Reliability is related to the consistency of a measurement. Psychologists look at three types of consistency: over time (reliability of tests), beyond the elements (internal consistency) and beyond different researchers (inter-board reliability). Imagine as an informal example that you have been dieting for a month. Your clothes seem softer, and several friends have asked if you have lost weight. If your bathroom scale indicated in this place that you had lost 10 pounds, it would make sense and you would continue to use the scale. But if that suggests you took 10 pounds, you`d be right to conclude that it was broken and that it was repaired, either get rid of it. When evaluating a measurement method, psychologists take into account two general dimensions: reliability and validity. For example, if two researchers observe “aggressive behaviors” of children in kindergarten, they would both have their own subjective opinion about what the aggression entails. In this scenario, they would be unlikely to record aggressive behaviors in the same way and the data would be unreliable. A typical assessment would involve giving participants the same test on two different occasions.

If the same or similar results are obtained, external reliability is established. The disadvantages of the retest testing method are that it takes a long time to get results. If researchers measure a construct that they think is consistent over time, the results they get should be consistent over time. The reliability of test tests is the extent to which this is actually the case. For example, intelligence is generally thought to be consistent with time. A very smart person today will be very smart next week. This means that every good level of intelligence should provide next week pretty much the same values for that person as they do today. It is clear that a measure that produces very inconsistent values over time cannot be a very good measure of a construction that must be consistent. For example, the Minnesota Multiphasic Personality Inventory has subscales that measure different behaviors such as depression, schizophrenia, social introversion….