Reliability refers to the consistency of a measure of a concept. There are three factors researchers generally use to assess whether a measure is reliable:
- Stability (aka test-retest reliability) – is the measure stable over time, or do the results fluctuate? If we administer a measure to a group and then re-administer it and there is little variation in the results changed over time, the measure can be staid to have ‘test-retest reliability.
- Internal reliability – are the indicators which make up the scale or index of a measurement consistent ?If the score a respondents according to one indicator of a measure are consistently related to the scores they achieve according to other indicators for that same measure, then the measure can be said to have ‘internal reliability’.
- Inter-rater reliability – how much agreement is there over which observed empirical phenomena fit into what indicator? If researchers have a high level of agreement over how observed behaviour ‘map onto’ the indicators of a measure, then we can say the measure has a high level of inter-rater reliability.
Bryman, Alan (2016) Social Research Methods