Educational Attainment and the Gender Gap in Childcare in Canada: A Decomposition Analysis

Abstract

Using the 2010 General Social Survey in Time Use (Canadian Time Diary data set, N = 1782), we explore the relationship between the education level of couples and the time they spend on childcare. We find that fathers and mothers with higher levels of educational attainment spend more time parenting children. However, the education childcare gradient is stronger for mothers than fathers. Consequently, the gender gap in childcare is much greater for couples with more educational attainment. The Blinder–Oaxaca decomposition results of these gender gaps by education level suggest very little can be attributed to how mothers and fathers at different levels of education differ on demographic and workplace characteristics. We argue that the differences in parenting time accompanying socio-economic status are more likely attributed to differences in parenting values.

JEL Classifications: I24, J13, J16, C10

Keywords

Educational attainment Gender gap Childcare Canada

1. Introduction

Despite the focus on social class in the social science literature, very little is known about how gender gaps in domestic work differ by levels of education. This understanding is especially important in an era where mothers continue to do on average the bulk of domestic work, irrespective of their market work commitments, although fathers are increasingly doing more domestic work. Amidst these persistent gender gaps, research suggests that parents are focusing more on their children and this focus is related to the socioeconomic characteristics of parents themselves (Nelson, 2010). How does the gender gap in parenting time differ by the education level of couples? For instance, to what extent does the time couples spend on market work lead to differences in the time they spend on childcare? The disentanglement of these different aspects of the gender gap in childcare time has implications for gender, work, family and attendant family welfare policies.

In this article, we investigate along socioeconomic and demographic profiles, particularly education, the differences in the time spent by Canadian fathers and mothers on childcare. In the first stage of the analysis, we examine the differences by gender and education level across all measures associated with and including time spent in childcare. We then regress childcare time on all variables by gender and education level. As a final aspect of this analysis, we decompose the gender gap for both education groups by implementing the Blinder–Oaxaca method. This technique allows us to separate the gender gap into the part that is attributable to differences on the predictors (the explainable component) and the part that is a result of differential influence of the predictors (the unexplained component) as well as unobserved factors.

The explained aspects of the gender gap in parenting is widely studied and understood. Traditionally, fathers are more likely to focus on their careers while mothers make career sacrifices more often to devote more time to parenting. This trend is explainable by fathers’ typically higher levels of work hours, on average, and mothers’ higher childcare hours. However, these explanations do not reveal the whole story of the gender gap in parenting time. By decomposing the part of the gap explained by mean differences from the part which contains the differential impacts of those differences on parenting time, we are able to fully understand the gendered nature of the gap. The research questions we aim to answer in this article are:

To what extent does the gender gap in parenting manifest itself differently for couples with higher levels of educational attainment?

To what extent is the gender gap in parenting across educational levels left unexplained by the differences in work and family characteristics?

We answer these two intriguing questions by applying the Blinder–Oaxaca decomposition technique to the 2010 Canadian Time Diary data set. To preview, our results indicate a positive relationship between educational attainment and parenting time. However, we find that the difference in the increase in parental time by educational attainment is less for fathers than mothers; fathers have a weaker education childcare gradient. This means that the gap in the time couples spend on childcare is largest among those with the most educational attainment. This article is the first of its kind filling a gap in the literature by studying the relationship between educational level and the gender gap in parenting time in Canada.

The second section is an overview of the literature on the major existing theoretical perspectives. In the third section, the methodological approach is presented. In this section, we explain the measures as well as the Blinder–Oaxaca decomposition approach. The results are then presented in the fourth section. In the fifth section of the article, the findings are interpreted and the theoretical implications are addressed.

2. Review of the Existing Literature

2.1 Education, Socioeconomic Status and Parenting

One of the most important components of socioeconomic status is educational attainment (Evans et al., 1997). Since the 1960s, the relationship between social class and parenting has been extensively studied. Kohn (1969) illustrated that working-class parents were more likely to value obedience to authority while middle-class parents fostered independent decision-making. This study sparked a sustained sociological debate linking work context to parenting and children’s outcomes in the social structure and personality theoretical approach (Kohn & Schooler, 1973).

More recently, Lareau (2011) has resurrected the connection between parenting and class in her book Unequal Childhoods. Lareau argues that middle-class parents practise parenting that she calls ‘concerted cultivation’; this involves nurturing children’s talents as well as helping them develop their voice. Working class and poor families, on the other hand, seek to achieve the accomplishment of natural growth in their children. Generally, these parents do not spend as much time scheduling organised activities but allow their children more free reign. Where middle-class parents may spend more time reasoning with their children, working-class parents are more likely to offer clear directives. Increased education of parents has been linked to increased parenting time comparatively (Dotti Sani & Treas, 2016). In addition, parents in more developed countries, being more educated, have been associated with valuing nurturing of independence over obedience in their children (Park & Lau, 2016).

