Regression NaN

[ray]

Hello,

This is a weird one. If anyone has any ideas please let me know.

I'm running a linear regression and one of my variables is marked as NaN in the results table (and therefore also making the post hocs all come up NaN).

I don't think the variable is in any way different from the other variables that are working: the data type is the same, the distribution/linearity is roughly the same, the variable name has no spaces or characters, the data hasn't any more zeros that the other variables. I'm running the latest software. Then I tried running another linear regression and this time none of the variables would go into the table. I'd select dependent and covariates, the little blue loading symbol would spin for a second and then nothing. Re-started everything and the same sequence of events happened. Used the browser version and the same thing happened again.

The next thing I can think of trying is redoing the original spreadsheet but it would be a copy and paste job so unless I know whats wrong I'll copy the mistake too.

Any help is much appreciated. I'm getting to the talking-to-the-screen stage of frustration.

Ray

[reason180]

I see that this happens for a predictor that is perfectly correlated with another predictor.

tempPerfCorr.png (37.51 KiB) Viewed 2183 times

[ray]

Thanks for the reply, it's not perfectly correlated with another variable.

I've retyped all the data in by hand but its still coming up the same.

[reason180]

Can you post a copy your jamovi file, or a copy of a similar jamovi file that shows the NaN error?