Conclusion

This paper has presented an analysis of persistence in postsecondary education in Atlantic Canada based on the Postsecondary Student Information System (“PSIS”) database that has been developed at Statistics Canada using administrative files provided to them by PSE institutions in the region. Given the unique nature of this database, the analysis carried out is original in a number of important ways.

First, the PSIS data – and this analysis based on it – covers all PSE students in Atlantic Canada in public PSE institutions at the college and university levels, and therefore represents the overall record for all PSE students in the region. The analysis is generally broken down by level of study, with the focus on college and bachelor’s level students, while master’s, Ph.D.., and first professional degree students are also covered in terms of the basic persistence dynamics. Patterns are broken down by sex, age (as of the start of the program), province, and year of study (the latter throughout the analysis) and some interesting patterns are revealed.

Second, the longitudinal version of the PSIS that has been constructed for this analysis (the “L-PSIS”) allows us to conduct a full and proper analysis of persistence, which is an inherently dynamic (and complex) process. The approach used here consists of identifying individuals as they enter a new PSE program and then following them over time on a year by year basis. Individuals are then classified as either graduates, continuers, switchers or leavers on a year by year (and cumulative) basis. The full coverage of the PSIS is key to capturing all these dynamics, including separating switchers from true leavers from PSE, something which can not be done with institution-specific data alone. The full coverage of the Atlantic region also means we capture those who move to a program at an institution in a different province (as long as it is in Atlantic Canada) which is not possible with province-specific datasets, and to break the patterns down and make comparisons along the provincial dimension.

Third, the longitudinal aspect also allows us to identify not only “first transitions” (graduates, continuers, switchers and leavers) but also to see where “switchers” (defined as those who move to a new PSE program in a different institution) go in terms of their level (college versus university) and place of study (in the original province or a different one).

Fourth, again thanks to the longitudinal element and full coverage of the PSIS, we are also able to identify how many leavers return to PSE, and (again) where they do so with respect to level and place – although due to the relatively short time period currently spanned by the data (four years), we can only identify those who return within two years of leaving.

Fifth, we are able to identify PSE graduates and see how many continue with their studies at the PSE level either immediately, or after a break of one or two years. And once again we look at the level (college or university) and place (province) of those further studies. The limitation of the data to Atlantic Canada, however, means that we are able to capture (as with the other dynamics covered in the analysis) only those who continue (or return to) their studies in this same region, which obviously limits the analysis to some degree.

All this represents unique and original evidence on persistence in PSE which should be of interest not only to those concerned with student pathways in Atlantic Canada, but also elsewhere in the country, and even at the international level, at least partly because our findings might help place existing evidence in a new perspective. The analysis may be of interest to administrators, policy makers, academics, students themselves, and others.

We would suggest three principal directions for future work, although many more could be identified. First would be to carry out the additional data checks proposed in the paper (see Appendix 4 for details), including linkages with other longitudinal datasets (the YITS and LAD) which would i) help verify the quality of the PSIS (as well as the other datasets with respect to their tracking of PSE profiles), and ii) in some cases (particularly with the LAD) allow further research of a very interesting type based on the linked files. This could include incorporating the family background information available in the LAD into our analysis of persistence rates, and then following individuals in the LAD after they left PSE.

The second extension we suggest is to adopt a modelling-econometric approach to analysing the dynamics in question. This would include applying the multi-nomial logit model used in related work on the same questions carried out by the authors using the YITS database. In this way, more variables could be brought into the analysis, and the various factors that affect persistence could be looked at simultaneously. The model could include both individual characteristics (sex, age, and so on) and institution and program characteristics, perhaps going as far as to incorporate the remarkably detailed program- and course-level information available in the PSIS.

The third line of research would be to attempt to identify within-institution program changers and to add this dynamic to the analysis. This would likely require working closely with institutional representatives so that the information available in the PSIS is correctly interpreted in terms of identifying such changes. Given the differences across institutions in this respect, such an analysis might be best restricted to a limited number of institutions, at least to start.

Many other possible research projects could surely be identified with this rich and unique dataset, including not only those that continue in the persistence topic, but also others related to different PSE issues. We hope this paper has provided a useful first step which has provided interesting and useful new evidence on persistence in PSE in Atlantic Canada, has offered a helpful assessment of the strengths of the main strengths and limitations of the PSIS data, and has pointed the way to new work that could be undertaken.