So my data:

I have 172 countries, and a time span of 10 years. -> I have some missing values, so not all data is available for all countries for all years.

Coming from theory, I have the following variables:

1x DV (a continuous variable ranging from -2.5 to 2.5, values of i.e. -1.952458 are possible)

2x IV (both continuous ranging from 0 to 500)

1x IV (dummies: 4 categories)

1x IV/Mediating: continous, I am not sure, so I want to test whether its an IV or Mediating

2 moderating variables which affects all relationships between IV and DV. 1 moderating variable is a dummy (0,1,2,3) and one is a continous variable

and 4 control variables, all continuous

So, I understand from theory, I have to do the following:

-Mundlak model to test whether i have to use fixed- or random effects analysis

- Hausman test to confirm the above test

- either the fixed or random effects analysis

- and then?

I know from simple regression that if i want to test mediation I do the Sobel test or the Baron & Kenny test. But how do i do that now?

I also know how to moderation in simple regression, but how does it works now?

And lastly, I played a little bit with a simple model in fixed effects, the output looks very much different then from simple regression, how does one interpret?

If possible, I would like to do the Vector AutoRegression, so I can say *something* about probability of possible causality.

So, long story short, are all tests mentioned and/or needed even possible in SPSS?

If so, in what order should I proceed?

how do fix my mediation/moderation problem?

in other words: what should I do?

Thanks for reading, and for any reply