Johnson

A single test score feels definitive. You see an 82 on a math exam and assume that number tells the whole story. But every test score carries measurement error, and a student who scored 82 today might have scored 76 or 88 under slightly different conditions. That is where a confidence interval single test score … Read more

Developing a new questionnaire means nothing if your instrument does not actually measure what it claims to measure. That is where the content validity index comes in. If you are learning how to run a content validity index for a new questionnaire, this guide walks you through every step with real numbers, formulas, and interpretation … Read more

Every researcher who has ever developed a questionnaire, selected a psychological assessment, or designed an educational test has faced the same fundamental question. Does this instrument actually work? Understanding criterion validity provides the answer. Criterion validity is one of the most important concepts in psychometrics and research methodology because it tells you whether your measurement … Read more

If you have ever run a survey or built a measurement scale, you have probably reported Cronbach’s alpha as proof your items hang together. But many researchers in are switching to composite reliability, and for good reason. Composite reliability accounts for something alpha quietly assumes away: the fact that your items rarely relate to the … Read more

Item-total correlation is one of the most informative statistics you can compute when developing or refining a measurement scale. It tells you whether each individual item on your questionnaire actually contributes to the construct you are trying to measure. When you are cleaning a scale, this single number helps you decide which items to keep, … Read more

You ran your Cronbach’s alpha in SPSS, and the number staring back at you is 0.62. Not great. You notice the “Alpha if Item Deleted” column shows that removing one or two items could push that number to 0.75 or higher. Now comes the question every researcher faces at some point: do you drop those … Read more

If you have ever developed a questionnaire, an educational assessment, or any multi-item scale, you have probably wrestled with one stubborn question: how many items do I actually need for reliable scores? The answer is not intuitive, and doubling your test length does not double your reliability. That nonlinear relationship is exactly what the Spearman-Brown … Read more

If you have ever built a survey, a questionnaire, or a psychological test, you have probably asked yourself one question: can I trust these scores? That is the core question behind split-half reliability, one of the oldest and most intuitive methods for checking whether a measurement instrument is internally consistent. Split-half reliability divides a test … Read more

Structural equation modeling (SEM) lives or dies on one foundational assumption: the constructs in your model must measure genuinely distinct concepts. That assumption is called discriminant validity, and the most reliable tool for testing it today is the heterotrait-monotrait ratio of correlations, or HTMT. If you are working through a measurement model for a thesis, … Read more

Measurement invariance is a psychometric property that assesses whether a construct is measured in the same way across different groups or time points. When researchers compare groups, they need to know that observed score differences reflect true differences in the underlying trait, not artifacts of how the instrument behaves in each group. Without this equivalence, … Read more