Population indepenedence

How do I identify if two population samples are independent? Is a good test to think like: Given the mean of a variable in population one, it must not predict the mean of any variable in population two? Would really appreciate practical tips to identify this in real life, thank you!