Martin Hicks and Fiona Deller – An Homage to Good Data
Those of us focused on facts rather than anecdotes have known all along that the alleged collapse of employment opportunities for higher education graduates is untrue. At HEQCO, we have published on this. And now comes a major breakthrough, a significant step forward, in our data-driven understanding of outcomes for Canadian college and university graduates. We are referring to Barista or Better, the newest release from Ross Finnie’s research shop at the University of Ottawa.
Canada has surveyed graduates in the past to ask about their jobs and incomes after graduation. Over many years the results have consistently pointed to their overall success. But as an approach to generating information about income, surveys are second best. The longer one waits to go out in the field, the lower the response rate. The less time one waits, the more myopically short term the findings. And regardless of when one conducts the survey, how representative are those who choose to participate and how accurately do they recall their past incomes?
Finnie, his colleagues and partners (disclosure — we were one), federal officials at Employment and Social Development Canada and Statistics Canada, and 14 willing colleges and universities did it a better way. They partnered up, marrying institutional graduate files (who graduated and from what programs) with tax data (what they earned in their first eight years out). No need to find graduates, contact them and pester them with questions. The results: rich, accurate and complete data that verify that the job outcomes for graduates are very good and that reveal the patterns between cohorts, programs of study and years on the job.
But the best part of the story is not the results. The best part is that this happened at all. It happened, in Canada, even though there are entrenched barriers to combining existing datasets in innovative ways that make a difference. We see too much of a culture of protective data ownership and anxiety about data imperfections. We see a pervasive misinterpretation of privacy legislation that leads us to question whether this issue is really being raised to avoid disclosure. We see a lack of imagination about the uses of data to make better decisions, to generate better students outcomes and to spend public dollars more wisely. And we see a lack of leadership ready to overturn the protected status quo. The actual data work, complex as it may look, is easy peasy in comparison to the human barriers. Congratulations to the partners on this study for vaulting over all of that.
Our interest ought now to shift from whether graduates are, on average, doing well (answer is yes), to what is happening along the distribution curve of individual outcomes. Who and how many are near the bottom, doing not so well? Where did they come from? Can they be identified early on or are they a random phenomenon? What of those who don’t show up at all in a study of graduates because they never got to the front door of higher education in the first place?
We have the data to investigate these and many more questions. In Ontario, we have collected data on students all the way back to childhood and all the way up to graduation from postsecondary; data rich with background information about the students and their academic performance. If we married this with Finnie’s files, we could generate instantaneous longitudinal studies stretching both backwards and forwards in time, and glean insights into those who never got there.
The connecting thread is called the Ontario Education Number. It is the key that stitches these data together. Every student in the province has such a number, but at this point we make very little use of the information it gives us because of all the reasons outlined above. Other jurisdictions across North America have similar education numbers and make wonderful use of the data. They use it to figure out who is going to PSE and who isn’t; who is succeeding and who needs help; how students are moving between institutions and programs and into the labour market, and where they are dropping out of the system; and how programs should be designed to help students and which are succeeding in their efforts.
You know — evidence-based policy making. With a little bit of resolve and focus, we could do the same.
Martin Hicks is HEQCO’s executive director, data & statistics; Fiona Deller is HEQCO’s senior executive director, research & policy.