Reliability is the consistency of the experiment / instrument.
Internal reliability refers to how consistent the items within the experiment / instrument all relate to each other and if they are all equally testing the same thing. For example, if you're given a questionnaire about sleep deprivation, if it has internal reliability, all of the questions should relate to sleep deprivation.
Construct validity refers to if the method of the experiment / instrument reflects the theory that is desired to be tested. For example, if you're testing the theory that there is a capacity of 5 - 9 items in short-term memory, your experiment will have construct validity it is designed in a way that supports this theory, and isn't instead testing the duration of short-term memory.
External validity refers to if the results of the experiment can be appropriately generalised to the population and other similar research/studies.
Does that make sense?
