Hi,
First of all, thank you for jamovi. I have been using it with my B.A. students for the past 2 years and it's great.
I would like to use it also with my graduate students, but for that I need to estimate panel data models (pooled OLS, fixed effects, and random effects). This requires that jamovi allows for panel data models, i.e., in which a given individual appears in different moments of time. I suppose it is what some of you call long format. Is it possible to have this in jamovi? I have also tried to install the plm package for R in jamovi, but sadly I failed. Does/will jamovi have this feature? If not, any ideas on how can I do this with jamovi?
Thank you for your help.
Panel data / Long format in jamovi
Re: Panel data / Long format in jamovi
GAMLj3 supports long format data. will that do what you need?
jonathon
jonathon
Re: Panel data / Long format in jamovi
Thank you. I have read about that jamovi module and, if I am not wrong, that allows for interesting models but different from the ones I am interested in. In panel data analysis (usual in economics, management, international business or international relations), fixed and random effects models are used to control for unobserved, time-invariant characteristics between units (e.g., individuals, firms, cities). Fixed effects assume these characteristics are fixed for each unit, while random effects treat them as random variables drawn from a distribution, assuming that the random effects are not correlated with the observed variables.
This is different, from what I understand, from the models with fixed or random coefficients that are allowed by GAMLj3, although both assume we have several observations for each individual (long format). I know there is a lot of information regarding these panel data models in the internet; yet, the following pdf covers the matter in a very nice way (and shows how to implement it in R) and may be helpful to understand what I mean: https://www.schmidheiny.name/teaching/panel2up.pdf
If you could help me understand how can I do these models in jamovi, that would be greatly appreciated.
This is different, from what I understand, from the models with fixed or random coefficients that are allowed by GAMLj3, although both assume we have several observations for each individual (long format). I know there is a lot of information regarding these panel data models in the internet; yet, the following pdf covers the matter in a very nice way (and shows how to implement it in R) and may be helpful to understand what I mean: https://www.schmidheiny.name/teaching/panel2up.pdf
If you could help me understand how can I do these models in jamovi, that would be greatly appreciated.