2.2 Bargaining Within the Family

Game theoretic bargaining models have been developed to help elucidate the gendered divisions of the time allocated to household work, and by extension childcare, within the family (Lundberg & Pollak, 1996). These models arose out of a need to move away from common preference models of family behaviour. Common preference models are limited in that they impose a unitary household utility function wherein decision-making is largely devoid of conflict and is driven by a singular shared interest within the family (Becker, 1981; Samuelson, 1956). On the other hand, game theoretic bargaining models shed light on the allocation of household work within the family unit allowing for the relative power of family members and their possibly different and conflicting interests. These models also consider the possibility that family decisions taken at one point in time can have an impact on the relative bargaining power of spouses at future points in time (Basu, 2006).

The most widely used bargaining models use the notion of threat points, namely those external and those internal. External or divorce threat points centre on the notion that a spouse’s bargaining power comes from a determination of which spouse would be worse off if the union were to be dissolved (McElroy, 1990). The prediction is that each spouse would evaluate, in a utility-maximising way, the resources that he/she would have if the union were to be terminated.¹ This evaluation, which Bittman, England, Sayer, Folbre and Matheson (2003) note is done in the ‘shadow of divorce’, determines each spouse’s relative bargaining power over the division of labour within the household. In contrast, internal threat points focus on bargaining within the family unit without considering the possibility of the family unit itself being terminated because of a relationship breakdown arising from irreconcilable differences or otherwise (Lundberg & Pollak, 1996).Without the threat of divorce, power for each spouse comes from resources that each can withhold within the marriage to achieve a desired distribution of household work.

Within the context of our article, educational attainment proxies both for income and earning power. Consequently, the education level of mothers and fathers is one source of their bargaining power over which they use to influence who does what, when and how much within the household. While we attempt to homogenise our educational groupings in our analysis, there is still a role for bargaining power to affect childcare allocation decisions within these groups. This is driven in part by a persistent gender wage gap in favour of men in Canada (Boudarbat & Connolly, 2013; Drolet, 2001). Consequently, for men and women of equal educational attainment, even if they are at the same occupational level, our expectation is that women on average will have less bargaining power over how much time each spouse allocates to childcare. In addition, even among those women who have more bargaining power, there is the persistence of certain gendered expectations. These expectations mean that among some women who earn more than men, and by extension could be more highly educated, these women respond to this gender wage gap ‘aberration’ by actually not fully exerting their bargaining power over the time allocation decisions of household work within the family (Bittman et al., 2003).

2.3 Parenting in Canadian Context

Over the last few decades, Canadian fathers have been spending a significant amount of time in childcare. Concomitant to this change, the proportion of all stay-at-home fathers has also increased from a modest 1 per cent in 1976 to 13 per cent in 2011 (Hoffman, 2012). Such an increase in the number of stay-at-home fathers is partly attributed to the rise in mothers’ participation in the labour market. Parental leave benefits are also important determinants of increased paternal time with children (Boll, Leppin, & Reich, 2014). Relatedly, studies have also found that low levels of state-supported family provisions are associated with high income inequality in Canada (Gauthier, Smeeding, & Furstenberg, 2004; Pronovost, 2007; Sayer et al., 2004; Turcotte, 2007). In Canada, on average, among employed parents, fathers devoted about three-quarters of the time to childcare than mothers in 1998, while fathers only spent 60 per cent of the time that mothers spent on childcare in 1986 (Zuzanek, 2001).

Compared to most developed countries, Canadian fathers are in an advantageous position since they have more access to parental leave. Data suggest that there has been a dramatic rise in the percentage of eligible Canadian fathers taking parental leave in the last few decades (Statistics Canada, 2011). Canada is the second most socially progressive country behind Finland (Social Progress Index, 2016). Canadian society can be seen as a fairly egalitarian society as it ranks 30 among 145 countries in the global index of gender equality (World Economic Forum, 2015).

The aforementioned benefits among Canadians represent increased opportunities to parent. Given the developed nature of the Canadian economy, the impact of educational attainment on parenting practices is an intriguing question. Canada is considered a progressive country. At the same time, research suggests that while more time is spent on parenting in general, a gender gap in parenting remains (Buchanan, McFarlane, & Das, 2016, 2017). It is essential to have a better understanding of the role educational attainment, as a key component of socioeconomic status, plays in the time allocated to childcare by Canadian parents. Not only is education a key component of mobility but parents’ educational attainment is a crucial predictor of their children’s socioeconomic status. Therefore, our understanding of this issue not only has implications for inequality and mobility but also addresses the salience of gender and egalitarianism in Canadian society.

2.4 Contribution

Given the increased focus of middle-to-upper-class parents on parenting (Nelson, 2010), it is essential that we develop our understanding of how increased parenting time is related to the gendered division of labour in more-educated families. This article contributes to this area of research by decomposing both overall parenting time as well as the relationship of gender and education to it. We challenge the orthodox assumption that higher educational status produces greater egalitarian relationships among parenting couples. To our knowledge, there is no research that decomposes the dynamics of educational attainment and the gender gap in parenting time in the Canadian context. We contribute towards filling this gap in the existing literature.

