search

stats4sd / Sampling-Decision-Assistant

GNU General Public License v3.0
0 stars 0 forks source link

Question 2.5 answered "No" #35

Closed cebarahona closed 6 years ago

cebarahona commented 6 years ago

We may need to provide some help to guest-estimate the variance of the indicator. I would like to strongly encourage a numerical answer here. If the user does not have any idea, we should suggest:

  1. find out from previous studies what is the variance of the estimate
  2. Ask someone who may know
  3. If everything fails, please take the difference between the minimum value and the maximum that came from 2.4, divided it by 5 and square the result. This will yield a very rough estimate of the variance that assumes that the distribution is roughly symmetric around the mean value.
chrismclarke commented 6 years ago

I've tried by changing 2.5 to yes/no and then having different questions render for 2.6 depending on response (either input as text or number the values if known, or input as number with calculation guidance if not known).

We could change 2.4 to be a numerical min and max input and have the app automatically calculate if wanted, however I'm not sure if this would be necessary. Let me know what you think (live from v0.2.3)

cebarahona commented 6 years ago

Let me see how it works. I don't want an automated computation because everyone will go for it. Let's t me think about it

Get Outlook for Androidhttps://aka.ms/ghei36

From: Chris Clarke notifications@github.com Sent: Thursday, December 7, 2017 7:29:50 PM To: stats4sd/Sampling-Decision-Assistant Cc: Carlos Barahona; Author Subject: Re: [stats4sd/Sampling-Decision-Assistant] Question 2.5 answered "No" (#35)

I've tried by changing 2.5 to yes/no and then having different questions render for 2.6 depending on response (either input as text or number the values if known, or input as number with calculation guidance if not known).

We could change 2.4 to be a numerical min and max input and have the app automatically calculate if wanted, however I'm not sure if this would be necessary. Let me know what you think (live from v0.2.3)

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/stats4sd/Sampling-Decision-Assistant/issues/35#issuecomment-350069924, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AgIxcswAqLrUejGndyyB3bNXjMqNZ25cks5s-DyugaJpZM4Q5YAz.

chrismclarke commented 6 years ago

based on previous discussion assume implementation is fine, closing for now