A cross-national analysis of parental involvement and student literacy

Abstract

Using data from the Programme for International Student Assessment (PISA), we examine the association between parental involvement and student literacy in 21 countries. We consider how the nature of the association between parental involvement and student literacy varies in direction and magnitude across national borders and across multiple dimensions of parental involvement and measures of literacy. Across the 21 countries, we observe that, in general, increased social and cultural communication with parents is associated with higher levels of student literacy, although the association is most consistent in the area of reading literacy. Specifically, for students residing in eight countries (Australia, Austria, Canada, Denmark, Finland, Ireland, Norway and the United Kingdom), there are consistent reading literacy benefits when their parents engage in various forms of social and cultural communication. Consistently across all 21 countries, students have significantly lower literacy scores the more frequently parents assist with homework. This finding provides robust cross-national support for the reactive hypothesis.

Keywords

Cross-national ecological systems theory literacy achievement parental involvement PISA

Introduction

Children’s background characteristics and family resources continue to eclipse extra-familial resources (e.g. school factors) in explaining many outcomes, including academic achievement (Alexander and Entwisle, 1988; Coleman et al., 1966, 1982; Fryer and Levitt, 2004; Jencks et al., 1972; White, 1982). Parental involvement is a key family resource that is linked to child well-being across a range of outcomes including educational well-being (e.g. Moore et al., 2009; Scaramella et al., 1998; Young et al., 1995; Zaslow et al., 2006). Moreover, a large quantity of research conducted primarily in the United States but also in other countries documents the importance of parental involvement to student success (e.g. Abd-El-Fattah, 2006; Barnard, 2004; Hango, 2007; Jeynes, 2005a, 2005b, 2007; Salazar et al., 2001; Spera, 2005). Despite the abundant body of country specific research examining the role of parental involvement in child well-being, little cross-national research has emerged on this topic. As such, researchers understand less about how parental involvement, or its association with educational well-being, varies across national borders.

Using data from the Programme for International Student Assessment (PISA), we examine the association between parental involvement and literacy among students 15 years of age in 21 countries. Literacy, as measured by PISA, differs from educational outcomes examined in other studies; rather than measuring solely academic or curriculum-specific knowledge and skills, literacy measures ‘the ability to apply a range of knowledge and skills to a variety of problems with real-life contexts’ (Lemke et al., 2004: 5). Specifically, we consider how the nature of the association between parental involvement and student literacy varies in direction and magnitude across national borders and across multiple dimensions of parental involvement. As the effects of parental involvement may vary across different literacy measures, we include three types of student literacy: reading, mathematics and science.

This study seeks to address the existing gap in cross-national research in this area. To date, much of the parental involvement literature has either been United States-based or single-country analysis conducted in disparate countries around the world, over varying periods, and using various measures. Given that much research in this area is conducted in the United States, it is of particular interest to see if the association between parental involvement and educational outcomes in other countries is similar to the research conducted using United States samples. This is possible by drawing upon international assessment data. Additionally, because PISA literacy scores measure students’ real-world knowledge rather than curriculum knowledge, PISA scores can be interpreted as measuring the education yield of countries as students are preparing to enter the workforce. As such, this study contributes to the field by using literacy as the international measure of achievement, rather than the curriculum-specific measure, which may be less comparable internationally.

Based on the research to date, we posit that parental involvement will generally have a positive association with student literacy across countries. At the same time, we expect that differences in the levels of parental involvement and the magnitude of its effects may help explain some of the variation in student literacy across countries.

Theoretical framework

Bronfenbrenner’s (1979) ecological systems theory of development is a useful conceptual model for examining the effects of country context on the association between parental involvement and students’ educational outcomes. According to ecological systems theory, children learn and grow in the context of multiple nested systems (i.e. the microsystem, mesosystem, exosystem and macrosystem), all of which interact with each other and with the child. Persons, processes and societal institutions at all levels, ranging from the family to schools they attend, to local school boards and to cultural norms, affect the child’s development.

