Construct Invalidity Comes from Method Effects

Method effects are the sources of the construct invalidity in social-behavioral measuring procedures. Some method effects come from inadequacies in the protocol-the sequence of behaviors that is selected and identified solely by the set of conditions under which it occurs. The protocol may be inappropriate to the construct as specified or it may be insufficient because its conditions restrict the range of possible manifestations of construct-relevant behavior. Other method effects are technique effects-errors of commission occurring when a measuring operation is applied to a protocol. These errors may come from the actor's reactions to being measured, from the data-producer, or from the instrument used by the data-producer. Although detecting the presence of method effects is relatively easy, identifying specific effects is more difficult. Such identification usually requires the rescoring or recoding of the data on attributes suspected of contributing systematic but unwanted effects. Method effects also can arise from interactions among such conceptually irrelevant variables or between one of them and the construct of interest. When investigators can understand and improve their methods, they can more adequately understand and interpret their empirical findings.