//
sign in
Post
by @danabra.mov
PostEmbed
by @danabra.mov
Record
by @jimpick.com
Record
by @atsui.org
+ new component
Post
Indeed! Sigma divided measures are terrible. The denominator can change from study to study in unpredictable ways that don’t carry any useful causal message. They are only, if ever, justified if the outcome scales vary from study. They should be regarded as a measure of last resort.