Differences and similarities between 130 volunteers who remain for more than eight years in the same non-profit organization and 110 volunteers who quit during the first year were analyzed in this paper. Both groups were chosen from a sample of 851 volunteers that were working as volunteers when we assessed the independent variables (Time 1). After a 12-month follow-up (Time 2), 209 (25%) of them had dropped out and 642 (75%) continued in the same organization. Using the previous time, we formed two groups made up of those who dropped out and had been in the organization less than a year and those who continued and had been in the organization more than 8 years. Results show that differences and similarities between both groups are coherent with the three-stage model of volunteer's duration (Chacón, Vecina, & Dávila, 2007). This model includes the functional approach of volunteers' motivations (Clary & Snyder, 1991), and the role identity approach (Callero, 1985), and indicates that people will remain as volunteers insofar as this satisfies the motivations that are relevant for them at the first stage, they develop organizational commitment at the second stage, and they develop role identity as volunteers at the third stage. More specifically, results show that it is possible to predict 85% of the cases correctly using seven variables. Volunteers who remain after eight years feel a higher level of emotional exhaustion, a higher level of organizational commitment, and a strong role identity as volunteers. They are also highly satisfied with the friendships in the organization and have a stronger intention to remain at the long term (2 years).

Risk propensity is the stable tendency to choose options with a lower probability of success, but greater rewards. Its evaluation has been approached from various perspectives: from self-report questionnaires to objective tests. Self-report questionnaires have often been criticized due to interference from voluntary and involuntary biases, in addition to their lack of predictive value. Objective tests, on the other hand, require resources that make them difficult to administer to large samples. This paper presents an easy-to-administer, 30-item risk propensity test. Each item is itself an objective test describing a hypothetical situation in which the subject must choose between three options, each with a different gain function but equivalent in expected value. To assess its psychometric fit, the questionnaire was administered to 222 subjects, and we performed a test of its reliability as well as exploratory factor analysis. The results supported a three-factor model of risk (Sports and Gambling, Long-term Plans, and Loss Management). After making the necessary adjustments and incorporating a global factor of risk propensity, confirmatory factor analysis was done, revealing that the data exhibited adequate goodness of fit.