Fixed or Random Effects for Panel Data of countries?º

This is a more theoretical/statistics question rather than coding. I have aggregate data of countries from the World Values Survey (WVS) that results in an unbalanced panel dataset with 300 observations of 85 countries.

I want to see the effect of GDP on a DV, so using the plm() and glmer functions I have used the followiung specifications:

model1 <- plm(dv ~ gdp, data=df, index=c("country", "year"), model="within")
model2 <- plm(dv ~ gdp, data=df, index=c("country", "year"), model="random")
model3 <- glmer(dv ~ lgni + (1|country) + (1+lgni|country), data=df)

The Hausman test from models 1 and 2 suggests that I should use random effects (0.67).

I have a few short questions, because I have read kind of contradictory things about FE and RE. Based on the dataset, unbalanced panel data of the same countries observed over time, should I estimate FE or RE?

Under which conditions should I use random slopes? Would it apply in this case? Thanks.