3. Data, Measures, and Methods

3.1 Data

We use the 2010 General Social Survey on time use in Canada. This survey is based on time diaries, and it samples information from Canadians 15 years of age and older excluding residents of the Yukon, Northwest Territories, Nunavut and full-time residents of institutions. Respondents were asked to report the time spent in minutes on various activities over the 24-hour period starting at 4:00 am prior to the interview day. In the case of more than one activity, the respondents had to determine the main activity. The time spent on market work and childcare are the time use categories of interest. For the purpose of our estimation, we included the respondents, who are 20–64 years old, married or common-law, have a child 17 years old or younger, and at least one partner working full-time. A more detailed description of the survey is given in Buchanan et al. (2016).

The traditional or questionnaire approach relies on accurate recollection of the time spent on some activity in a reference period. As such, it is subject to recall error even if the period is recent. Further, respondents will often recall the period when the activity in question was most prominent. Therefore, the potential exists for the respondent to overestimate the actual amount of time spent performing the activity (Carlin & Flood, 1997). Time diaries, on the other hand, ask respondents to document in detail the time spent on various activities usually over a 24-hour period. Consequently, it forces the respondent to really think about when various activities occurred and how much time spent was spent on each in relation to each other. Therefore, time diaries, although more burdensome for the respondent, tend to provide more accurate information on the time allocated to various activities.

3.2 Measures and Methods

The time spent on childcare is the sum of basic care (e.g., baby care), recreational childcare (e.g., playing with children) and educational childcare (e.g., reading to children) when designated the main activity by the respondent. We utilise information on the work hours, work characteristics and educational attainment of spouses in the analysis of childcare by restricting the sample to only those who reported their marital status as married or common law. We group respondents into two distinct categories of educational attainment. First, we categorise both mothers and fathers who have completed a bachelor’s degree or more into one group. We compare this group to couples in which both mothers and fathers have not completed a bachelor’s degree. We also constructed several measures of individual and household income, but none were significantly related to childcare hours or the gender gap in childcare.

In the first stage of our analysis, we regress childcare hours on several measures separately by gender and education level. We control for age (20–29 compared to all other age groups), number of children (1, 2, 3, 4 or more), region where the respondent lives (Atlantic compared to all others), market work hours, whether or not the respondent regularly hired for childcare, respondent’s and spouse’s shift, access to flex time, whether or not the respondent works at home, and the respondent’s and spouse’s household labour hours.

In the second stage, we implement the Blinder–Oaxaca decomposition analysis to fully explore the gender gap in childcare by gender and education levels. The explained variation is the amount accounted for by the mean differences on our measures. The unexplained variation is attributable to the impact of those differences in addition to unobserved factors. We argue that, like similar models used to predict discrimination, within this unexplained portion is encapsulated the gendered aspects of time gap in childcare. In other words, the explained portion of the gap is attributable to mothers and fathers having different characteristics. The unexplained component of the decomposition is associated with the differential influence of the characteristics on childcare hours for mothers and fathers. Discrimination literature often suggests that the unexplained component is gender discrimination (Brown, Troutt, & Prentice, 2011). While we do not necessarily take a similar stance, we argue that the unexplained gap does consist of differential impacts. Therefore, this is the component containing the most theoretical intrigue.

4. Results

Table 1 displays the means and percentages by gender and education level for all variables. The results are not surprising for the demographic control measures. A smaller number of female parents in the more-educated group are younger parents. Also, a larger percentage of higher-educated mothers are from Ontario and fewer are from the prairies. This is not surprising, given that Ontario has more jobs requiring more education. For household labour, the more-educated group, both mothers and fathers, do fewer hours. In addition, a higher percentage of more-educated fathers and mothers work a regular shift, have a partner who works a regular shift, and have higher access to flexible working hours. More-educated fathers are significantly more likely to work at home for some of their market labour hours.

Table 1
Mean and percentage differences by education level on Demographics, Respondent Characteristics, and Spouse’s Characteristics (N = 1782)

