In much of social research, the variables used are

Much of social research uses non-manipulable variables that cannot be actively altered by the researcher.

Traits like caste, income or attitudes exist prior to the study.

They are not manipulable, since the researcher cannot assign them.

They are not experimental, because no treatment is imposed on them.

Deviational is not a recognised variable type here.

Such variables are studied through correlational and survey designs.