The microsystem, which is most proximal to the child, includes individuals and societal institutions, such as the family, school and peers, which most directly interact with the child through interpersonal relationships and patterns of activity (Bronfenbrenner, 1989). Family and home environment are believed to have the largest influence on children, in particular at earlier ages as they represent the primary socializing influence. The interactions between individuals and settings of the microsystem make up the child’s mesosystem (Bronfenbrenner, 1989), where relationships between multiple settings have an additional influence on the child. The strongest connection in the mesosystem is the relationship between the family and the school, which has an influence on development over and above their individual contributions. The exosystem includes individuals, institutions and public policies that may affect children indirectly through their school and families (Bronfenbrenner, 1989). The macrosystem is the most distal system of influence in a child’s development and includes overarching patterns of values, beliefs and practices that characterize the other systems of influence and the broader social context. Although the effect is indirect, these larger societal processes permeate all stages and domains of development (Bronfenbrenner, 1989). In the current work, we examine how characteristics of the more distal systems, which are expected to vary between countries, moderate relationships in the child’s microsystem.

Parental involvement and educational outcomes

Consistent with Bronfenbrenner’s ecological systems theory, much research from the United States to date indicates that children benefit from parental involvement. Research links parental involvement with improved academic attainment (Barnard, 2004; Garnier and Stein, 1998) and achievement (Baker and Stevenson, 1986; Epstein, 1987, 1991; Fan and Chen, 2001; Jeynes, 2005a, 2005b, 2007; Lareau, 1987; Muller, 1993; Spera, 2005), in particular with respect to grades (Keith, 1991; Muller, 1993, 1995) and reading achievement (Epstein, 2001; Miedel and Reynolds, 1999). Furthermore, research shows that the quality of parent–child relationships, parent–child interactions and parental involvement in schools is positively related to children’s academic attainment and achievement (Muller, 1993; Steinberg, 1996; Teachman et al., 1996).

The results of other studies, however, call into question the assumption that parental involvement is associated with improved educational outcomes (see Epstein, 1991; Keith, 1991; McNeal, 1999, 2012; Muller, 1993, 1995). It is possible to summarize the results of these studies along three lines. The first set of studies suggests that some types of parental involvement are a response to poor school performance and thus, although they do not lead to poor outcomes, are associated with lower educational achievement. Epstein (1988), Muller (1995) and Horn and West (1992) find evidence of what has been coined the ‘reactive hypothesis’. Each researcher reveals that increased levels of parental homework supervision and assistance are negatively associated with academic achievement. The second set of studies suggests that the effect of parental involvement varies across educational outcomes. For example, Muller (1993) discovers that parental involvement (i.e. Parent Teacher Organizations) is positively associated with school grades, but demonstrates a weak and negative relationship with mathematics achievement (1995). Likewise, Keith (1991) ascertains that frequent parental involvement in school is significantly associated with higher grades but unrelated to reading and mathematics achievement scores.

The third body of research highlights the need to examine racial and gender differences when considering the impact of parental involvement on educational outcomes.¹ Specifically, Milne et al. (1986) discovered that the negative relationship between increased parental assistance in homework and educational achievement is limited to white students, and does not extend to blacks. McNeal (1999) found that white students benefit most consistently from parental involvement, while blacks experience limited benefits and Hispanics and Asians experience few, if any, benefits at all. Turning to gender, Muller (1993) reported a negative association between paternal school-related discussions for girls, but not for boys. Moreover, she found evidence to suggest that maternal discussions differ from those of fathers, with fathers’ discussions more likely to result from poor school performance (Muller, 1995).

In a more recent study, Graves and Wright (2011), using the ECLS-K data, established that European American parents were more likely to participate in home-based parental involvement activities than African American parents were. In turn, African American parents were more likely to participate in school-based activities. However, this research found very little evidence to suggest that parental involvement had differing effects on achievement by race/ethnicity.

Cross-national comparisons

A small but growing body of research suggests cross-national variations in the associations between parental involvement and other family factors and child outcomes. This is in accordance with ecological systems theory, since children growing up in different countries will have diverse influences on their development based on nations’ social, cultural, public policy and institutional differences. More specifically, differences in the association between parental involvement and child outcomes might be due to cross-country variation in levels of parental education, parental occupation, immigrant status, educational resources or family structure; differences in attitudes towards education and other relevant cultural characteristics; differences in educational systems and other social institutions; or differences in educational and social policies.

Among the limited body of research conducted in countries beyond the United States, most studies reveal positive educational outcomes related to parental involvement. For example, using longitudinal data from Britain, Hango (2007) as well as Flouri and Buchanan (2004) reported that parental involvement predicted educational attainment. Salazar et al. (2001) indicated that Canadian Filipino adolescents had higher school involvement when their parents’ involvement was greater. Another study found that in nine out of 12 countries,² children who reported that their mothers thought it was important for them to ‘do well’ in mathematics in school had higher average educational aspirations (Buchmann and Dalton, 2002).

