jamovi

Hi, I hope this will make sense, I´m hoping for a little guidance on missing values and how to treat them within my data set.

I have a dataset imported from Qualtrics in which three categorical questions were made conditional, so I now have two columns that are half-full. I need to run statistics using these two variables as the outcome but if treated as normal missing values in Jamovi it says sample size of 21 vs when treating them as a -99 value the full sample of 51.

Do I need to compute/transform, set up my data differently, add in values or is it correct to just leave them as missing for Jamovi not to use them?

Thank you!

hi,

this sounds like jamovi is doing the right thing, but come back to me if i'm misunderstanding.

if you've got a dataset with 51 rows, and 30 of those values are -99, you designate the -99's as missing values, that leaves you 21 observations.

you can't use a "full sample" of 51 values if 30 of them are missing.

jonathon

Hi Jonathon,

Thank you. I assumed Jamovi was just being smart and doing its thing, but was initially advised to designate the value to run analyses on the groups since they are system-missing (design) values.

I suppose that will be why my chi-square tests and a few others are returning NaN as the sample then becomes too small or comparing the two against each becoming impossible as the sample becomes zero when the missing are disregarded in both variables (since they are the opposite of each other from the conditional question)