4. Returning to school

Archived Content

Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please "contact us" to request a format other than those available.

4.1 Characteristics of dropouts and returnees

Young men and young women have very different incidences of dropping out and of returning to school. In general, profiles of female and male dropouts differ on a few significant dimensions. Table 4.1 presents the characteristics by gender for all dropouts aged 18 to 20 in December 1999. The characteristics are those that may influence the decision to return to school.

Box 4.1

Data and Definitions

The data for this analysis come from the first two cycles of the Youth in Transition Survey, cohort B. The first cycle of this survey took place in 2000, collecting information from respondents who were aged between 18 and 20 years in December 1999. The second cycle took place in 2002 when the same youth were aged between 20 and 22 years.

For the purpose of this analysis, a dropout is any respondent who was not attending school in December 1999 and who had not completed high school.

A returnee or a return-to-school consists of a dropout who attended high school or a postsecondary institution, between January 2000 and December 2002.

All of the personal and family characteristics of returnees and non-returnees used for the analysis are those they had as of December 1999, i.e. previous to their return to school.

For more details, see Appendix B.

The characteristics presented in Table 4.1 are those that are believed to influence the decision to return to school. Two characteristics are used to capture potential budget constraints: living alone and having a child. Having a child (or children) captures the financial responsibilities the individual faces. How the individual values school is captured by parental education. As noted above, parents who have postsecondary credentials tend to value education highly and to impress upon their children the importance of education.

To capture the costs of returning to school, the last grade completed, the time elapsed since leaving school, and having repeated a grade in primary school are used. The more credits required to complete, the longer it will take the returnee to complete his or her schooling. The time elapsed since leaving school captures the costs associated with lost studying habits, and for those working, the opportunity cost of losing the acquired seniority in the current employment by going back to school. Having repeated a grade can be thought of as a proxy for having lower ability and potentially lower self-confidence for academic activities, increasing the costs of returning to school. Finally, the analysis also controls for the reasons dropouts gave for leaving school.

Table 4.1
Descriptive statistics, all high school dropouts¹ and dropouts who returned to school between January 2001 and December 2002

Three variables are used to capture the opportunity cost of returning to school. The first two capture work activities in the fall of 1999. A dummy variable indicates whether the individual worked at all during the period and if so, how many hours they worked. The last variable, the unemployment rate among those 15 years old or more in the economic region¹⁴ of residence captures the potential supply of jobs.

An intentionally-temporary dropout is captured through the postsecondary education aspirations. The underlying rationale is that an individual wishing to pursue postsecondary education probably did not envisage dropping out. Postsecondary aspirations are measured through two dummy variables: having taken as the last course of mathematics a postsecondary preparatory one (thereafter termed prepostsecondary mathematics) and having educational aspirations greater than a high school diploma. Having taken a pre-postsecondary mathematics course can be interpreted as past postsecondary aspirations and thus as intentions to complete high school. Pre-postsecondary mathematics courses are more demanding than those designed for students intending to work upon graduation. An individual contemplating to leave school without graduating would likely not choose to take a more-demanding mathematics course.

As can be seen in Table 4.1, columns 1 and 2, fewer women drop out than men and thus, they make up only 39% of the sample. They also return in a greater proportion (35% versus 26%). The circumstances surrounding dropouts are slightly different for men and women. Men drop out at an earlier stage of their secondary schooling than women. Almost half (49%) of male dropouts leave school prior to having completed grade 11, that is, in grade 10 or before compared to 43% of female dropouts. Male dropouts have been out of high school longer, almost 6 quarters compared to 5.2 for women.¹⁵

Another obvious gender bias lies with the proportions that declared having dropped out due to personal reasons versus want/need to work. The proportion citing leaving school due to personal reasons is four times higher among young female dropouts than among young men. Personal reasons encompass expecting or caring for a child, health problems and having problems at home. In contrast, the proportion of young men who cited wanting/needing to work as the reason for dropping out is double the proportion of young women. Young male dropouts appear more eager to join the labour force than their female counterparts. Alternatively, the expressed need or desire to work on the part of some young men may be driven by or linked to having to provide for a child. It is worth noting that male dropouts engage more in work and work more hours. These work activities do not appear to be driven by local labour market conditions. Indeed, the average local unemployment rate for men was 0.3 percentage points higher than the one for women.

In terms of educational experiences and aspirations, no clear pattern emerges between male and female characteristics. Young men in the sample experienced repeating a grade in primary school in greater proportion. Their postsecondary aspirations are slightly lower than those of female dropouts. Yet male dropouts more frequently choose to take pre-postsecondary mathematics courses, while a greater number of female dropouts have not pursued mathematics beyond grade 9.

A few more differences emerge in the personal and family characteristics. Female dropouts are more likely to have children and to live alone. A greater proportion of male dropouts are unaware of their parent's education levels.¹⁶ As noted earlier, male and female dropouts face relatively similar labour market conditions as indicated by the unemployment rates. Finally, both male and female dropouts come disproportionately from Quebec and Alberta.

