Validity

Validity

Introduction

Validity is the most critical component of measurement integrity. Simply defined, measurement validity is the degree to which meaningful inferences may be drawn from test data. Although measurement validity is a unitary concept, social scientists have developed a variety of approaches for gathering validity evidence, including content-related and construct-related approaches. Assessment of content-related validity is extremely important in test development and is not to be confused with “face” validity (i.e., the degree to which the content of test items appears prima facie to reflect a content domain). Face validity has typically been treated pejoratively by experts in the field of social science measurement compared with the process of content validation, which is considered an integral part of instrument development.

Discussion

Validation entails ascertaining the degree to which multiple lines of evidence are consonant with the intended inference (or “construct”), while establishing that alternative inferences are less well supported. This judgment takes into account the meaning of test scores as well as their relevance and utility for particular applied purposes, their value implications, and the social consequences of using them for applied decision making. Messick's definition of validity signals that validity is a property of inferences based on test scores, not tests themselves; that it is a matter of degree, not all or none; and that it evolves as new evidence is brought to bear (Shepard, 1993).

He uses the term test to refer not just to tests as typically conceived but to any means of observing or documenting consistent responses, behaviors, or attributes; scores subsume qualitative and quantitative summaries of these consistencies and can refer to groups, situations, or objects, as well as persons. The term construct refers not just to unobservable qualities assumed to account for performance on a test, as originally conceived, but to whatever con-cept(s) the test is designed to measure. Thus, construct, score meaning, and intended inference or interpretation can be used interchangeably, as can construct validity and validity . Construct theory encompasses (a) the definition of the construct (or “construct domain”) that specifies the boundaries and facets of the construct as well as (b) the expectations (hypotheses) about responses processes, consistencies in item responses, relationships among variables, and so on (Messick, 1996).

Internal Validity

Internal validity refers to the accuracy of statements made about the causal relationship between two variables, namely, the manipulated (treatment or independent) variable and the measured variable (dependent). Internal validity claims are not based on the labels ...