[Internal validity and reliability of an instrument for evaluating quality of nursing care services for patients, nurses, and family members].

Aim: To analyze the internal validity and reliability of the instrument for evaluating the quality of services adapted to three interest groups: patients, nurses, family members, and primary caregivers.

Material and method: Our research was conducted meticulously, employing a mixed methodology with two phases: qualitative, using the focus group for internal validation of the instrument, and quantitative. Subsequently, the survey was passed to the interest group of 430 patients, 525 relatives, and 298 nurses. Chronbac's alpha reliability analysis, the multiple linear regression model as a point estimator of the parameters, and exploratory factor analysis with a maximum likelihood factor using Kaiser-Meyer-Olkin to analyze the constructor and its indicators were performed to validate the adaptation. Finally, confirmatory factor analysis determines their respective measurement models' unidimensionality, validity, and reliability.

Results: The result shows that the factor loading of each subconstruct is more significant than 0.5 in the three models, which indicates that the aptitude indices of the model were met. In addition, the model meets the discriminant validity criteria. The behavior of the SERVPERF questionnaire was analyzed in terms of consistency, Cronbach's alpha=0.94.

Conclusion: The scale items' discrimination concerning the questions on global satisfaction is confirmed. This shows that the instrument is valid, reliable, and useful.