Missing pre- or post- data

[Mitch Henderson]

Hi team,

Loving this site, So useful!

I have paired data (pre-post) but for some participants we weren’t able to get either their pre- or post- measures (leaving only one measure).

I see that the paired student’s t-test (and associated paired estimation plot) requires arrays of the same length on estimationstats.com. Would this be the same if I provided data with NA’s into {dabestr} in R?

Do you have any advice regarding how to navigating the issue of missing data such as this?

[Joses Ho]

Hi Mitch,

Thanks for your comments and your question.

Unfortunately, your problem of missing data can't be solved by including NAs in the paired data. The best way forward, I think, is to only retain participants with pre-post responses. If that leaves out too many data points, then you can perform an unpaired analysis instead.

Hope this helps!

Joses