Additionally, studies found differences across countries in the relative importance of parental involvement and of specific indicators of parental involvement. A study of adolescents in Egypt reported that parental involvement was a stronger predictor of academic performance than the level of parental education, school engagement or family structure (Abd-El-Fattah, 2006), in contrast to studies in the United States which found that family structure and parental education are stronger predictors of academic performance. Likewise, Hung (2007) discovered a positive relationship between the family home learning environment and Taiwanese children’s academic achievement. Yet, while a positive link exists between parental aspirations and involvement in the home and children’s self-concept, there was no significant relationship with academic achievement (Hung, 2007). In a cross-national study of 18 countries, Hampden-Thompson (2009) found that both social and cultural parental involvement along with students’ socio-economic status were key predictors of reading literacy for students who resided in single-parent households in Canada, Great Britain, Ireland, New Zealand and the United States. However, these results were not consistent across the other 13 countries in the study.

A small number of studies examined the role of national context in children’s outcomes, including the manner in which microsystem and mesosystem relationships and practices are influenced by the national context. Pong et al. (2003), for example, established that living with a single parent rather than two parents was more detrimental to student achievement in the United States and New Zealand than in nine other countries. They concluded that the gap in mathematics and science scores between children in single- and two-parent families is attenuated in countries where family policies are implemented to equalize resources between families with different structures. Similarly, Hampden-Thompson and Pong (2005) concluded from a study of 14 European countries that national family and welfare policies could alleviate the negative educational outcomes that are associated with having a single parent. Similarly, Hampden-Thompson (2013) found that the association between the reading, mathematics and science literacy gaps between students from single-mother and two-parent families is influenced by a country’s economic policy environment.

In their study of 12 countries, Buchmann and Dalton (2002) discovered that only in the three countries with highly stratified educational systems were neither parent nor peer attitudes significant predictors of students’ educational aspirations. Park’s (2008) study of 14 countries concluded that in countries with high levels of educational standardization, parent–child communication is more beneficial to students with low socio-economic status (SES) than those of a high-SES background. This research adds and complements this previous work by including a larger number of countries and focuses on more parental involvement measures.

Measurement of parental involvement

Although there is little consensus about how to measure parental involvement, and indicators are numerous and vary widely, most studies measured it along two dimensions. The first centers on whether parental involvement is school- or home-based (see Muller, 1993). The second recognizes the importance of the impetus to parental involvement, and thus distinguishes involvement that seeks to enhance or manage a child’s educational career from that which is crisis or problem motivated. Ho and Willms (1996) distinguish further home-based involvement as either monitoring out-of-school activities or facilitating school performance through parent–child-centered activities and relationships.

This analysis focuses only on parental involvement within the boundaries of the home. The dimensions of parental involvement included in this study are social communication, cultural communication and parental assistance with homework, which is a type of parental monitoring of school-based activities. By concentrating on those behaviors that take place within the home, external factors such as school policies that may influence parental involvement can be controlled (see Muller, 1995).

Research questions

Data from PISA offer several advantages including: (1) data from numerous countries; (2) multiple measures of parental involvement; and (3) multiple measures of student literacy. Using data from PISA 2000 we address the following research questions:

Is parental involvement associated with student literacy? If so, does this association vary across the three dimensions of parental involvement considered? We hypothesize that the association between parental involvement and educational outcomes observed in prior studies will extend to student literacy. Thus, we expect that parental involvement will be associated with student literacy in most countries. We also expect that there will be variation across the three dimensions of parental involvement considered. We extrapolate from existing research (mostly nationally based studies) and hypothesize that parental involvement as measured by social and cultural communication, in general, will be positively associated with student literacy across countries. Consistent with the reactive hypothesis, we hypothesize that parental assistance with homework will be negatively related to student literacy.

Does the association between parental involvement and student literacy differ across the three domains – reading, mathematics and science – of student literacy examined? We hypothesize that the association between parental involvement and student literacy will differ across the three domains – reading, mathematics and science – of student literacy examined. In particular, studies suggest that there is a linkage between parental involvement and reading literacy (Epstein, 2001; Miedel and Reynolds, 1999). As such, we expect to discover a strong association between parental involvement and reading literacy and a weaker (or no) association with science and mathematics literacy.