Turning to the characteristics of returnees (columns 3 and 4 of Table 4.1), men and women returning to school tend to differentiate themselves from other dropouts along the same characteristics. Focusing on the variables of interest, that is, the PSE-aspiration variables capturing return intentions, returnees are more likely to have taken a pre-postsecondary mathematics course and to want a postsecondary certification. There is no clear pattern between male and female returnees for these variables. Thus, if the PSE-aspiration variables are to explain the gender gap in return rates, the difference most likely will emerge in the influence of these factors – the past and current postsecondary aspirations – since similar proportions of male and female dropouts have these traits. The next section presents the regression results for the determinants of return controlling for all the characteristics at once for men and women separately and discusses the influence of return intentions.

4.2 What influences young dropouts to return to school?

The decision to return to school is analyzed for young men and women separately using the estimation technique of linear probability modelling (see Box 4.2 for further details). Two models are estimated: one basic model and a second, the full model which includes the educational aspirations and reasons for having dropped out. Table 4.2 presents the results of both models.

Box 4.2

Estimation methodology

The linear probability model (LPM) is an estimation technique that produces results that are very easy to interpret. Results indicate the number of percentage points a given characteristic increases or decreases the probability of return. For example, a coefficient of -0.10 for having a child implies that this characteristic decreases the probability of return by 10 percentage points from the average. This ease of result interpretation justifies the use of the LPM technique despite the possibility of predicting negative and greater-than-one probabilities. The proportion of probabilities falling outside the zeroone range is calculated and reported. For validation of the results, the model is also estimated using the probit model.

Two specifications are estimated: one excluding the current postsecondary aspirations and the reasons for dropping out (the basic specification) and the other which includes these two variables (the full specification). These two variables may potentially introduce endogeneity problems. Current postsecondary aspirations may have changed between the dropping out of school and the time of the survey. Individual aspirations may have been altered by the same factors as those that influence the decision to return to school. Similarly, the reasons the individual gives at the time of the survey may have been altered by the individual's experience out of school and the decision to return or not. The estimation strategy takes into account this potential problem by estimating the two specifications and testing for endogeneity.

LPM results are presented here; results for the alternative estimation technique, the probit analysis, are found in Appendix Table A.3.1. The p-values are found below the coefficients. The proportion of predictions outside the 0-1 interval is indicated at the bottom of the table. Overall, the LPM produces no more than 5% of the predictions outside the 0-1 interval and all the erroneous predictions are negative return probabilities.

The results from the basic regression model reveals that the determinants of school returns are perfectly gender-specific, i.e. no one factor significantly affects both the male and the female decisions, including the proxy for intentionallytemporary departures and having taken a pre-postsecondary mathematics course (found under the category "Educational experience and aspirations"). Male dropouts who took a pre-postsecondary math course are 15 percentage points, or 60%, more likely to return to school than those who did not take such a course. This suggests that a non-negligible proportion of male dropouts had the intention to pursue some form of postsecondary education and thus most likely left high school before graduation with the intention of returning. Women, on the other hand, are not influenced by past school preparation. If past pre-postsecondary mathematics courses influenced women the same way as they do men, the gender gap in returns would have been greater than its current 10 percentage-point difference.

Besides past academic preparation, men are also influenced by a few factors while women by only two – how long they have not been in school and being unaware of their parent's education level. Indeed for every quarter that young women stay out of school, their likelihood of returning is 6.7 percentage points lower than if they had returned the quarter before. As important, ignorance of parental education levels is associated with a lower probability of return. While ignoring the parental education probably does not influence directly the decision to return, it probably is indicative of the emphasis put on education in the home environment. No other determinant appears to influence women's decision to return to school, including having a child or the number of hours worked in a month.

In contrast, various factors influence men's decision to return. The most important factor is having dropped out in grade 12 rather than in an earlier grade. Dropouts with some grade 12 courses are only a few credits/courses away from graduation, implying lower opportunity costs of obtaining their high school diploma or of pursuing college studies. They are twice as likely to return as young men who dropped out in grade 11 or earlier.

Other factors influencing men include having a child and/or having a parent with a postsecondary diploma or degree. Having parental responsibilities negatively affects the decision of returning to school for young men, diminishing the return probability by 10 percentage points. The need to earn a living probably explains the influence of having a child on the decision. Interestingly, parental obligations do not influence young women's decision to return to school.

Unlike having parental responsibilities, working does not constitute a very influential deterrent for men. Working an additional 40 hours in a month, the equivalent of a week, lowers the probability of return by only 4 percentage points. On the other hand, those who do not work at all are also less likely to return to school. This result suggests that the reasons leading young men not to work may also stand as a barrier to return to school.¹⁷

Finally, parental education exerts as strong an influence as grade repetition or having a child, such that it could nullify either. Indeed, a parent with postsecondary certification encourages returns, increasing the return probability by 18 percentage points or 70%.

