Shih-Chieh Lee, PhD, is Postdoctoral Researcher, School of Occupational Therapy, College of Medicine, National Taiwan University, Taipei, Taiwan.
Gong-Hong Lin, PhD, is Assistant Professor, Master Program in Long-Term Care, College of Nursing, Taipei Medical University, Taipei, Taiwan. At the time this article was submitted, Lin was Postdoctoral Researcher, School of Occupational Therapy, College of Medicine, National Taiwan University, Taipei, Taiwan.
Am J Occup Ther. 2021 Jan-Feb;75(1):7501205140p1-7501205140p11. doi: 10.5014/ajot.2020.043463.
The most frequently used measures of facial emotion recognition (FER) are insufficiently comprehensive, reliable, valid, and efficient; moreover, the impact of gender on scoring has not been controlled.
To develop a computerized adaptive test of FER for adults with schizophrenia.
First, we selected photographs from a published database. Second, items that fitted well to a Rasch model were used to form the item bank. Third and last, we determined the best administration mode for prospective users to achieve both high reliability and efficiency.
Psychiatric hospitals and the community.
Adults living with schizophrenia (n = 351) and adults without diagnosed mental illness (n = 101).
After removal of misfit items (infit or outfit ≥1.4), the remaining 165 items were selected to form an item bank. Among them, 39 showed severe gender bias, so the item difficulties were adjusted accordingly. On the basis of the item bank, two administration modes were recommended for prospective users. The reliable mode required approximately 128 items (nearly 20 min) to achieve reliability (.72-.81), similar to that of the entire item bank. The efficient mode required approximately 73 items (approximate 11 min) to provide acceptable reliability (.69-.73) for the seven domain scores.
Our newly developed measure provides comprehensive, valid, and unbiased (to examinees' gender) assessments of FER in adults living with schizophrenia. In addition, the administration modes can be flexibly changed to optimize the reliability or efficiency for prospective users.
This newly developed FER measure can help occupational therapists identify deficits in recognizing specific basic emotions and plan corresponding interventions to manage the impact on their clients' social functions.
目前最常使用的面部情绪识别（FER）测量方法不够全面、可靠、有效；此外，评分的性别影响尚未得到控制。
为精神分裂症患者开发成人面部情绪识别的计算机自适应测试。
首先，我们从已发表的数据库中选择照片。其次，适合 Rasch 模型的项目被用来形成项目库。第三，也是最后，我们确定了最适合预期使用者的管理模式，以实现高可靠性和高效率。
精神病院和社区。
精神分裂症患者（n = 351）和未确诊精神疾病的成年人（n = 101）。
剔除不适合的项目（拟合或 outfit ≥1.4）后，剩余的 165 个项目被选入项目库。其中，39 个项目存在严重的性别偏差，因此相应调整了项目难度。在此基础上，为预期使用者推荐了两种管理模式。可靠模式需要大约 128 个项目（约 20 分钟）来达到可靠性（.72-.81），与整个项目库相似。有效模式需要大约 73 个项目（约 11 分钟），即可为七个领域分数提供可接受的可靠性（.69-.73）。
我们新开发的测量方法为精神分裂症患者的面部情绪识别提供了全面、有效和公正（考察性别）的评估。此外，管理模式可以灵活改变，以优化预期使用者的可靠性或效率。
这个新开发的 FER 测量方法可以帮助职业治疗师识别识别特定基本情绪的缺陷，并计划相应的干预措施，以管理其客户社交功能的影响。