Does the association between parental involvement and student literacy vary across countries? If so, does it vary across the dimensions of parental involvement? We expect the association patterns of the types of parental involvement to vary across countries. The theoretical basis for this hypothesis can be found in Bronfenbrenner’s (1989) ecological systems theory: differences between countries in terms of institutions and public policies as well as the overarching cultural context will moderate the effects of schools and families on children’s outcomes. Our literature review also found some empirical support for this hypothesis (Buchmann and Dalton, 2002; Hampden-Thompson, 2013; Hampden-Thompson and Pong, 2005; Pong et al., 2003).

Data and methods

Sponsored by the Organization for Economic Co-operation and Development (OECD), PISA 2000 is a unique international study conducted in 32 countries,³ including OECD countries (US Department of Education, 2001). The main goal of PISA is to provide assessments of general and cross-curricular competencies of nationally representative samples of youth in reading, mathematics and scientific literacy, in order to enable participating countries to gauge their success at meeting educational objectives at both a national and international level. Data collection occurs every three years (i.e. 2000, 2003, 2006, 2009, 2012 and so on), with the focus of each cycle rotating between the three literacy areas; the first cycle (2000) focused on reading literacy. We do not use data from the 2003 PISA onwards in this study because measures of parental involvement were not included in all countries’ student background questionnaires. Thus, the 2000 PISA provides data on parental involvement for the greatest number of countries. To increase international comparability, and in order to assess the educational yield of countries, target respondents in PISA are 15-year-old students enrolled in secondary schools: that is, students who have reached, or are near reaching, the end of compulsory schooling.

Selection of countries

Of the 32 countries that participated in the first cycle of PISA in 2000, we excluded 11 countries from our study. We excluded the Netherlands, for example, because their sample of students did not meet the minimum reporting standards set by PISA. The Southeast Asian countries of Japan and Korea did not include key socioeconomic variables, such as parents’ education and/or immigrant status, in their publicly released data sets. Therefore, we excluded those countries. We excluded Liechtenstein because of its small size and lack of variation. Brazil and Mexico were also excluded because we were mainly interested in comparing Western industrialized countries and not emerging economies. We include the following 21 countries: Australia, Austria, Belgium, Canada, the Czech Republic, Denmark, Finland, France, Germany, Greece, Hungary, Ireland, Italy, New Zealand, Norway, Portugal, the Russian Federation, Spain, Sweden, United Kingdom (UK) and the United States.⁴ It is worth noting that in addition to being Western industrialized countries, with the exception of the Russian Federation, all of the countries included in this study are member countries of the OECD. By focusing on these, we are able to increase the comparability of countries and minimize, in part, the variability in economic development.

Measures and analysis plan

Reading, mathematics and science literacy are the dependent variables for this analysis. Reading literacy was measured in terms of students’ ability to use texts ranging in style from forms to manuals, newspaper articles and entertainment pieces in order to perform three types of processes: retrieving information; interpreting texts’ meaning and drawing inferences; and relating texts to their own experience, knowledge, and ideas (US Department of Education, 2001). Mathematics literacy involved applying ‘big ideas’ of mathematical content in areas such as space and shape and demonstrating mathematical competencies such as modeling and problem-solving. Science literacy involved applying scientific concepts such as energy conservation and decomposition, drawn mainly from the fields of physics, biology and chemistry, and demonstrating process skills such as identifying evidence, and drawing, evaluating and communicating conclusions (Organization for Economic Co-operation and Development, 1999; Schleicher and Tamassia, 2000; for sample items see Organization for Economic Co-operation and Development, 2009).

Using item response theory (IRT), plausible values were generated for each student based upon their responses to the test items they were given and the performance of students with similar characteristics to other test items. In total, five plausible values were randomly estimated for each student in each domain. This generated a score on a scale from 0 to 1000. Across all 2000 PISA surveys, the mean was 500 and the standard deviation was 100 (US Department of Education, 2001). In order to ensure that standard error estimates included a component of sampling variance and measurement error, analyses conducted for this study combined all five plausible values as suggested in the PISA User Manual (OECD, 2000: 24).