Table 4.2
Linear probability model results for the decision to return to school

Box 4.3

Alternative model specifications

Other regression models were estimated, including one without the number of quarters since out of high school. This variable, along with the last grade enrolled in, characterize the timing of dropping out and could create problems of multicollinearity. No evidence of the latter was found. Anther specification included a proxy for ability, the gradepoint average (GPA). This categorical variable did not exhibit any explanatory power, with p-values ranging from 40% to 95%. A few things may explain this lack of explanatory power. First, the GPA is self-reported and hence, individuals may not report it accurately, especially if embarrassed by it. Second, the GPA reported is for the last grade completed. The individuals left school at different points of their schooling (anywhere between grade 9 and grade 12). Obtaining a B in grade 9 versus in grade 11 probably does not mean the same thing about the individual's abilities. Having stayed in school until grade 11 probably indicates better abilities. Thus, the last grade completed was retained in the final specification, while the GPA was dropped.

The full specification, in the last two columns of Table 4.2, introduces current postsecondary intentions (labelled "wants to get PSE certification" under "Educational experience and aspirations") and the reasons that led dropouts to leave school. The inclusion of current postsecondary aspirations and reasons for dropping out does not seem to alter significantly any of the results. Hausman tests cannot reject the hypothesis that the inclusion of these variables creates problems of endogeneity. The p-values are 0.998 and 0.247 for men and women, respectively. In other words, the full specification provides valid results.

The results point to a large influence of wanting to have a postsecondary education for both men and women. The influence of postsecondary aspirations is as large as that found for parental education for young men. It can be further noted that this factor appears to be more important for female dropouts than for males. The difference, however, is not statistically significant.¹⁸ Assuming that postsecondary aspirations have not or have only minimally changed since the time of dropping out, this result offers some evidence that those who deemed their school departure to be temporary are indeed more likely to return.

The reasons that drove young men to drop out do not affect their decision to return. In contrast, the reasons for dropping out distinguish which young women return. Those who left due to personal circumstances – including expecting or having a child – as opposed leaving to go work are more likely to return. They return more often than those who left to go to work by 15 percentage points. Those who quit school due to academic reasons are also more likely to return by 10 percentage points – the result is significant at 11% – than those who quit due to work-related reasons. This result appears to indicate that a good proportion of females who left for academic reasons re-consider the benefits and costs of schooling once they leave and assess the net benefits of schooling as being positive. Possibly, those who left school for work may reconsider the net benefits, but still evaluate them as nonpositive or face financial constraints prohibiting them from returning to school.

The other determinants of the decision to return to school remain mostly unaltered for both men and women. For young men, the determinants that were significant at the 10% level in the basic specification lose some explanatory power (having a child and not working) but overall, the results are left unaffected by the introduction of the current postsecondary aspirations and the reasons for having dropped out.

It is interesting to note that the state of the labour market does not influence the return decision of either men or women. Various measures of labour market conditions have been tested: the unemployment rate of 15 to 24 year-olds, the industrial composition of the labour market, and the demand for skills as proxied by the percentage of workers with a postsecondary diploma or degree. The results remain the same. This lack of influence on the return to school contrasts with findings for the causes of dropping out. Parent (2006) explains the work decisions of high school students using the unemployment rate and shows that the better the labour market conditions are, the higher the hours worked by high school students leading to higher dropping out rates. However, Ferrer and Lauzon (2005) present evidence that younger cohorts have become less sensitive to labour market conditions in their decision to drop out. The lack of influence of this variable in the present results might be caused by an insufficient variation in the measure of labour market conditions. Further research with additional years of data would be necessary to have sufficient variation in the data to conclusively establish the role of labour market conditions in returning to school.

In closing, returns to school are determined by different factors for young men and women. Yet they share having significant postsecondary aspirations. Their postsecondary aspirations and the past academic preparation for men increase the probability of returning to school, suggesting that dropping out was viewed by the individual as a temporary absence from school rather than as a permanent departure. All of the above suggests that dropping out is, for a significant number of individuals, only a temporary interruption of schooling.

4.3 And were the returns successful?

Dropouts may return to school, but the question is: was the return successful in leading to a completed diploma or certificate? Table 4.3 presents the proportion of returns by school type and school status in December 2002.

Roughly two thirds of returnees chose to go back to high school while the remainder preferred to pursue a postsecondary diploma. Women chose in a higher proportion than men to pursue postsecondary studies (43% and 33%, respectively). In aggregate, some 26% and 32% of male and female returnees, respectively, successfully completed their studies and graduated. Almost 40% left again without completing their certification and some 37% of men and 33% of women were still attending school. The high proportion of failed returns suggests that individuals face further hurdles to completing than just returning to school.

Table 4.3
Outcome of the return to school, as of December 2002

Excluding those still attending school, the proportion of second-time dropouts account for more than half and about half of male and female returnees. Further research should pay attention to the reasons why not all returns to school are successful.