Psychometric Characteristics of Assessment Procedures

Abstract

In this chapter by Wasserman and Bracken, the most salient psychometric characteristics of psychological tests are described, incorporating elements from both classical test theory and item response theory. Guidelines are provided for the evaluation of test technical adequacy across a wide array of psychological tests in the areas of sampling, norming, scaling, validity, reliability, and fairness. Sampling, norming, and scaling guidelines address appropriate sample sizes, the accuracy and recency of norms, item and scale gradients, and floor and ceiling effects. Evidence of test score validity is described as occurring in two broad classes (internal and external), both of which ultimately are concerned with construct validity. Evidence of test score reliability includes the extent to which measurement results are precise and accurate; free from random and unexplained error; and consistent, accurate, and uniform across occasions, time, observers, and samples. Fairness is described as an important additional area of psychometric emphasis. A holistic and systemic approach to test score fairness is described, extending from test conception through applied consequences. Finally, the study of psychometrics is criticized for its historical over‐reliance on internal sources of evidence, and recommendations are made for an increased focus on external, applied sources of evidence of psychometric adequacy.

IRRODL Volume 14, Number 4

“…Cronbach's alpha for "before" (0.6870) is higher than the value suggested by Wasserman and Bracken (2003). Cronbach's alpha for "after" (0.9332) exceeds the values suggested by Wasserman and Bracken (2003) and George and Mallery (2003).…”

“…Quantitative data were collected from distance students using a Likert scale. Cronbach's alpha is 0.8612 and 0.9487 for the "before the class" and "after the class" data, respectively, both higher than the values suggested by Wasserman and Bracken (2003) and George and Mallery (2003).…”

“…Cronbach's alpha is a coefficient that is used for ordered rating scale survey instruments, such as a Likert scale that measures participants' attitudes. Wasserman and Bracken (2003) suggested that Cronbach's alpha of 0.60 or higher indicates acceptable coefficient of internal consistency for group assessment. George and Mallery (2003) proposed that Cronbach's alpha of 0.7-0.8 is acceptable; 0.8-0.9 is good; and ≥0.9 is excellent.…”

Adaptation of the Scale Knowledge Perception Scale in Primary Education

“…De notar que vários autores recomendam que os valores da consistência interna deveriam ser acima de 0.5, tendo definido que valores entre 0.5 até 0.7 são medíocres, de 0.7 até 0.8 são bons, valores entre 0.8 e 0.9 são muito bons e acima de 0.9 excelentes (Costello & Osborne, 2005;Hutcheson & Sofroniou, 1999;Kaiser, 1974;Pestana & Gageiro, 2008;Sharma, 1996;Wasserman & Bracken, 2003).…”

Do Cyber Victimization and Traditional Victimization Form Separate Factors? Evidence from a Preliminary Study

“…Internal consistency was acceptable for all variables, ranging from .74 for cyber victimization to .86 for direct victimization [19]. To compare the relative frequency of the three different types of victimization, total scores for the three victimization scales were calculated as the average rating across all items in the scale.…”