As noted above, we consider three types of parental involvement: (1) social communication; (2) cultural communication; and (3) parental assistance with homework. Social communication with parents is measured through students’ reports of the frequency with which they (a) eat a meal with a parent and (b) talk about general issues. Parent–child cultural communication is measured by responses to items on the frequency with which students discuss (a) political or social issues and (b) books, films or television with their parents. Last, we measure parental involvement in school through students’ responses to an item asking them how frequently they receive help from their parent with homework. Response options for each of the items are: never, a few times a year, about once a month, several times a month, several times a week, and every day. Even though the same types of involvement were measured in each country, differences in cultural context might confound cross-national comparisons. For example, ‘eating a meal together’ might have a different meaning across countries. However, these standardized measures have been used in cross-national surveys before PISA 2000 (e.g. Third International Mathematics and Science Study) and provide the opportunity for cross-national differences in parental involvement to be examined.

Rather than create a composite scale for each type of parental involvement, we include each of the items separately in the analyses. The reasons for doing so are threefold. First, we want to ensure that the underlying association between student literacy and, for example, social communication is not masked by the use of a composite scale. For instance, it is possible that a strong positive association between the first indicator of social communication and student literacy will be obfuscated by a weak or negative association with the second indicator. Second, we are interested in whether the observed association varies either in direction or in magnitude across the different indicators of social and cultural communication and across countries. Third, the results of exploratory factor analysis show some variability in the extent to which the indicators of social and cultural communication load together across countries, suggesting that a composite scale may be appropriate for some countries but not for others.

Additional independent variables are included in the OLS regression analysis in order to control for factors that research has shown to be strongly associated with achievement. These include gender, student grade level, parents’ occupational status, parent education, family structure, immigrant status, and number of books in the home. Parents’ occupational status, parent education and number of books in the home are standard measures of socio-economic status used in international surveys (see Buchmann, 2002). PISA 2000 was sampled based on age and not grade. Therefore, across countries there is a broad range of grade levels for the 15-year-olds in those countries. Consequently, any analysis of the PISA data should include the student’s grade level as a control variable. Cross-national research indicates that patterns of gender differences in academic achievement are often dependent upon subject matter (Mullis et al., 2000). Analysis of PISA 2000 shows that, on average, in every country, girls demonstrate higher levels of reading literacy than boys. In mathematics, boys outperform girls in about half of the countries, while in science literacy the differences are small with no statistically significant differences between the sexes in the OECD countries (Organization for Economic Co-operation and Development, 2001). Finally, immigrant status is included in the OLS regression as a control due to prior research indicating that it is also a significant predictor of academic achievement (Levels et al., 2008).

The data analysis consists of two distinct stages. We begin by exploring the levels of parental involvement across the 21 countries and go on to use OLS regressions to estimate the association between parental involvement and reading, mathematics and science literacy for each country included in this study separately. In the OLS models, the standard error estimates are corrected using replicate weights and are conducted in the statistical package WesVar. The WesVar software, which was specifically designed to use with the PISA 2000 data, estimates sampling variances for complex design (i.e. two-stage sampling designs) through replication methods. Fay’s Balanced Repeated Replication method was used in the preparation of the PISA data sets, with the data files containing 80 replicates. We used these replicate weights to estimate the sampling variance for the computed statistics. For a complete description of the WesVar software and the Fay’s Balanced Repeated Replication method, see the user’s guide that accompanies the WesVar software (Westat, 2000).⁵

Results

Table 1 contains the weighted means for the main variables of interest by country. Across the 21 countries, Finland has the highest average literacy scores in reading and science literacy (547 and 538 score points, respectively) with New Zealand students having the highest mathematics literacy scores (538 score points). The Russian Federation, Greece and Portugal have the lowest average point scores for reading (426), mathematics (447) and science literacy (459), respectively.

Table 1.

Mean literacy achievement and frequency of parental involvement, by country.

