Dichotomous Covariates in Meta Analysis

[christopherhillcsusb]

I am attempting to run a covariate in the meta analysis function that is dichotomous. I have coded the variables as 0 or 1 in order to obtain an interpretable effect size estimate. I have done this in the past in JASP but it is giving me an error today. I have attached the screen shot for reference.

[akashrajkn]

@christopherhillcsusb Thank you for reporting the issue. Which operating system are you using? And, can you attach the corresponding .jasp file?

[christopherhillcsusb]

I am using a Mac OS and the file is attached.
Meta Analysis Combined Effects.jasp.zip

[akashrajkn]

@christopherhillcsusb The analysis is not able to handle the header Measurement (1 =Device) which has special characters. We'll look into fixing this; for now, you can run the analysis if you change the header names to only include alphanumeric characters. (Attached csv: I renamed the headers)

meta_analysis_combined_effects.zip

[AlexanderLyNL]

The problem here is the meta-analysis does not use base64 encoding for the covariates. The issue is solved by changing the headers to standard text, see also:

#193

[AlexanderLyNL]

@christopherhillcsusb if you'd like to regress the meta-analysis on a dichotomous variable, please add the variable to the "Factors" box. On the other hand, if you'd like to regress on a continuous variable, then you should add it to the "Covariates" box.

Cheers,
Alexander

[christopherhillcsusb]

@AlexanderLyNL and @akashrajkn thank you for the help. I appreciate it.

[akashrajkn]

Fixed in PullRequest: jasp-stats/jasp-desktop#2932