College of Education, Criminal Justice, and Human Services, School of Human Services, University of Cincinnati, Cincinnati, OH 45221, United States of America.
College of Education, Criminal Justice, and Human Services, School of Human Services, University of Cincinnati, Cincinnati, OH 45221, United States of America.
J Subst Abuse Treat. 2020 Aug;115:108041. doi: 10.1016/j.jsat.2020.108041. Epub 2020 May 28.
Many studies have employed the Short Alcohol Attitudes Problem Perception Questionnaire (SAAPPQ) to measure professionals' attitudes toward working with people experiencing alcohol problems. However, research has not confirmed the internal structure of the SAAPPQ, leaving open the possibility that scale scores are not valid empirical proxies for underlying attitudes. It is also unclear if measurement invariance holds over time or across groups. This is an important gap in light of evidence that (a) experiences can change the structure of scales and (b) structures of some popular scales differ dramatically between groups. Thus, measurement bias may confound comparisons of composite SAAPPQ scores between measurement occasions or groups. To address these gaps in the literature, we conducted a longitudinal examination of the psychometrics of the SAAPPQ using data from 241 human services professionals-in-training. We found a four-factor structure, rather than the previously reported five factors. Configural, metric, and a strong degree of scalar invariance held longitudinally. Few factor intercorrelations were significant at time one and they strengthened over time, suggesting that the SAAPPQ should not be aggregated at higher-order levels (i.e., global scores should not be used). Findings suggest four SAAPPQ subscale scores may be used as valid proxies for attitudes toward alcohol users; however, additional confirmatory studies are needed to ensure that measurement bias does not compromise inferences based on SAAPPQ scores.
许多研究都采用了《短式酒精态度问题知觉问卷》（SAAPPQ）来衡量专业人士对与有饮酒问题的人合作的态度。然而，研究尚未证实 SAAPPQ 的内部结构，这使得量表分数是否是潜在态度的有效经验代表变得不确定。也不清楚测量的不变性是否随时间或群体而变化。鉴于以下证据，这是一个重要的差距：（a）经验可以改变量表的结构，以及（b）一些流行量表的结构在群体之间存在显著差异。因此，在不同测量时间或群体之间，测量偏差可能会混淆对复合 SAAPPQ 分数的比较。为了解决文献中的这些差距，我们使用 241 名人力资源服务专业人员的培训数据，对 SAAPPQ 的心理测量学进行了纵向研究。我们发现了一个四因素结构，而不是以前报告的五因素结构。配置、度量和很强的标度不变性在纵向具有一致性。很少有因素间相关性在时间上是显著的，而且随着时间的推移它们会增强，这表明 SAAPPQ 不应该在更高阶水平上进行聚合（即，不应该使用总体分数）。研究结果表明，四个 SAAPPQ 分量表分数可以作为对酒精使用者态度的有效代表；然而，需要进一步的验证性研究来确保测量偏差不会影响基于 SAAPPQ 分数的推论。
