**Addendum Preregistration ‘Disadvantaged Youth in Higher Education: Health and Match with the Environment’** Because of valuable comments and suggestions from reviewers on previous versions of the manuscript, we have deviated from the preregistration in several important aspects. In the final version of the manuscript, the study was conducted in line with the preregistration with regard to research question 2, the dataset, inclusion criteria, and independent variables. Below, we describe the deviations. ****Focus on satisfaction**** Reviewers expressed their concerns about the measures for health. First, the health measures are self-reported by the students and therefore, the reliability of these measures is uncertain. Second, the dataset contains no information about the onset of the health problems (e.g., the health problems may have developed during childhood) which makes it difficult to examine the relation between undermatching and health. Because of these concerns regarding health measures, and with the aim of streamlining the manuscript, research question 1 (health) was removed from the study. Instead, we focus solely on research question 2 (satisfaction). Consequently, research question 1, as well as all the measures and analyses regarding health were removed. ****Changes in scales for satisfaction**** Because of reviewers’ concerns regarding the internal consistency of the scales and because of concerns regarding the fit with the theoretical underpinnings, we have reconsidered the scales for satisfaction. **Satisfaction with the social environment.** According to our preregistration, the scale for satisfaction with the social environment consisted of two items (1. General atmosphere, and 2. Students’ attitude towards fellow students) with a Cronbach’s alpha of 0.70. In order to elevate the internal consistency of this scale, and because of the fit with the theoretical underpinning (according to our theoretical model, teachers are also a part of the social environment in college), we decided to add one item (i.e., Teachers’ attitude towards fellow students) to our scale. The Cronbach’s alpha of our new scale is 0.73 (changed from “acceptable” to “respectable”). **Satisfaction with the academic environment.** According to our preregistration, the scale for satisfaction with the academic environment consisted of three items: 1. The content of the program, 2. The general skills learned in the program, and 3. Quality of teaching. We reconsidered this scale because of the fit with our theoretical underpinning. In the theoretical model that we build on, teachers can also be considered as part of the social environment rather than the academic environment. Therefore, satisfaction about the quality of teachers may bias our scale for satisfaction with the academic environment. Moreover, the literature about the satisfaction among undermatched students is especially focused on the academic environment that may not be enough challenging in less selective institutions. Therefore, we decided to remove the item about quality of teaching and to replace this with an item about satisfaction with the degree to which the academic environment is challenging. These changes resulted in a scale with a Cronbach’s alpha of 0.72 (a respectable level of internal consistency). The items in the final scales were included in the questionnaire in the most recent years 2013, 2014, and 2015 (N = 21,452). ****Propensity Score Matching and Split Sample Method**** Because of comments and suggestions from reviewers related to the possibility that pre-existing differences between students who match and undermatch may confound our findings, we decided to apply Propensity Score Matching (PSM) in our main analyses. PSM addresses issues regarding confounders more rigorously than linear regression models. When applying PSM, it is recommended to take into account all observed factors that may hold pre-existing differences between the treatment and control group. In our preregistration, we mentioned the covariates age, gender and grades in high-school. However, the dataset contains more observed factors that may hold pre-existing differences between matched and undermatched students. Therefore, we applied the preregistered covariates age, gender and grades in high-school, as well as immigrant status, home language, disability or chronicle disease, and motivation before entrance to higher education as predictors for undermatching. Consequently, following the recommended PSM procedure, in the main analyses testing the interaction between undermatch and SES, the pre-registered covariates were not included (because these were applied in the first stage of PSM, creating the treatment and control group). In addition, following reviewers’ suggestions, since experiences with college major may confound students’ experiences after being matched or undermatched, we decided to add ‘college major’ as covariate in the main analyses testing the interaction between undermatch and SES. The choice to apply PSM reduced our sample from 21,452 respondents to 8,356 respondents (from whom 2,628 undermatched students). We decided not to apply the Split Sample Method on this smaller sample because of the loss of power. Instead, we conducted our analyses twice: on the dataset resulting from the matching procedure (results-section), and on the full sample (sensitivity analyses-section). The analyses that were preregistered for research question 2a and 2b were reported in the sensitivity analyses-section.