M. Martínez-Gómez1, J.A. Marín Garcia1, M. Giraldo O´Meara2

1Universidad Politécnica de Valencia (SPAIN)
2Universidad de Valencia (SPAIN)
The purpose of this article is to test the complete invariance between two different techniques of data collection to see if they can be regarded as equivalent. Therefore, we assumed a previously validated global model which analyses the existing relationship between the features of the teaching methodology used with university students, the motivation and the satisfaction. The total sample was constituted by 535 students of a Spanish public university in different qualifications, courses and academic years. 294 questionnaires were completed in paper form and 241 were completed digitally through the questionnaire developed with a free software programme called LimeSurvey. A Confirmatory Factor Analysis was carried out from the multigroup structural equation models perspective in order to analyse complete invariance, using the program EQS 6.1. Besides the Cronbach's alpha coefficient, several goodness-of-fit measures were assessed and a proper fit was achieved both in the web-based and paper-based questionnaires samples, so it can be accepted that the factor structure of the proposed model does not depend on the administration mode of the questionnaire, with equal factor loadings, equality of variances in the factors error terms and equality of covariances across factors.