Reliability, If a researcher administers a questionnaire measuring political conservatism today and again two weeks later, expecting political orientation to be stable, the degree of correlation between the two sets of results indicates the stability of the measure., A researcher developing a scale to measure anxiety might use ten questions. By randomly splitting these questions into two groups of five , they check if respondents score consistently across both halves. If the halves yield different results, the measure suffers from a problem of reliability., In a study evaluating the feasibility of using thoracic ultrasound during flight, researchers must account for the high potential for variable reliability in clinicians interpreting the point of care ultrasound (POCUS). To mitigate this, they implement standardised training for all clinicians to ensure consistent application of the measurement procedure., Validity, If attempting to measure a concept like "mathematical ability," the test must cover the full range of meanings associated with that concept (e.g., addition, subtraction, multiplication, division), not just addition. Similarly, measuring "attitudes to school" requires ensuring questions are appropriate, such as avoiding questions about parental relationships., The College Board exams (such as the SAT) is demonstrated by their ability to predict the subsequent college success of students., A researcher measures "marital satisfaction" and hypothesizes that satisfied partners are less likely to cheat on their spouses. If the measurement relates to marital fidelity in the expected manner, this provides evidence of construct validity. If satisfied partners are as likely to cheat as dissatisfied ones, the validity of the measure is challenged. In educational research, establishing construct validity may involve checking if hypothesized subscales (e.g., in a self-concept scale) exist and are measured by the variables intended.

Reliability & Validity

由
打印
嵌入

排行榜

视觉风格

选项

切换模板

恢复自动保存：？