Variable Male
Female

< Bach Bach or > < Bach Bach or >

Age of Respondent (20-29 = 1) (pct) 11.64 1.09 * 9.54 3.33

Age of Respondent (30-39 = 1) (pct) 44.18 51.63 43.68 47.04

Age of Respondent (40-49 = 1) (pct) 40.41 42.93 38.17 41.48

Age of Respondent (50-64 = 1) (pct) 3.77 4.35 8.60 8.15

Region (Atlantic =1) (pct) 21.23 20.65 21.10 18.89

Region (Quebec =1) (pct) 16.95 16.85 12.77 16.67

Region (Ontario =1) (pct) 29.62 41.30 28.63 33.33

Region (prairie =1) (pct) 23.12 13.04 ** 26.21 19.63 *

Region (BC =1) (pct) 9.08 8.15 11.29 11.48

Number of Children 1.65 1.72 1.69 1.71

Market Work Hours 30.86 34.05 41.24 42.06

Household Labour Hours 23.27 19.71 * 15.92 13.28 *

Regular Shift (pct) 78.94 86.96 * 72.72 85.93 *

Works at Home (pct) 20.89 26.09 20.97 32.22 *

Flextime (pct) 36.47 46.74 * 41.67 58.52 *

Spouse’s Childcare Hours 2.62 2.75 3.60 3.33

Spouse’s Household Labour Hours 2.56 2.64 3.82 3.56 *

Spouse’s Market Work Hours 36.15 38.70 22.73 24.32

Spouse’s Shift (regular =1) (pct) 65.92 83.15 * 49.06 64.07 ***

Regularly Hire for Childcare (pct) 29.62 32.07 24.46 31.85 *

N 584 184 744 270

Variable	Male	Female
Age of Respondent (20-29 = 1) (pct)	11.64	1.09	***	9.54	3.33
Age of Respondent (30-39 = 1) (pct)	44.18	51.63		43.68	47.04
Age of Respondent (40-49 = 1) (pct)	40.41	42.93		38.17	41.48
Age of Respondent (50-64 = 1) (pct)	3.77	4.35		8.60	8.15
Region (Atlantic =1) (pct)	21.23	20.65		21.10	18.89
Region (Quebec =1) (pct)	16.95	16.85		12.77	16.67
Region (Ontario =1) (pct)	29.62	41.30	**	28.63	33.33
Region (prairie =1) (pct)	23.12	13.04	**	26.21	19.63	*
Region (BC =1) (pct)	9.08	8.15		11.29	11.48
Number of Children	1.65	1.72		1.69	1.71
Market Work Hours	30.86	34.05		41.24	42.06
Household Labour Hours	23.27	19.71	*	15.92	13.28	*
Regular Shift (pct)	78.94	86.96	*	72.72	85.93	***
Works at Home (pct)	20.89	26.09		20.97	32.22	***
Flextime (pct)	36.47	46.74	*	41.67	58.52	***
Spouse’s Childcare Hours	2.62	2.75		3.60	3.33	**
Spouse’s Household Labour Hours	2.56	2.64		3.82	3.56	***
Spouse’s Market Work Hours	36.15	38.70		22.73	24.32
Spouse’s Shift (regular =1) (pct)	65.92	83.15	***	49.06	64.07	***
Regularly Hire for Childcare (pct)	29.62	32.07		24.46	31.85	*
N	584	184		744	270

Source: Authors’ calculation based on 2010 General Social Survey in Time Use.

Note: *p <= .05, **p<= .01, ***p<= .001

Next, we regress childcare hours on a number of controls separately for fathers and mothers by educational groupings. The results of this analysis are displayed in Table 2. As expected, childcare hours increase for fathers and mothers as the number of children increases, although for more-educated fathers the increases is not statistically significant at conventional levels. Additionally, the time spent on childcare as the number of children increases is greater for mothers in the more-educated group. Surprisingly, fathers and mothers in families in which someone is regularly hired to do childcare spend significantly more time in childcare. This could be related to households in which there is more sharing due to more time spent by both parents at work. For all groups, childcare hours are reduced as market work hours increase. The variation in childcare (about its mean) is reduced by 33 per cent in the model for mothers where both partners have bachelor’s degrees or higher. The other models reduce variation in childcare by less than 20 per cent.

Table 2

Ordinary Least Squares Regression of Hours Spent in Childcare Time on Demographics, Respondent Characteristics, and Spouse’s Characteristics (N = 1782)

Variable	Female		Male
	< Bach	Bach or >	< Bach	Bach or >
Age of Respondent (20-29 = 1)	5.436**	10.185	–1.189	–2.021
Region (Atlantic =1)	–0.740	2.250	1.687	–0.146
Number of Children	3.323***	6.730***	2.468***	0.226
Market Work Hours	–0.190***	–0.297***	–0.115***	–0.128***
Household Labour Hours	–0.167***	–0.514***	–0.118***	–0.062
Regular Shift	–1.440	–1.545	0.287	–3.650***
Works at Home	–3.629**	4.970	–1.125	1.856
Flextime	1.472	–0.148	1.673	0.423
Spouse’s Childcare Hours	0.367	–1.435	1.498***	–0.191
Spouse’s Household Labour Hours	0.408	–1.062	–1.376**	0.608
Spouse’s Market Work Hours	0.069*	–0.009	-0.024	0.029
Spouse’s Shift (regular =1)	–1.430	4.551	–1.350	–1.330
Regularly Hire for Childcare	1.178	3.535	4.141***	5.227**
R2	0.174	0.329	0.159	0.173
F	10.68	5.22	10.64	4.12
N	584	184	744	270

Source: Authors’ calculation based on 2010 General Social Survey in Time Use.

Note: *p <= .05, **p<= .01, ***p<= .001

In the final stage, we decompose the gender gap in childcare hours by education level. We subsequently analyse the gap in childcare time between mothers in couples, both with bachelor’s degrees or higher compared those in which both do not have a bachelor’s degree or higher. The Blinder–Oaxaca decomposition results are displayed for the general models for each educational grouping in Table 3.

