PurposeThis study aims at developing an instrument to measure the latent trait propensity to drop out in face-to-face higher education.Design/methodology/approachBased on Tinto's student integration theory, a 27-item scale was created to measure student propensity to drop out of undergraduate programs. Item response theory was used to evaluate the psychometric analysis of the items. Furthermore, different methodologies were used to evaluate and provide evidence for content validity, response process validity, internal structure validity and criterion-related validity.FindingsWith the support of specialists in the construct, the interpretation of the scores for the use of the scale was defined in four levels of propensity: high, moderate, low and very low.Research limitations/implicationsThe latent trait propensity to drop out in face-to-face higher education allows the inclusion of new items and aspects in the instrument. Thus, it can be adapted to distance education.Practical implicationsThe students' propensity to drop out score can be useful for researchers and administration units in colleges and universities in the planning of permanent institutional actions and programs to take preventive measures.Social implicationsMinimize dropout in order to raise the educational level of the population and make better use of the resources invested in education.Originality/valueThis study points out when, why and how propensity to drop out can be measured and how scores can be interpreted in the context of the problem.