	Literacy achievement			Parental involvement
	Reading	Mathematics	Science	Eat Meal with Parent	Talk with Parent	Discuss Politics or Social Issues with Parent	Discuss Books, Films, and TV with Parent	Parent Helps with Homework
Australia	528	534	527	3.34	3.08	1.58	2.07	2.05
Austria	508	515	519	3.44	2.98	1.57	1.94	1.79
Belgium	507	519	496	3.77	3.13	1.28	2.01	1.52
Canada	534	533	530	3.46	3.06	1.60	2.39	1.74
Czech Republic	492	497	512	3.52	3.54	1.18	2.36	2.08
Denmark	497	514	481	3.77	3.58	1.80	2.38	2.35
Finland	547	537	538	3.37	3.66	1.42	2.41	1.86
France	504	518	501	3.79	3.36	1.91	2.70	1.91
Germany	484	490	486	3.65	2.91	1.50	2.06	1.76
Greece	474	447	460	3.48	3.36	1.99	2.40	1.36
Hungary	480	488	496	3.56	3.85	1.59	2.92	2.01
Ireland	527	503	513	3.53	3.29	1.39	2.32	1.58
Italy	487	459	478	3.88	3.76	2.26	2.81	1.02
New Zealand	529	538	528	3.15	3.08	1.73	2.31	2.25
Norway	506	499	500	3.77	3.37	1.40	1.98	1.82
Portugal	470	453	459	3.74	3.53	1.36	2.35	1.65
Russian Federation	462	478	461	3.78	3.59	1.51	2.67	2.12
Spain	493	476	491	3.66	3.29	1.50	2.83	1.48
Sweden	515	510	511	3.75	3.13	1.30	2.22	2.36
United Kingdom	524	530	533	3.25	3.34	1.33	2.71	2.11
United States	504	493	500	3.20	3.33	1.91	2.56	1.73

Note: Weighted means.

Turning to our indicators of parental involvement, our descriptive analysis indicates that across national borders children and their parents frequently interact with one another in a variety of ways. The most common of these is eating meals together as a family – an activity that many have noted as important in structuring family time and interactions (e.g. Guzman and Jekielek, 2004). The data also suggest that the frequency with which parents and students interact, and the activities they most commonly engage in, does vary across countries. In all but five countries (the Czech Republic, Finland, Hungary, United Kingdom and the United States), eating meals together is the activity that students report engaging in most frequently with their parents. Talking with parents about general issues follows closely, while cultural communication, as measured by discussions about books, films or TV, or political and social issues, and homework assistance are reported to occur on a less frequent basis. Comparing mean levels of activities across the countries, we see that youth in Italy report eating meals and discussing politics or social issues more frequently with their parents than their counterparts in other countries. Students in Hungary have more frequent discussions about general issues and about books, films or TV with their parents than those in other countries. Students in the Scandinavian countries of Denmark and Sweden receive more frequent help from their parents on homework than youth in other countries.

Single country OLS regression analyses

Table 2 presents the results of the OLS models regressing parental involvement on reading literacy by country. The models include control variables for gender, school grade, immigrant status, family structure, parental occupation, parents’ education, and the number of books in the home.⁶ As seen in Table 2, with the exception of Italy (talking with parents about general issues), our two indicators of social communication – eating meals together and talking about general issues – are positively associated, although not always significantly, with reading literacy in all 21 countries included in this study. Specifically, we discovered that eating meals together with a parent is significantly associated with higher levels of reading literacy in all but four countries (Hungary, New Zealand, Spain and the United States). For example, a one-unit increase in the frequency with which students report eating meals together with their parents is associated with a 16-point increase in reading literacy among students in Denmark.⁷ The association between talking with parents about general issues and reading literacy is less consistent across countries. In just under half of the 21 countries, frequent conversations with parents about general issues are associated with significant increases in reading literacy.

Table 2.

Reading literacy OLS regression coefficients for parental involvement, by country.

