Test-retest reliability is when someone or something is tested more than once in the same way (Myers & Hansen, 2012) while inter-rater reliability is one test but is observed by two different people during that test. The difference is mainly focused on how many times the test is being done.
Content validity is focusing on whether the sample we are using of the variable is fair or not and checking that it represents what we want it to represent. Predictive validity is seeing if we can predict the validity by looking at the information we have. Construct validity is the checking of the overall and validity of the information, the concrete information (Myers & Hansen, 2012). The main difference between predictive and construct is that one is the prediction or the possibility, and the other is the overall validity of the information that is given.
Internal validity focuses on the cause and effect of the study, that there is no reason or explanation for the findings besides what is given in the study, while external validity focuses on how the study can be applied to the world outside of the study or in multiple ways (Myers & Hansen, 2012). The main difference is internal focusing on the actual study, while external focuses on uses outside of the study.
A test can be reliable without being valid. Since reliability is focused on the outcome of the study, it does not have to be valid. On the other hand, if a test is not reliable, it cannot be valid. Since the test in the second question, is not reliable, it cannot be valid due to not having that consistency or reliability that would be needed.