search

MIDS-at-Duke / unifying-data-science-final-project-primaries

unifying-data-science-final-project-primaries created by GitHub Classroom
0 stars 2 forks source link

choosing states (unit of analysis variables) #1

Closed katevcoulter closed 4 years ago

katevcoulter commented 4 years ago

treatment states: minnesota, idaho, colorado, maine

*** will also assign shota when he joins this

katevcoulter commented 4 years ago

added shota

aberman6 commented 4 years ago

I just did a very messy first pass at the d&d analysis. I came out with a couple of questions.

  1. Which states make sense to include as our treatment group?
  2. Are there any states we should exclude from the analysis all-together?
  3. Are we only including states that have complete data from 2008-2020? If so this is very limiting.
katevcoulter commented 4 years ago
  1. treatment = our original 4 states (Minnesota, Idaho, Maine, Colorado)
katevcoulter commented 4 years ago
  1. exclude = needs more work, but we are 100% excluding Hawaii, Alaska, California, Ohio and any other state affected by covid-19 thus having their primary moved
katevcoulter commented 4 years ago
  1. not necessarily but also depends on which vars are missing
katevcoulter commented 4 years ago

i'm going through the data now and adding a new treatment column to mark the ones we've assigned to treated, control or exclude if it's blank then we still should double check to decide if we should include or not

katevcoulter commented 4 years ago

shoot i meant wisconsin not ohio my b

aberman6 commented 4 years ago

I updated the cleaning_2.R file and the turnout_v3.csv with new exclusion criteria. I excluded:

That leaves us with 25 states (4 treatment states and 21 control states).

Control = Alabama, Arizona, Arkansas, Florida, Iowa, Massachusetts, Michigan, Mississippi, Missouri, Nevada, New Hampshire, North Carolina, North Dakota, Oklahoma, South Carolina, Tennessee, Texas, Utah, Vermont, Virginia, & Washington

Treatment = Colorado, Idaho, Maine, & Minnesota

katevcoulter commented 4 years ago

perfect!! will confirm that we don’t need to exclude any other states like Cali by tomorrow morning :) excited to chat with you and Shota at 10 am

On Apr 16, 2020, at 2:15 PM, Anna Berman notifications@github.com wrote:

I updated the cleaning_2.R file and the turnout_v3.csv with new exclusion criteria. I excluded:

States that are strange like Cali States whos primaries were effected by COVID States that didn't have Democrat turnout or VEP numbers for 2008, 2016, and 2020. That leaves us with 25 states (4 treatment states and 21 control states).

Control = Alabama, Arizona, Arkansas, Florida, Iowa, Massachusetts, Michigan, Mississippi, Missouri, Nevada, New Hampshire, North Carolina, North Dakota, Oklahoma, South Carolina, Tennessee, Texas, Utah, Vermont, Virginia, & Washington

Treatment = Colorado, Idaho, Maine, & Minnesota

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub https://github.com/MIDS-at-Duke/unifying-data-science-final-project-primaries/issues/1#issuecomment-614815082, or unsubscribe https://github.com/notifications/unsubscribe-auth/AKLBR577UKVLG2PR3VSXHR3RM5DN7ANCNFSM4LV2D6IA.

shttksm commented 4 years ago

Thank you. It's awesome. I also want to know how the number of samples in the control and treatment groups, in this case, 21 and 4, affect the result statistically. I will ask Nick.