This work aimed to explore the factorial structure and the measurement invariance between groups of the emotional dependence questionnaire (EDQ). Participants included a random sample of 569 university students, 67.8% women and 51.2% had a romantic partner. The multigroup confirmatory factor analysis (MG-CFA) showed metric invariance for the EDQ. Particularly, the findings indicate that five of six scales of the questionnaire have an average variance extracted higher than 53% and a composite reliability of more than 0.75; the borderline expression scale was excluded. It was found that the EDQ has a strong measurement invariance degree between genders, χ2(365) = 750.14, p<.05; RMSEA= .07; CFI= .92; TLI= .92; and the presence of a partner, χ2(365) = 750.30, p<.05; RMSEA= .06; CFI= .92; TLI= .92.

El objetivo de este estudio fue evaluar la estructura factorial y la invarianza entre grupos del Cuestionario de Dependencia Emocional (CDE) mediante una muestra aleatoria de 569 universitarios, 67.8% mujeres y 51.2% con pareja. El Análisis Factorial Confirmatorio Multigrupo (AFCMG) mostró invarianza métrica para el CDE. Específicamente, los resultados indicaron que cinco de las seis escalas del cuestionario poseen varianzas extraídas promedio superiores al 53% y una fiabilidad compuesta superior a .75. La escala de Expresión límite fue excluida porque no cumplió con el criterio de validez convergente. Se encontró que el CDE posee invarianza métrica fuerte entre sexos, χ2(365) = 750.14, p<.05; RMSEA= .07; CFI= 0.92; TLI= .92; y presencia o no de pareja, χ2(365) = 750.30, p<.05; RMSEA= .06; CFI= .92; TLI = .92.

This work aimed to explore the factorial structure and the measurement invariance between groups of the emotional dependence questionnaire (EDQ). Participants included a random sample of 569 university students, 67.8% women and 51.2% had a romantic partner. The multigroup confirmatory factor analysis (MG-CFA) showed metric invariance for the EDQ. Particularly, the findings indicate that five of six scales of the questionnaire have an average variance extracted higher than 53% and a composite reliability of more than 0.75; the borderline expression scale was excluded. It was found that the EDQ has a strong measurement invariance degree between genders, ?2(365) = 750.14, p<.05; RMSEA= .07; CFI= .92; TLI= .92; and the presence of a partner, ?2(365) = 750.30, p<.05; RMSEA= .06; CFI= .92; TLI= .92.