In practice, regressions in the social sciences are more likely than not to violate at least one of the assumptions for linear regression. Nonetheless, you should be aware of (some of) the assumptions and be able to identify conditions where they would not be met.