Country	Parental involvement															Intercept		R ²	Number of Observations^a
Country	Eat Meal with Parent	SE		Talk with Parent	SE		Discuss Books, Films or TV	SE		Discuss Politics or Social Issues	SE		Parent Helps with Homework	SE		Intercept		R ²	Number of Observations^a
Australia	2.73	1.39	*	3.67	1.31	**	6.21	1.44	**	15.05	1.57	**	−10.55	1.16	**	205.6	**	0.32	5140
Austria	4.23	1.72	**	2.97	1.25	**	5.53	1.09	**	10.09	1.29	*	−13.65	0.88	**	208.0	**	0.36	4680
Belgium	10.29	1.73	**	0.88	1.11		2.04	0.89	*	7.12	0.89	**	−11.82	0.89	**	160.5	**	0.52	6615
Canada	7.73	0.69	**	4.76	0.76	**	5.60	0.71	**	9.07	0.76	**	−13.86	0.65	**	260.7	**	0.28	29,375
Czech Republic	6.06	1.29	**	1.13	1.41		5.11	0.90	**	9.30	1.00	**	−12.39	0.93	**	154.9	**	0.47	5345
Denmark	16.21	2.78	**	3.22	1.66	*	6.60	1.25	**	11.82	1.35	**	−11.32	1.22	**	157.5	**	0.35	4210
Finland	3.85	1.75	*	7.38	2.24	**	6.65	1.33	**	10.62	1.11	**	−14.06	1.35	**	288.8	**	0.28	4860
France	4.71	1.86	**	2.36	1.35		2.67	0.98	**	6.05	0.86	**	−9.08	0.83	**	218.3	**	0.52	4645
Germany	5.59	1.53	**	1.84	1.20		3.96	1.13	**	6.88	1.42	**	−11.85	1.28	**	118.6	**	0.52	5010
Greece	3.68	1.65	*	3.63	1.97		2.95	1.27	*	9.66	1.41	**	−12.56	1.17	**	171.5	**	0.30	4635
Hungary	1.30	1.58		6.43	2.57		3.82	1.15	**	8.14	1.12	**	−11.73	1.02	**	200.6	**	0.48	4850
Ireland	3.03	1.57	*	5.90	1.55	**	5.80	1.13	**	7.68	1.10	**	−13.89	1.08	**	141.9	**	0.30	3825
Italy	10.37	2.79	**	−0.22	2.33		2.80	1.13	**	8.66	0.90	**	−11.56	0.92	**	149.6	**	0.30	4955
New Zealand	0.31	1.37		6.72	1.56	**	0.06	1.46		12.05	1.44	**	−11.46	1.45	**	41.6		0.32	3650
Norway	11.94	2.86	**	3.89	1.89	*	3.95	1.38	**	15.26	1.35	**	−13.30	1.38	**	96.0	*	0.27	4080
Portugal	5.10	1.43	**	1.97	1.54		5.02	1.06	**	9.02	0.94	**	−6.10	0.77	**	234.0	**	0.59	4550
Russian Federation	9.25	1.98	**	2.54	1.59		3.26	0.83	**	7.91	0.72	**	−9.30	0.77	**	171.8	**	0.30	6685
Spain	2.70	1.50		1.52	1.20		7.48	1.22	**	6.71	0.95	**	−7.95	0.72	**	218.5	**	0.45	6075
Sweden	6.23	1.87	**	2.11	1.30		6.20	1.10	**	11.48	1.24	**	−15.10	1.22	**	242.6	**	0.30	4380
United Kingdom	4.60	1.08	**	5.96	1.50	**	8.93	1.14	**	7.26	1.31	**	−15.89	1.13	**	168.7	**	0.35	9230
United States	2.26	1.31		3.98	1.79	*	7.37	1.78	**	3.96	1.24	**	−14.57	1.46	**	264.3	**	0.35	3845

** = p .01, * = p. 05.

Number of observations represents the unweighted average.

Note: Models include controls for student gender, grade, immigrant status, family structure, parental occupation, parent’s education, and number of books in the home.

Turning to our two indicators of cultural communication, we observe that frequent discussions about books, film or TV, as well as political or social issues, are significantly associated with higher scores in reading literacy in all of the countries examined, with the exception of New Zealand (discuss books, films or TV). We also discover that frequent parental assistance with homework is associated with significantly lower levels of reading literacy across the 21 countries. This finding is consistent with the hypothesis that increased parental involvement in this area is a response to poor school performance (i.e. the reactive hypothesis: see McNeal, 2012).

Table 3 presents the results for models estimated for mathematics literacy by country. Turning first to social communication, we observe that in the majority of the countries − 13, out of 21 − the frequency with which students eat meals with their parents is significantly associated with higher levels of mathematics literacy. On the other hand, the association between the frequency with which students talk with parents about general issues and mathematics literacy is significant in only four countries (Finland, Hungary, New Zealand and the United States).

Table 3.

Mathematics literacy OLS regression coefficients for parental involvement, by country.

