meta-analysis blog post

ggupta2 / matcha_codejam

fun codejam with Charles and Matt

1 stars 1 forks source link

meta-analysis blog post #4

Open softloud opened 3 years ago

softloud commented 3 years ago

for Canadian pdoc application

example from Charles' job application

[ ] choose a dataset from metafor:: (meta-analysis package)
[ ] recreate the analysis
[ ] explain to oneself what is happening
[ ] interpret the results
[ ] browseVignettes("metafor")

softloud commented 3 years ago

And create a blog to put it on

options

distill:: - for science
blogdown:: - most conventional
hugodown:: - in development but better than blogdown in some ways
pkgdown::- for packages

set up blog

Go through Chapter 1; it ends with a recommend workflow for setting up a blog.

[ ] create the blog
[ ] create the post
[ ] deploy on github pages

softloud commented 3 years ago

Note the "vignettes" for metafor:: are in the massive pdf on the package. Not friendly to find. But if you can bear drilling in to find one example, just focus on that, recreating the code.

Another option is to look for someone else blog post on meta-analysis and recreate that.

softloud commented 3 years ago

Another tip. Look for data that has control & treatment continuous measures, mean and sd. Easiest to do a standardised mean difference meta-analysis.

softloud commented 3 years ago

In particular three functions from metafor:: are of interest:

::escalc for calculating the effect and the variance of the effect (odds-ratio, risk-ratio, standardised mean difference)
::rma for random effects meta-analysis, i.e., the modelling function
::forest for plotting the forest plot