Campo-Arias Adalberto, Herazo Edwin, Oviedo Heidi Celina
Instituto de Investigación del Comportamiento Humano (Human Behavioral Research Institute), Bogotá, Colombia; Programa de Medicina, Facultad de Ciencias de la Salud, Universidad del Magdalena, Santa Marta, Colombia.
Instituto de Investigación del Comportamiento Humano (Human Behavioral Research Institute), Bogotá, Colombia.
Rev Colomb Psiquiatr. 2017 Jan-Mar;46(1):31-35. doi: 10.1016/j.rcp.2016.03.005. Epub 2016 May 13.
The process of evaluating measurement scales is an ongoing procedure that requires revisions and adaptations according to the characteristics of the participants. The Homophobia Scale of seven items (EHF-7) has showed acceptable performance in medical students attending to two universities in Colombia. However, performance of some items was poor and could be removed, with an improvement in the psychometric findings of items retained.
To review the psychometric functioning and refine the content of EHF-7 among medical students from two Colombian universities.
A group of 667 students from the first to tenth semester participated in the research. Theirs ages were between 18 and 34 (mean, 20.9±2.7) years-old, and 60.6% were females. Cronbach alpha (α) and omega of McDonald (Ω) were calculated as indicators of reliability and to refine the scale, an exploratory (EFA) and confirmatory factor analysis (CFA) was performed.
EHF-7 showed α=.793 and Ω=.796 and a main factor that explained 45.2% of the total variance. EFA and CFA suggested the suppression of three items. The four-item version (EHF-4) reached an α=.770 and Ω=.775, with a single factor that accounted for 59.7% of the total variance. CFA showed better indexes (χ=3.622; df=1; P=.057; Root-mean-square error of approximation (RMSEA)=.063, 90% CI, .000-.130; Comparative Fit Indices (CFI)=.998; Tucker-Lewis Index (TLI)=.991).
EHF-4 shows high internal consistency and a single dimension that explains more than 50% of the total variance. Further studies are needed to confirm these observations, that can be taken as preliminary.
评估测量量表的过程是一个持续的程序，需要根据参与者的特征进行修订和调整。七项恐同量表（EHF - 7）在哥伦比亚两所大学的医学生中表现出了可接受的性能。然而，一些项目的表现不佳，可以删除，保留项目的心理测量结果会有所改善。
回顾EHF - 7在哥伦比亚两所大学医学生中的心理测量功能并完善其内容。
一组来自第一至第十学期的667名学生参与了研究。他们的年龄在18至34岁之间（平均20.9±2.7岁），60.6%为女性。计算克朗巴哈α系数（α）和麦克唐纳欧米伽系数（Ω）作为可靠性指标，并为完善量表进行探索性因素分析（EFA）和验证性因素分析（CFA）。
EHF - 7的α系数为0.793，Ω系数为0.796，一个主要因素解释了总方差的45.2%。EFA和CFA建议删除三个项目。四项版本（EHF - 4）的α系数为0.770，Ω系数为0.775，一个单一因素占总方差的59.7%。CFA显示出更好的指标（χ = 3.622；df = 1；P = 0.057；近似均方根误差（RMSEA）= 0.063，90%置信区间，0.000 - 0.130；比较拟合指数（CFI）= 0.998；塔克 - 刘易斯指数（TLI）= 0.991）。
EHF - 4显示出高内部一致性和一个解释总方差超过50%的单一维度。需要进一步研究来证实这些观察结果，这些结果可作为初步结论。
