| Sumario: | The writer considers the use of the null hypothesis significance test, as usually carried out using the t test of the difference between means, and he refutes those who say it is an impediment to progress in psychology. He points out that the t test, based on the sample—not the true—standard deviation, does not apply exclusively to distance between means. He argues that statistical procedures were created to standardize and reveal the logic underlying the conclusions reached about data, and though accompanied by error, the error is likely to be less than that based on previous subjective procedures. He contends that in view of the usual, but limited, evidentiary, not probative, role of the null hypothesis test, prohibiting its use in publications will do solely that; researchers who regard it as useful will continue to use it, but they will not refer to it in manuscripts submitted for publication.
|
|---|