
         | Mturk
     Age |         1 |        NA | Row Total |
------------- ----------- ----------- ----------- 2 90 1 91 0.092 2.090 0.233 0.059 0.223 0.002

3 116 8 124

0.065 1.483

0.300 0.471

0.287 0.020

------------- ----------- ----------- -----------

4 73 7 80

0.172 3.922

0.189 0.412

0.181 0.017

------------- ----------- ----------- -----------

5 65 0 65

0.120 2.735

0.168 0.000

0.161 0.000

------------- ----------- ----------- -----------

6 36 1 37

0.009 0.199

0.093 0.059

0.089 0.002

------------- ----------- ----------- -----------

7 4 0 4

0.007 0.168

0.010 0.000

0.010 0.000

------------- ----------- ----------- -----------

8 1 0 1

0.002 0.042

0.003 0.000

0.002 0.000

------------- ----------- ----------- -----------

NA 2 0 2

0.004 0.084

0.005 0.000

0.005 0.000

------------- ----------- ----------- -----------

Column Total 387 17 404

0.958 0.042

------------- ----------- ----------- -----------

Statistics for All Table Factors

Pearson's Chi-squared test

Chi^2 = 11.1961 d.f. = 7 p = 0.130291

On Mon, Jul 30, 2012 at 3:34 PM, Talitha Yen tkooyen@gmail.com wrote:
ok,

problem : I want to compare the variable WillToLive between the TTO and Slider, and between the MTurk and Hangout

code I'm using :

new.Mturk <-car::recode(Mturk, "1 = 'yes'; NA = 'no'") t.test(WillToLive~new.Mturk)
    Welch Two Sample t-test
data: WillToLive by new.Mturk t = -1.0914, df = 20.827, p-value = 0.2876 alternative hypothesis: true difference in means is not equal to 0 95 percent confidence interval: -5.829770 1.818218 sample estimates: mean in group no mean in group yes 20.05882 22.06460

there is no error, but this data gives me the mean of all the Turkers that answered the questionnaire regardless if they've answered the TTO or the Slider

On Mon, Jul 23, 2012 at 3:59 PM, Ricardo Pietrobon < reply@reply.github.com

wrote:

so, i would suggest that you post your questions using a three-point structure, just to make sure i can answer exactly what you need:

what the problem is

which code you used

what error message you got

would also be good for us to meet, can you shoot me an invitation?

On Mon, Jul 23, 2012 at 12:40 PM, tkooyen < reply@reply.github.com

wrote:

ok, i'll upload the census data into the data set the way you've described

i've already worked around the commands that you've sent below but just to make sure i understand, the t test and the chisquare test are ways to filter and statistically analyze the dataset, right ? i'll use one or the other dependig if my variable is continuous or categorical ...

i'm getting the hang of it, just give me some time ... the problem is I have to stop to think that I'm in excel ! if i don't get it right by the end of today, i'll send you an invite

On Mon, Jul 23, 2012 at 1:28 PM, Ricardo Pietrobon reply@reply.github.com wrote:

ok, so just to make sure i understand the question. how do you include census info: would put the average values for each the variables that are in your data set and census. also, check how buhrmester and others have done this

Do you need the commands to get summary statistics for the variables below?

if so, did you check the section below -- it's in your script. Talitha, we probably need to set a time to go over this. please shoot me an invitation
###########################################################################################

TABLE 1: DEMOGRAPHICS

###########################################################################################

describes your entire dataset

describe(templateData)

summary(variable) qplot(variable)

t.test, where outcome is a continuous variable and predictor

is a dichotomous variable t.test(outcome~predictor)

chi square test where both outcome and predictor are

categorical variables CrossTable(outcome, predictor, chisq=TRUE, missing.include=TRUE, format="SAS", prop.r=FALSE)

On Mon, Jul 23, 2012 at 11:59 AM, tkooyen < reply@reply.github.com

wrote:

so, I've already erased all my tests what I want to analyze in my first table is the demographic aspect of the Mturk population compared to the US census the only problem is the database for demographic variables like age, gender, etc ... includes not only the Mturk population but the hangout population as well, so before doing the analysis i've "filtered" the Mturk population the variables for this first table are age gender race education income marital status comorbidities the US census data is not in the database, if needed tell me how to include this data into the database

