A measure of reliability of a scale or research instrument such as a questionnaire in which the results from half the items on a multi-item measure are compared with the results for the remaining items. Sometimes the division of items is made randomly;
while at other times even items are assumed to form one half and the odd the other half of the instrument. The total on the even items is then correlated with the total score obtained from the odd items. If there is substantial variation between the groups, the reliability of the instrument is in doubt.