Table 3

Blinder-Oaxaca Decomposition Results; Contribution of Statistically Significant Characteristics to the Explained and Unexplained Components of the Differential in Childcare by Education Level and Gender

	< Bach	Bach or >		Females Only
Overall Decomposition
Male Child Care Hours	8.726***	9.056***	Less than Bachelors	12.690***
Female Child Care Hours	12.690***	16.132***	Bachelors or higher	16.131***
Difference	3.963***	7.075***	Difference	3.442*
Explained	–0.107	–0.749	Explained	–0.369
Unexplained	3.856***	6.326***	Unexplained	3.073*
Explained Component	Coeff	Coeff		Coeff
Age of Respondent (20-29 = 1)	–0.049	0.049		0.578*
Number of Children	0.122	–0.078		–0.247
Market Work Hours	–1.327***	–1.957**		–0.142
Household Labour Hours	1.017***	1.724**		–0.611
Works at Home	–0.002	0.149		0.068
Spouse’s Childcare Hours	0.986**	–0.489		0.012
Spouse’s Shift (regular =1)	0.287*	–0.052		0.078
Regularly Hire for Childcare	–0.146*	–0.009		–0.044
Unexplained Component
Age of Respondent (20-29 = 1)	–0.697*	–0.228		–0.056
Number of Children	–1.440	–11.170**		–5.837
Market Work Hours	2.623*	5.900*		3.312
Household Labour Hours	1.002	8.097**		7.393**
Works at Home	0.524	–0.848		–2.120*
Spouse’s Childcare Hours	3.492*	3.889		5.014
Spouse’s Shift (regular =1)	–5.688*	4.873		3.898
Regularly Hire for Childcare	–2.698*	1.309		3.039
N	1328	454		768

Source: Authors’ calculation based on 2010 General Social Survey in Time Use.

Note: *p <= .05, **p<= .01, ***p<= .001

The overall decomposition results are presented at the top of the table. First, we explain the results for the gender gap by education-level models. For those respondents in couples in which both do not have a bachelor’s degree, the mean hours fathers spent in childcare is 8.7, and for mothers it is 12.7. Therefore, the gender gap in childcare hours after controlling for socioeconomic characteristics is approximately 4 hours. The decomposition results indicate that about 0.1 hours of this gap is explained by the father–mother mean differences on the characteristics in the model. The amount of the gap left unexplained by the mean differences on characteristics is 3.9 hours. This part of the gap is attributable to the differential influence of the characteristics on childcare time as well as the unobserved factors.

The last part of the table displays the breakdown of the explained and unexplained contributions to the gap by specific variables. Variables which are not significant in any of the models are not displayed. While some coefficients in the general model are significantly related to childcare hours, they do not contribute to explaining the gender gap. For example, market work significantly predicts childcare hours, but given that mothers and fathers do not differ on this component, the impact does not contribute to the explained portion of the gender gap. Overall, the small explained portion of the gap is not significant.

Next, the same process is performed with the sample for those in couples in which both partners have a bachelor’s degree or higher. Surprisingly, the gender gap in childcare is greater for this group. Furthermore, the explained portion of the gap is neither substantive nor statistically significant. The decomposition reveals that approximately 90 per cent, or 6.3 of the 7.1 hours, remains unexplained by the mean differences on characteristics between fathers and mothers. This suggests that differences in, for instance, market work hours between fathers and mothers account for less than the differential impact working has on the childcare time. More of the unexplained component measures are significant. Most importantly, the impact of increased numbers of children is more dramatic for mothers with higher levels of educational attainment. This is not surprising when considering the results from the ordinary least squares (OLS) regressions. The coefficient was much larger for this group of mothers compared to fathers. For the Blinder–Oaxaca results, this suggests that the impact of the number of children on childcare hours is much greater for mothers in couples in which both partners have a bachelor’s degree or higher. In other words, the return on children to childcare hours is greater for this group. These mothers devote more time to children than fathers.

The gap is greater in childcare between mothers of different educational configurations. While fathers of higher education levels do slightly more childcare, the difference is marginal. We point out that while fathers’ child care commitment may be increasing generally, mothers are consistently doing more. We find it interesting that more-educated mothers are not only experiencing a greater gap within their parental relationship but that they are doing more than less-educated mothers. Therefore, as a final step, we perform a Blinder–Oaxaca decomposition analysis on the childcare by education for mothers. The results suggest that there exists a weekly gap of 3.4 hours. This gap is not explained by differences in characteristics in the model between groups. In other words, the two groups of mothers have differential returns on their characteristics, such as the number of children. This is provoking given that more-educated fathers only do marginally more than other fathers. This suggests that more-educated mothers are doing the bulk of the increased parenting time.

This is surprising on two levels. First, for both more- and less-educated parents, the gender gap in childcare is left unexplained by the factors in the model which are typically related. Finally, the unexplained gap in childcare is larger for more-educated mothers. The Blinder–Oaxaca decomposition method is critical to our analysis because it confirms this gender gap is due to factors which are beyond the usual cluster of variables typically related to parenting time.

In addition to the Blinder–Oaxaca decomposition analysis, we also performed an OLS regression pooling both mothers and fathers and both education groups. Results from this analysis are located in Appendix A1. There are two models in this table. The first model is a baseline analysis of the predictors. For the second model, we include a gender-by-education-level interaction. The interaction model corroborates our earlier findings. The interaction between gender and education is statistically significant and indicates that mothers with higher levels of education spend more time parenting than do less-educated mothers.