On Thu, Jul 19, 2012 at 9:20 PM, Ricardo Pietrobon reply@reply.github.com wrote:

perfect, in that way i can get into your code and add examples, and then create some other instructional material for you to interpret the results that will look like http://goo.gl/jKPjy

On Thu, Jul 19, 2012 at 8:18 PM, tkooyen < reply@reply.github.com

wrote:

the code is really confused because i was doing a bunch of tests to understand how the program worked i'm going to take down the codes that i did that have nothing to do with what i'm trying to analyze than i'll send you a list of the variable names

On Thu, Jul 19, 2012 at 7:35 PM, Ricardo Pietrobon reply@reply.github.com wrote:

Talitha, could you just list side by side the variable names of what you are trying to compare? i tried to go through the code but i am somewhat confused

On Thu, Jul 19, 2012 at 4:20 PM, tkooyen < reply@reply.github.com

wrote:

what are the statistical tests I have to do for these comparisons ?

On Thu, Jul 19, 2012 at 5:17 PM, Ricardo Pietrobon reply@reply.github.com wrote:

got, but what was the specific question for me?

On Thu, Jul 19, 2012 at 3:15 PM, tkooyen < reply@reply.github.com

wrote:

Hi Ricardo, I've built these tables based on the analysis we've discussed previously.

**

external validity in relation to 2010 US census tract (Table 1)

MturkUS censusageunder 35 = 53%under 18 = 24% 18 to 44 = 36.5% 45 to 64 = 26.4% 65 and over = 13% http://goo.gl/BqL7ogendermale 46 %male 49.2% http://goo.gl/slCCLracewhite = 81.8%white 72.4% african american 12.6% native american 0.9% asian 4.8% pacific islander 0.2% some other race 6.2% two or more races 2.9% http://goo.gl/nbP4sethnicityeducation2 year college or more 56.6%population 25 years and older high school graduate or more 83.4% some college or more 50.1% bachelor degree or more 25.2% http://goo.gl/ajPAqincome over 20,000 70.4% http://goo.gl/U7Moe http://goo.gl/ZYnDMmaritalstatus non single 50.6%married 54.4% widowed, divorced or separated 18.5% never married 27.1% http://goo.gl/F6Ukpcomorbiditiesno 63.3% http://goo.gl/8b5Zr

feasibility

yesnoattention question84.2%15.8%understanding100%missing dataaccepted due duration91%9%

*

*validation MturkHangoutTTOSliderTTOSliderWill to live22.67021.81818.62521.333

these values i got from analyzing the data in excel ... On Jun 15, 2012 11:19 PM, "Ricardo Pietrobon" < reply@reply.github.com> wrote:

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7109517

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114035

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7114104

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7117405

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119175

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7119215

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7182205

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183045

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7183349

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7187558

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7378131

Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7387672
Reply to this email directly or view it on GitHub:

https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-7506329

-------------	-----------	-----------	-----------	2	90	1	91		0.092	2.090			0.233	0.059			0.223	0.002
3	116	8	124
0.065	1.483
0.300	0.471
0.287	0.020
-------------	-----------	-----------	-----------
4	73	7	80
0.172	3.922
0.189	0.412
0.181	0.017
-------------	-----------	-----------	-----------
5	65	0	65
0.120	2.735
0.168	0.000
0.161	0.000
-------------	-----------	-----------	-----------
6	36	1	37
0.009	0.199
0.093	0.059
0.089	0.002
-------------	-----------	-----------	-----------
7	4	0	4
0.007	0.168
0.010	0.000
0.010	0.000
-------------	-----------	-----------	-----------
8	1	0	1
0.002	0.042
0.003	0.000
0.002	0.000
-------------	-----------	-----------	-----------
NA	2	0	2
0.004	0.084
0.005	0.000
0.005	0.000
-------------	-----------	-----------	-----------
Column Total	387	17	404
0.958	0.042
-------------	-----------	-----------	-----------

rpietro commented 12 years ago

if you are looking for a comparison across three variables, you can just use http://goo.gl/vJr2h

sorry, still didn't get my computer back, and my laptop doesn't have all my gdrive completely synchronized yet, which means i can't get into your script. let me know if that doesn't work

On Mon, Sep 3, 2012 at 10:32 AM, tkooyen notifications@github.com wrote:

