Detecting Construct-Irrelevant Variance: A Comparison of Network Psychometrics and Traditional Psychometric Methods Using the HEXACO-PI Dataset

Construct-irrelevant variance (CIV), defined as excessive variance that is unrelated to the intended construct, poses a significant threat to the validity of test interpretations and applications. CIV can arise from two notable sources: construct-irrelevant items, which include items with content unrelated to the construct being measured, and redundant items, which repeat information already captured by other items, leading to individual fatigue and inflated reliability estimates. This study explores the detection of CIV using network psychometrics in comparison with traditional psychometrics methods. The study utilizes the HEXACO-PI dataset, focusing on six personality dimensions, and applies network psychometrics techniques such as Exploratory Graph Analysis and Unique Variable Analysis. These techniques are compared against traditional methods, such as reliability and factor analysis, to assess their effectiveness in identifying CIV items. Results highlight the potential of network psychometrics as a complementary approach for enhancing the structural integrity of psychological instruments, with 89% agreement in the number of identified CIV items. This comparison provides insights into alternative methods for improving psychological instruments, with implications for the future of psychometric evaluation.