5. Discussion and Concluding Remarks

Following the Blinder–Oaxaca example of a discrimination study, the explained aspects of the decomposition represent the role of differences on the predictor variables on impacting gender gap in childcare. The unexplained component includes the unobserved factors as well as the differential influences of those characteristics on childcare hours. We suggest that the unexplained gap can be interpreted partially as the difference in the parenting attitudes of the more-educated couples. While all of the unexplained portion of the gap cannot be attributed to differential gendered processes, we purport that the greater the portion of the gap that is significantly unexplained, the greater the existence of gendered processes and parenting attitudes unique to these different educational status levels. Following the same argument, it appears that education plays a strong role in the prediction of childcare for families.

Broadly speaking, there are two important pieces to these findings. First, there is the comparative amount of childcare time associated with different levels of education. Second, there is the comparative gender gap in childcare time by education. Both components have critical implications for our understanding of gender inequality across the different levels of socioeconomic status. Our surprising and main finding is that more-educated mothers have a more positive education childcare gradient than fathers. Consequently, the gender gap is larger for more-educated couples. If more-educated couples were contributing more to childcare and the gender gap was smaller, that would be less surprising. We offer several explanations of this educational impact on childcare and the childcare gap.

First, on average, the more-educated group does have slightly more flexibility in their jobs (Duncan & Pettigrew, 2012). This includes greater access to flexible work schedules and higher percentages working a regular shift. However, the differences do not seem to be explained by these characteristic differences in the Blinder–Oaxaca analyses.

Another way of looking at our findings is through the fathering lens. Given the aforementioned flexibility, why are fathers in the more-educated grouping not doing more childcare? Fathers who are more educated may potentially resist increasing childcare time. This may be because they feel they are doing enough and also could be a passive aggressive rebellion against egalitarian values. Because the expectation of the more educated is egalitarianism, educated fathers take advantage of this in how much time they allocate to childcare. An even more subconscious perspective on this notion is what Kulik and Tsoref (2010) describe about the onset of parenting. They argue that couples, even those that openly express egalitarian values, find themselves in very unequal domestic work configurations once they become parents. Either way, the increased childcare of more-educated mothers correlates with an increased gender gap.

Mothers may resist more participation and sharing by fathers in order to protect their maternal roles (Zuo & Tang, 2000). This has historically been applied to threats related to the mother’s role. The threat hypothesis is linked to gender role identity (Maume, 2011; Monna & Gauthier, 2008) which asserts that childcare time is determined not by egalitarianism but by gender norms which determine who should spend most time. According to the status characteristics theory, mothers are culturally socialised to succumb to a family devotion schema while males adopt a provider schema of family devotion (Ridgeway, 2011). This is further related to the concept of maternal gatekeeping. According to this perspective, it is the mothers that protect the traditional mother role.

Allen and Hawkins (1999) conceptualised maternal gatekeeping as a combination of reluctance to let go of responsibility coupled with rigid standard setting, a desire for ultimate accountability for domestic work and the expectation and confirmation that work involving family was critical to the identity of a mother. They found in their study that approximately 21 per cent of Western American mothers exhibited gatekeeping behaviours. Kulik and Tsoref (2010) found that more traditional gender role attitudes among mothers were associated with gatekeeping as well as with mothers’ satisfaction with fathers’ care. Recently, Pedersen and Kilzer (2014) found mothers’ gatekeeping impacts fathers’ work-to-family balance negatively. According to a study by Gaunt (2008), gatekeeping mothers reported lower self-esteem, higher religiosity and stronger feminine gender orientation and maternal identities; they also worked and earned less, had lower work attachment and were less educated than non-gatekeeping mothers.

Whereas gatekeeping studies focus on very specific interactions and behaviours between mothers and fathers, our study focuses more on the overall time spent in parenting. Therefore, bargaining theory is a more appropriate theoretical model for interpreting our results. More-educated mothers may be more subtle in their maintenance of parenting power by maintaining the lion’s share of parenting time. Particularly, if fathers of similar education levels earn more from their work, mothers can increase their household bargaining power by spending more time on childcare (Basu, 2006).

Additionally, more-educated mothers may be feeling the pressures of investing in parenting their children along with the time bind that goes along with working. Given that fathers are more likely to adhere to a breadwinner schema, mothers are more likely to pick up the slack on increased parenting pressure. Irrespective of sharing an egalitarian ethic or not, fathers are not going to feel the same cultural pressure to share increased parenting.

We also suggest that expectations constantly experienced by working mothers play a role in these findings. Maybe a more-educated mother feels societal pressure if she does not do enough at home even after higher levels educational attainment. As we know, university education teaches people progressiveness (Reese, 2001; Schoon, Cheng, Gale, Batty, & Deary, 2010). Because they are trained in a progressive way, they feel more peer pressure from friends and colleagues. Structurally, mothers who have more education have more flex time and work-at-home time which would necessitate more boundary setting with family roles. Schieman and Glavin (2016) found that middle-class fathers experience more role blurring. This could also be the case with more-educated mothers when it comes to childcare. More flexibility becomes a parenting trap.