Problem : trying to cross three categorical variables using the crosstable code, but is not working, specifically because it is not recognizing the recoded variable Code and error : CrossTable(Attention, Mturk*new.Age, chisq=TRUE, missing.include= TRUE, format="SAS", prop.r=FALSE)
     | Mturk * new.Age
Attention NA Row Total 0 61 61 0.151 ------------- ----------- ----------- 1 343 343 0.849 ------------- ----------- ----------- Column Total 404 404 ------------- ----------- -----------

Warning message: In Ops.factor(Mturk, new.Age) : * not meaningful for factors

OBS : this code works when used with two recoded variables, and it works with 3 not recoded variables, but doesn't work with 3 variables being 1 recoded

— Reply to this email directly or view it on GitHubhttps://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-8239937.

tkooyen commented 12 years ago

Almost done with the analysis but having problems with the t test ... problem : I need a test where the outcome is continuous variable and the predictor is a categorical variable (not a dichotomous) code and error : t.test(mt$WillToLive~mt$new.Age) Error in t.test.formula(mt$WillToLive ~ mt$new.Age) : grouping factor must have exactly 2 levels

joaovissoci commented 12 years ago

Please, try this code:

On data, put the name of the data set in the script

fit <- aov(mt$WillToLive ~ mt$new.Age, data=) summary(fit) #This will give you the main results for the ANOVA comparison TukeyHSD(fit) # Here you will fidn a comparison pair by pair

Please, send me the graphs that this comand will create, just to check the

aplicability of the Variance Analysis through ANOVA. layout(matrix(c(1,2,3,4),2,2)) plot(fit)

If you have any doubts, just ask.

2012/9/10 tkooyen notifications@github.com

Almost done with the analysis but having problems with the t test ... problem : I need a test where the outcome is continuous variable and the predictor is a categorical variable (not a dichotomous) code and error : t.test(mt$WillToLive~mt$new.Age) Error in t.test.formula(mt$WillToLive ~ mt$new.Age) : grouping factor must have exactly 2 levels

— Reply to this email directly or view it on GitHubhttps://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-8425411.

Joao Ricardo N. Vissoci Psicólogo CRP 08/12469 Prof. Ms. Faculdade Ingá Doutorando em Psicologia Social - PUCsp Grupo Pro-Esporte UEM/CNPq Núcleo de Estudos e Pesquisas sobre Identidade-Metamorfose - NEPIM/PUC/CNPq Research on Research Group - RoR - Duke University Tel. 4499298078 joaovissoci@gmail.com/jrvissoci@ig.com.br proesporteuem.blogspot.com.br researchonresearch.org

rpietro commented 12 years ago

Talitha, please post different problems as different issues, otherwise this will become really hard to find later on

On Mon, Sep 10, 2012 at 3:17 PM, Joao Ricardo N Vissoci < notifications@github.com> wrote:

Please, try this code:

On data, put the name of the data set in the script

fit <- aov(mt$WillToLive ~ mt$new.Age, data=) summary(fit) #This will give you the main results for the ANOVA comparison TukeyHSD(fit) # Here you will fidn a comparison pair by pair

Please, send me the graphs that this comand will create, just to check

the aplicability of the Variance Analysis through ANOVA. layout(matrix(c(1,2,3,4),2,2)) plot(fit)

If you have any doubts, just ask.

2012/9/10 tkooyen notifications@github.com

Almost done with the analysis but having problems with the t test ... problem : I need a test where the outcome is continuous variable and the predictor is a categorical variable (not a dichotomous) code and error : t.test(mt$WillToLive~mt$new.Age) Error in t.test.formula(mt$WillToLive ~ mt$new.Age) : grouping factor must have exactly 2 levels

— Reply to this email directly or view it on GitHub< https://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-8425411>.

Joao Ricardo N. Vissoci Psicólogo CRP 08/12469 Prof. Ms. Faculdade Ingá Doutorando em Psicologia Social - PUCsp Grupo Pro-Esporte UEM/CNPq Núcleo de Estudos e Pesquisas sobre Identidade-Metamorfose - NEPIM/PUC/CNPq Research on Research Group - RoR - Duke University Tel. 4499298078 joaovissoci@gmail.com/jrvissoci@ig.com.br proesporteuem.blogspot.com.br researchonresearch.org

— Reply to this email directly or view it on GitHubhttps://github.com/joaovissoci/Preference-Mturk/issues/5#issuecomment-8434879.