Relationship Satisfaction in Young Couples: Evidence for Validity of Short Scale Combining CFA and IRT

This study aimed to validate a short relationship satisfaction scale (RAS) in a sample of young people and adults in the Lima metropolitan area. There were a total of 806 participants aged between 18 and 30 years: 622 females (77.20%) and 184 males (22.80%). Every participant was currently in a love relationship of a minimum of three months. The methodologies of Confirmatory Factor Analysis (CFA) and Item Response Theory (IRT) were combined to determine the structure and assess the reliability of the instrument. The Omega coefficient (omega) was used from CFA, while the test information function and empirical reliability (r(xx)) were used from IRT. The results showed that the RAS could be interpreted as a one-dimensional scale when eliminating the items 4-7 and joining the items 1-2. This model shows high goodness of fit from a CFA-perspective. A similar situation occurs in IRT, except that error covariance is not considered. Nevertheless, assuming the elimination of items 4-7 is still a good option. The reliability in both CFA (omega = .838) and IRT (r(xx) = .862) was satisfactory. The evidence of relationship with other variable showed inverse and big relationship with violent spells of tension and difficulty.