Measure what matters, and measure it well. Learn how psychological tests are built and checked, try a reliability calculator with real numbers, and browse a library of free and open instruments for research in psychology, counseling, mental health and education.
Psychometrics is the science of measuring psychological qualities such as anxiety, belonging or conscientiousness. Because these cannot be seen directly, researchers build instruments and then gather evidence that the scores are consistent and mean what they claim to mean.
A construct is the idea being measured, such as "academic stress." An operational definition states exactly how it will be measured, such as total scores on the Perceived Stress Scale. A clear construct definition guides every later choice.
Reliability is a property of scores in a particular sample, so report it for your own data.
Validity is an argument built from many kinds of evidence (Messick, 1995).
Follow the developer's scoring rules exactly, including reverse-scored items (on a 1–5 scale, a reversed 2 becomes 4). Norms compare a score with a reference group; cutoffs flag scores worth a closer look. A cutoff from one population may not fit another.
A test can work differently across languages and cultures. Good practice includes careful translation and back-translation with community review (International Test Commission, 2017), checking measurement invariance before comparing group means, and looking for items with differential item functioning. Also ask whether the construct itself means the same thing in each culture.
See the Standards for Educational and Psychological Testing (AERA, APA, & NCME, 2014).
Paste survey responses: one row per person and one column per item, separated by spaces, commas or tabs. The sample below shows 12 people answering a 6-item scale (1–5). Item 4 is worded negatively, so try reverse-scoring it.
Your data stays in your browser. Nothing is uploaded.
Measures you can use in psychology, counseling, mental health and education research without buying a license. Each entry describes what the instrument measures, its length and audience, how it can be used, and the original citation.
Read the items and the developer's definition. "Stress," "distress" and "burnout" are different constructs.
Check age, language, culture, and clinical or community samples in the validation studies.
Look for internal consistency and, where relevant, test-retest evidence in published studies.
A measure built for research may not be suited to individual decisions or diagnosis.
Use published, validated translations rather than translating on your own when possible.
Confirm it is free for your purpose and whether you may change items or put it online.
Short forms reduce fatigue and missing data, especially in combined surveys.
Plan support and referrals for questions about suicide, trauma, abuse or discrimination, and get ethics board approval.
Short talks and peer-reviewed articles published in the last ten years (2016 or later) that support this lesson. Videos and article links need an internet connection.
How early intelligence tests were misused to justify eugenics and immigration limits, and why test fairness matters.
Why careful measurement, transparency and replication make findings trustworthy.
A ready-to-teach lesson that takes students from constructs to reliability analysis and fair instrument selection.
