• Aucun résultat trouvé

Research question three: How does the association of skill-based activities at home, readiness to learn in 2012, and the likelihood of being NEET in 2014 differ between youth initially identified as NEET in

2012 and non-NEET youth?

As discussed above, it is difficult to unravel the relationship between skill-based activities at home and NEET status with reference to a single time point, given that they occur simultaneously. Importantly, research question two is unable to determine if NEET youth engage less with skill-based activities at home because they are NEET or if having fewer skill-based activities at home is associated with an increased likelihood of being NEET. To further understand the connection between skill-based activities at home and NEET status, Figures 1 to 4 reveal how the predicted probability of NEET status in 2014 (i.e., AME generated from an underlying logistic regression model) differs by skill-based activity and readiness to learn index scores for a LISA sub-set of youth identified as NEET or non-NEET in 2012.

The results in Figure 1 reveal the importance of home-based reading activities, especially among youth who were NEET in 2012. It shows that youth who either were or were not NEET in 2012 had a lower probability of being NEET in 2014 if they had higher scores on the reading-based activities at home scale in 2012. Among youth who were NEET in 2012, those who engaged in fewer reading-based activities at home were more likely to be NEET in 2014. Notably, the probability of NEET status in 2014 begins to converge between the two groups at higher scores, suggesting that increasing reading-based activities at home may decrease the likelihood of NEET status over time.

Figure 1 Probability of 2014 NEET status by reading-based activities at home in 2012

Note: All results are calculated from Canadian LISA microdata for respondents aged 16 to 24 in 2012. Conditional marginal effects are generated from a logistic regression that controls for age, gender, immigration status, disability status, parental education, number of books at home at age 16, and literacy assessment score, as measured in 2012. To interpret the average predicted probability of NEET status among all LISA participants, all covariates are mean centred.

The analysis of numeracy-based activities at home shows a less pronounced trend than reading activities do. As illustrated in Figure 2, the predicted probability of being NEET in 2014 decreases slightly by numeracy-based activities at home index scores among NEET and non-NEET youth in 2012. Nevertheless, overlap in the confidence intervals highlights greater uncertainty compared to reading at home index scores. That is, although the lines signal slight differences when comparing the probability of being NEET in 2014 at the low and high ends of the index, the large confidence intervals surrounding these point estimates suggest a lack of confidence in this relationship. We cannot establish if NEET youth who engaged in more numeracy-based activities at home were or were not less likely to be NEET in 2014.

Figure 2 Probability of 2014 NEET status by numeracy-based activities at home in 2012

Note: All results are calculated from Canadian LISA microdata for respondents aged 16 to 24 in 2012. Conditional marginal effects are generated from a logistic regression that controls for age, gender, immigration status, disability status, parental education, number of books at home at age 16, and numeracy assessment score, as measured in 2012. To interpret the average predicted probability of NEET status among all LISA participants, all covariates are mean centred.

The analysis of ICT-based activities at home shows a similar trend to reading activities. Figure 3 demonstrates that the predicted probability of being NEET in 2014 decreases by ICT-based activities at home index scores among youth who either were or were not NEET in 2012. Although there is some overlap in the confidence intervals, differences can be seen between the low and high ends of the index, especially for youth who were NEET in 2012. This suggests that NEET youth who engaged in more ICT-based activities at home in 2012 were also less likely to be NEET in 2014.

Figure 3 Probability of 2014 NEET status by ICT-based activities at home in 2012

Note: All results are calculated from Canadian LISA microdata for respondents aged 16 to 24 in 2012. Conditional marginal effects are generated from a logistic regression that controls for age, gender, immigration status, disability status, parental education, number of books at home at age 16, and PS-TRE assessment score, as measured in 2012. To interpret the average predicted probability of NEET status among all LISA participants, all covariates are mean centred.

Compared to skill-based activities at home, higher scores on the readiness to learn scale in 2012 are not associated with the probability of being NEET in 2014. As shown in Figure 4, both NEET and non-NEET youth in 2012 with high and low readiness to learn scores had a similar probability of being NEET in 2014. The absence of a connection between 2012 readiness to learn and 2014 NEET status may be due to the small distinction between the scores of NEET and non-NEET youth in 2012. That is, NEET youth had average readiness to learn index scores that were only marginally smaller than non-NEET youth in 2012. As the discussion highlights next, what NEET youth report doing at home is more strongly associated with the later probability of being NEET than is their self-reported readiness to learn.

Figure 4 Probability of 2014 NEET status by readiness to learn index score in 2012

Note: All results are calculated from Canadian LISA microdata for respondents aged 16 to 24 in 2012. Conditional marginal effects are generated from a logistic regression that controls for age, gender, immigration status, disability status, parental education, and number of books at home at age 16, as measured in 2012. To interpret the average predicted probability of NEET status among all LISA participants, all covariates are mean centred.