The beginning of the parenting stage of any family represents a critical shift in the operation of the family. It is not only a rite of passage but also a juncture at which negotiation for time increases dramatically (Kulik & Tsoref, 2010) as work and family conflicts are resolved and negotiated. It is at this point that gender roles and their manifest form are likely to mark the beginning of a long journey of sustained expectations. Ridgeway (2011) argues that the home is the ultimate context in which gender relations are formed and persist. Findings from qualitative research suggest that fathers and mothers do have strong feelings about gendered parenting roles with their first child (Miller, 2011; Rose, Brady, Yerkes, & Coles, 2015).

5.1 Future Research and Limitations

Researchers should focus in more detail on specific domains of work–family balance satisfaction. The global nature of the measure is limiting in terms of interpretation. In addition, more attention should be paid to the status of the parents in assessing these attitudes. While we did find major differences based on educational status, less-limited measures of educational status could be used to fully explore this relationship. Future studies could also examine the implications of these different parenting configurations and inequalities on the quality of the marital/common law relationship. The trend for higher-educated mothers to increase parenting time should be monitored closely, given a growing backlash against helicopter parenting. While the extension of the Blinder–Oaxaca decomposition analysis to the gender gap in parenting time contributes to the literature on this topic, it should be pointed out that the counterfactual approach is limited theoretically and statistically. The unexplained component of the gap does contain differential influences of the characteristics, but it also includes unobserved factors which impact the gender gap.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship and/or publication of this article.

Footnotes

Appendix

Table A1

Ordinary Least Squares Regression of Hours Spent in Childcare Time on Demographics, Respondent and Spouse’s Characteristics, and Gender by Education Interaction (N = 1782)

Models	1	2
Gender (Female = 1)	4.440***	4.398***
Education (Bachelors or Higher = 1)	1.361	0.186
Age of Respondent (20-29 = 1)	2.320*	2.401*
Region (Atlantic =1)	0.745	0.740
Number of Children	2.881***	2.869***
Market Work Hours	–0.157***	–0.156***
Household Labour Hours	–0.162***	–0.161***
Regular Shift	–0.997	–0.953
Works at Home	–0.584	–0.545
Flextime	1.011	1.033
Spouse’s Childcare Hours	0.537*	0.511*
Spouse’s Household Labour Hours	–0.317	–0.350
Spouse’s Market Work Hours	0.028	0.027
Spouse’s Shift (regular =1)	–1.269	–1.430
Regularly Hire for Childcare	3.184***	3.226***
Gender * Education		2.844*
R2	0.164	0.166
F	23.14	21.99

Source: Authors’ calculation based on 2010 General Social Survey in Time Use.

Note: *p <= .05, **p<= .01, ***p<= .001

Notes

References

Allen

S. M.

& Hawkins

A. J.

(1999). Maternal gatekeeping: Mothers’ beliefs and behaviours that inhibit greater father involvement in family work. Journal of Marriage and the Family, 61(1), 199–212.

Basu

(2006). Gender and say: A model of household behaviour with endogenously determined balance of power. The Economic Journal, 116(511), 558–580.

Becker

(1981). A treatise on the family. Cambridge, MA: Harvard University Press. Enlarged edition.

Bittman

England

Sayer

Folbre

& Matheson

(2003). When does gender trump money? Bargaining and time in household work. American Journal of Sociology, 109(1), 186–214.

Boll

Leppin

& Reich

(2014). Paternal childcare and parental leave policies: Evidence from industrialized countries. Review of Economics of the Household, 12(1), 129–158.

Boudarbat

& Connolly

(2013). The gender wage gap among recent postsecondary graduates in Canada: A distributional approach. Canadian Journal of Economics, 46(3), 1037–1065.

Brown

L. K.

Troutt

& Prentice

(2011). Ten years after: Sex and salaries at a Canadian university. Canadian Public Policy, 37(2), 239–255.

Buchanan

Das

& McFarlane

(2017). Gender differences in within-couple influences on work–family balance satisfaction: When benefits become threats. Journal of Family Studies. doi: 10.1080/13229400.2017.1335225

Buchanan

McFarlane

& Das

(2016). A counterfactual analysis of the gender gap in parenting time: Explained and unexplained variances at different stages of parenting. Journal of Comparative Family Studies, 47(2), 193–219.

10.

Carlin

P. S.

& Flood

(1997). Do children affect the labor supply of Swedish men? Time fiary vs. survey data. Labour Economics, 4(2), 167–183.

11.

Dotti Sani

G. M.

& Treas

(2016). Educational gradients in parents’ child-care time across countries, 1965–2012. Journal of Marriage and the Family, 78, 1083–1096.

12.

Drolet

(2001). The persistent gap: New evidence on the Canadian gender wage gap (Statistics Canada Analytical Studies Branch Working Paper, No. 157). Retrieved from https://ssrn.com/abstract=310600orhttps://dx-doi-org.web.bisu.edu.cn/10.2139/ssrn.310600

13.

