Terms from Statistics for HCI: Making Sense of Quantitative Data

The extent to which the results of an experiment or study can be recreated by others. The language used for this varies, but the ASA distinguishes between reproducibility and replication. If others can use your data, etc., to obtain the same published result, that is reproducibility. For replication, the question is: if others follow your procedures with fresh subjects, etc., they will come to a similar result? Furthermore, some writers use the term 'repeatability' in a very specific sense: does the same scientist in the same laboratory with the same equipment get identical results? Some of these distinctions are far less clear in Human–Computer Interaction and human sciences; for example, due to learning effects if the same subject repeats a task we do not expect a similar outcome.

Used on page 138