search

R4EPI / R4EPIs-website

A repository to host the landing page of the R4EPIs website
https://r4epis.netlify.app
MIT License
5 stars 6 forks source link

Additions to R4epis website and survey templates #60

Open isidro-carrionmartin opened 4 years ago

isidro-carrionmartin commented 4 years ago
  1. I would suggest that we add scripts for sample size calculation considering all the parameters that we normally use: expected prevalence, precision, and Z-cut off value (or T cut off value for cluster surveys) and non response rate. Plus (in household cluster surveys): DEFF, household size, and percentage of target population in the total population. I am attaching an excel that I recently used for a household cluster sample size calculation in case it is helpful. sample size calcs_v2.xlsx
  2. I would suggest that we add scripts for the sampling:
    • Sampling for probability proportional to size (PPS) in clusters surveys using a a list of primary sampling units (e.g. villages or neighbourhoods). Depending on how straight forward it is, we can add that? Or just add the excel file? Or both?
    • Spatial sampling (with different options): Systematic random sampling using GPS random points in one area (e.g. refugee camp). Starting point for cluster (in cluster surveys) (e.g. we have to interview three clusters in one city, we draw the three clusters for the teams). We would need to explain that this would depend on the availability of GIS data (e.g. shape files for villages)
  3. Regarding the text, I had a more in depth look and I would suggest some changes. Please, see them in the attached. R4epis comments for website_survey templates.docx --> happy to have a chat if needed.. thanks!
aspina7 commented 4 years ago

hey hey

re sample size: agree necessary, has been in discussion for a while to create functions but was decided wouldn't fit within the initial funding (see this discussion for details)

re spatial sampling: yes this was also one of the things wanted to expand upon, have code for spatial sampling but needs updating and cleaning. A lot of this is quite simple now with the sf package using the st_sample function

re text changes: going to leave this for pat

aspina7 commented 4 years ago

so this is the website I mentioned with some sample size functions (and equations) - but seems to be only basic ones, not for survey designs. http://powerandsamplesize.com/Calculators/