Ordinal logistic in JASP: sign convention of cumulative logit (α_j-βX or α_j+βX)?
> I'm a Japanese university student currently using JASP to analyze the self-interest hypothesis of policy support. I'd like to know which parameterization (α_j-Xβ or αj_+Xβ) of the right-hand side of the cumulative logit used in JASP's Ordinal Regression, a type of GLM. I couldn't find any information on Github, but if R's mass:polr is used, is a negative sign correct? I've tried inputting variables whose inter-variable relationships are easily visible, and at this point, I think a positive sign is correct. However, I'd like to know more accurate information, so I'm posting this question.I apologize for my incomplete English due to translation issues. I look forward to reply.thank you for reading.
> respectfully,
Comments
This discussion is continued at https://forum.cogsci.nl/discussion/9968/interpreting-the-signs-of-glm-ordinal-logistic-regression#latest