A model and its fit lie in the eye of the beholder: Long live the sum score

Frontiers in Psychology

article

Peter Adriaan Edelsbrunner¹ 

[1] Department of Humanities, Social and Political Sciences

关键词: Latent variable;    Sum score;    Formative measurement;    model equivalence;    Socio-constructivism;   

DOI  :  10.3389/fpsyg.2022.986767

学科分类：社会科学、人文和艺术（综合）

来源: Frontiers

【 摘 要 】

A common way to build scores for statistical analysis from psychological or educational scales (e.g., knowledge or intelligence tests, attitude or motivation questionnaires,) is to sum up participants' scores across all items. For example, in a questionnaire assessing knowledge about a specific topic, participants might have to answer different questions probing their content knowledge. The correct answers on all items are then summed up to build a score for analysis that is meant to represent participants' knowledge (Edelsbrunner et al., 2018). Similarly, on an intelligence test, correctly solved items might be summed to yield an overall score that is transformed into IQ estimates (Raven et al., 1962). On a questionnaire measuring need for cognition, participants might indicate their agreement with different self-descriptions on a Likert scale and their agreement is summed up (or a mean is built, which is equivalent for the arguments brought up here) across all items (Beißert et al., 2014).

【 授权许可】

CC BY