Department of Psychology, University of Miami, Coral Gables, FL, 33124, USA.
Department of Educational Methodology, Policy, and Leadership, University of Oregon, Eugene, OR, 97403, USA.
Adm Policy Ment Health. 2021 Sep;48(5):921-935. doi: 10.1007/s10488-021-01139-1. Epub 2021 Apr 30.
Pragmatic instruments with psychometric support are important to advance dissemination and implementation (D&I) research, but few well-researched D&I instruments exist. Item response theory (IRT), an approach that is underutilized in D&I, can help with the development of actionable and brief instruments. This paper provides an overview of IRT for D&I researchers and examines an instrument of therapist attitudes using IRT measurement models. Eight items of the Attitudes Towards Individualized Assessment-Monitoring and Feedback (AIA-MF) Clinical Utility scale were fit to the Graded Response Model in a national sample of master's level therapists. Various IRT model characteristics including item threshold and discrimination parameters, information, and item and person fit were examined. Discrimination and thresholds parameters showed significant variability across the eight items. Item information curves also showed that each item contributed variably to the total test information, suggesting that items 4 and 5 reliably measure therapist attitudes across the latent continuum and items 3 and 6 warrant further investigation. Results suggest that IRT models can help D&I researchers examine existing instruments with greater specificity than traditional measurement methods, thus increasing measurement precision while lowering response burden, both important considerations for the field.
实用工具和心理计量学支持对于推进传播和实施(D&I)研究很重要,但具有良好研究基础的 D&I 工具却很少。项目反应理论(IRT)是一种在 D&I 中未得到充分利用的方法,它可以帮助开发可操作和简洁的工具。本文为 D&I 研究人员提供了 IRT 的概述,并使用 IRT 测量模型检查了一种治疗师态度的工具。在全国范围内的硕士水平治疗师样本中,使用等级反应模型拟合了《个体化评估-监测和反馈的态度量表》(AIA-MF)临床实用性量表的 8 个项目。检查了IRT 模型的各种特征,包括项目阈值和区分参数、信息以及项目和人员拟合。辨别力和阈值参数在这 8 个项目之间存在显著差异。项目信息曲线还表明,每个项目对总测试信息的贡献各不相同,这表明项目 4 和 5 在潜在连续体上可靠地衡量了治疗师的态度,而项目 3 和 6 需要进一步研究。结果表明,IRT 模型可以帮助 D&I 研究人员更具体地检查现有工具,从而提高测量精度,同时降低响应负担,这两者对于该领域都很重要。