Construct valdity relates to whether a particular psychometric assessment...
Lead consultant at Test Partnership, Ben Schwencke, explains construct validity.
Construct validity relates to whether a particular psychometric assessment genuinely measures the psychological construct it intends to measure. For example, if a psychologist has developed a numerical reasoning assessment, construct validity is concerned with whether this assessment measures numerical reasoning, or instead measures another construct entirely. Psychologists consider construct validity to be the primary form of psychometric validation, and the question of construct validity is always the primary concern of any psychometric R&D process.
In practice, there are two main ways to display construct validity in psychometric assessments. Firstly, concurrent validity involves completing a specific assessment alongside another, pre-validated assessment designed to measure that same psychological construct. If the correlation between the two scores is sufficiently high, this suggests that these two assessments measure the same underlying psychological construct, providing evidence of construct validity.
Conversely, an assessment can be said to show discriminant validity if that assessment shows little to no relationship with scores on a supposedly unrelated assessment. For example, if a psychologist develops an emotional intelligence (EI) questionnaire, they may want to show that it doesn’t measure cognitive intelligence (IQ). Therefore, the psychologist gives both the EI and an IQ assessment to participants, and the subsequent scores show a very weak relationship. This suggests that these two assessments measure distinct psychological constructs, providing evidence of discriminant validity. Had the relationship between the two been strong, that could suggest that the EI questionnaire lacks construct validity, as it in actuality measures IQ instead.