Duncan

& Pettigrew

R. N.

(2012). The effect of work arrangements on perception of work-family balance. Community, Work & Family, 15(4), 403–423.

14.

Evans

D. A.

Hebert

L. E.

Beckett

L. A.

Scherr

P. A.

Albert

M. S.

Chown

M. J.

… Taylor

J. O.

(1997). Education and other measures of socioeconomic status and risk of incident Alzheimer disease in a defined population of older persons. Archives of Neurology, 54(11), 1399–1405.

15.

Gaunt

(2008). Maternal gatekeeping antecedents and consequences. Journal of Family Issues, 29(3), 373–395.

16.

Gauthier

Smeeding

& Furstenberg

F. F.

(2004). Are parents investing less time in children? Trends in selected industrialized countries. Population and Development Review, 30(4), 647–671.

17.

Hoffman

(2012). Canadian dads @ home. Fatherhood Involvement Research Alliance. Retrieved from http://www.fira.ca/article.php?id=140

18.

Kohn

M. L.

(1969). Class and conformity: A study in values. Homewood, IL: Dorsey.

19.

Kohn

M. L.

& Schooler

(1973). Occupation experience and psychological functioning: An assessment of reciprocal effects. American Sociological Review, 38(1), 97–118.

20.

Kulik

& Tsoref

(2010). The entrance to the maternal garden: Environmental and personal variables that explain maternal gatekeeping. Journal of Gender Studies, 19(3), 263–277.

21.

Lareau

(2011). Unequal childhoods: Class, race, and family life (2nd ed.). London, UK: University of California Press, Ltd.

22.

Lundberg

& Pollak

R. A.

(1996). Bargaining and distribution in marriage. Journal of Economic Perspectives, 10(4), 139–158.

23.

Maume

D. J.

(2011). Reconsidering the temporal increase in fathers’ time with children. Journal of Family and Economic Issues, 32(3), 411–423.

24.

McElroy

(1990). The empirical content of Nash-bargained household behaviour. Journal of Human Resources, 25(4), 559–583.

25.

Miller

(2011). Falling back into gender? Men’s narratives and practices around first-time fatherhood. Sociology, 45(6), 1094–1109.

26.

Monna

& Gauthier

A. H.

(2008). A review of the literature on the social and economic determinants of parental time. Journal of Family and Economic Issues, 29(4), 634–653.

27.

Nelson

M. K.

(2010). Parenting out of control: Anxious parents in uncertain times. New York, NY: New York University Press.

28.

Park

& Lau

A. S.

(2016). Socioeconomic status and parenting priorities: Child independence and obedience around the world. Journal of Marriage and the Family, 78(1), 43–59.

29.

Pedersen

D. E.

& Kilzer

(2014). Work-to-family conflict and the maternal gatekeeping of dual-earner mothers with young children. Journal of Family and Economic Issues, 35(2), 251–262.

30.

Pronovost

(2007). Le temps dans tous ses états: temps de travail, temps de loisir et temps pour le famille à l’aube du XXIe siècle. Enjeux Public IRPP, 8(1), Février.

31.

Reese

(2001). The origins of progressive education. History of Education Quarterly, 41(1), 1–24.

32.

Ridgeway

C. L.

(2011). Framed by gender: How gender inequality persists in the modern world. New York, NY: Oxford University Press.

33.

Rose

Brady

Yerkes

M. A.

& Coles

(2015). Sometimes they just want to cry for their mum: Couples’ negotiations and rationalisations of gendered divisions in infant care. Journal of Family Studies, 21(1), 38–56.

34.

Samuelson

(1956). Social indifference curves. The Quarterly Journal of Economics, 70(1), 1–22.

35.

Sayer

L. C.

Gauthier

A. H.

& Furstenberg

F. F.

(2004). Educational differences in parents’ time with children: Cross-national variations. Journal of Marriage and Family, 66(5), 1152–1169.

36.

Schieman

& Glavin

(2016). The pressure-status nexus and blurred work–family boundaries. Work and Occupations, 43(1), 3–37.

37.

Schoon

Cheng

Gale

C. R.

Batty

& Deary

I. J.

(2010). Social status, cognitive ability, and educational attainment as predictors of liberal social attitudes and political trust. Intelligence, 38(1), 144–150.

38.

Social Progress Index (2016). Social progress imperative. Retrieved from http://www.socialprogressimperative.org/countries/can/

39.

Statistics Canada (2011). Labour force survey. Ottawa, Canada.

40.

Turcotte

(2007). Time spent with family during a typical workday, 1986 to 2005. Canadian Social Trends.

41.

World Economic Forum (2015). The global gender gap report, 2015. Geneva, Switzerland: World Economic Forum. Retrieved from http://www3.weforum.org/docs/GGGR2015/cover.pdf

42.

Zuo

& Shengming

(2000). Breadwinner status and gender ideologies of men and women regarding family roles. Sociological Perspectives, 43(1), 29–43.

43.

Zuzanek

(2001). Parenting time: Enough or too little? Canadian Journal of Policy Research, 2, 1–12.