Country	Parental involvement															Intercept		R ²	Number of Observations^a
Country	Eat Meal with Parent	SE		Talk with Parent	SE		Discuss Books, Films or TV	SE		Discuss Politics or Social Issues	SE		Parent Helps with Homework	SE		Intercept		R ²	Number of Observations^a
Australia	4.37	1.43	**	2.48	1.73		2.81	1.58		9.66	1.98	**	−8.35	1.61	**	300.7	**	0.28	2835
Austria	8.08	2.13	**	−0.50	1.73		3.54	1.71	*	7.51	1.85	**	−14.23	1.18	**	270.6	**	0.33	2605
Belgium	8.91	2.10	**	−0.84	1.56		0.90	1.27		3.54	1.20	**	−9.52	1.13	**	204.2	**	0.51	3750
Canada	8.91	0.90	**	0.40	0.86		3.77	0.81	**	8.07	0.82	**	−12.64	0.68	**	317.5	**	0.21	16,310
Czech Republic	8.67	2.19	**	−1.18	2.00		3.44	1.28	**	7.56	1.52	**	−13.78	1.30	**	274.0	**	0.34	3055
Denmark	9.87	3.48	**	5.31	2.91		3.99	1.42	**	8.72	1.55	**	−11.47	1.29	**	230.8	**	0.31	2370
Finland	5.45	1.69	**	6.38	2.27	**	0.87	1.43		10.47	1.12	**	−11.78	1.36	**	344.9	**	0.19	2700
France	5.41	2.14	**	2.00	1.56		1.35	1.34		2.63	1.08	*	−8.96	1.28	**	270.2	**	0.46	2580
Germany	7.94	2.37	**	−0.72	1.84		0.73	1.35		3.09	2.17		−12.09	1.36	**	235.7	**	0.45	2790
Greece	4.56	2.44		0.24	2.86		0.69	1.67		9.90	1.90	**	−13.43	1.80	**	139.3	**	0.29	2580
Hungary	4.09	2.26		7.74	3.88	*	−0.98	1.48		6.36	1.66	**	−11.59	1.24	**	221.1	**	0.37	2780
Ireland	3.69	1.68	*	3.46	1.98		2.43	1.62		3.90	1.35	**	−11.57	1.27	**	182.0	**	0.28	2110
Italy	5.73	3.78		4.55	3.07		−0.41	1.46		3.74	1.23	**	−12.58	1.42	**	143.9	**	0.23	2745
New Zealand	1.22	1.68		4.31	2.17	*	0.04	1.84		10.44	1.75	**	−10.83	1.84	**	125.2	**	0.27	2035
Norway	16.44	3.77	**	−0.81	2.38		1.22	1.88		13.04	1.85	**	−11.97	1.53	**	182.0	**	0.22	2265
Portugal	4.45	2.41		3.27	1.88		1.58	1.45		6.91	1.34	**	−5.06	1.20	**	229.7	**	0.52	2520
Russian Federation	12.20	2.51	**	1.92	2.38		1.99	1.52		6.31	1.36	**	−8.88	1.07	**	216.0	**	0.18	3710
Spain	−0.20	2.71		−0.02	1.92		5.73	1.49	**	4.15	1.32	**	−5.85	1.21	**	222.3	**	0.37	3350
Sweden	5.10	2.82		−2.25	1.60		4.74	1.60	**	9.12	1.84	**	−14.22	1.73	**	242.9	**	0.26	2445
United Kingdom	5.26	1.40	**	1.77	1.83		5.50	1.54	**	5.97	1.76	**	−13.44	1.28	**	248.5	**	0.30	5125
United States	3.17	1.98		6.07	2.18	**	3.51	2.05		3.72	1.71	*	−13.72	1.85	**	274.6	**	0.36	2135

** = p .01, * = p. 05.

Number of observations represents the unweighted average.

Note: Models include controls for student gender, grade, immigrant status, family structure, parental occupation, parent’s education, and number of books in the home.

As in the case for reading literacy, all but one country (Germany) the frequency of discussions about political or social issues is significantly associated with higher levels of mathematics literacy. In contrast, in only seven out of the 21 countries was there a significant association between the frequency with which students discuss books, films or television with their parents and mathematics literacy scores. Consistent with the results for reading literacy, students who report receiving frequent help from their parent with homework have significantly lower levels of mathematics literacy in all 21 countries.

For our third and last dependent variable, science literacy, results are summarized in Table 4. As in the previous two models, there is, in general, a positive association between our indicators of social and cultural communication and science literacy, although the relationship is less consistent across countries. For social communication, we observed significant positive associations between science literacy and the frequency with which students and parents eat meals together in only seven of the 21 countries. Frequent discussions about books, films and TV – one of our two indicators of cultural communication – is significantly associated with higher levels of science literacy in 12 of the 21 countries. Frequent discussions with parents about political or social issues (our second indicator of cultural communication) are significantly associated with higher scores in science literacy in all 21 countries. Also consistent with models for reading and mathematics literacy, the associations observed between homework assistance and science literacy are negative and significant across all of the countries. Parent assistance with homework was a consistent finding across all the countries and measures of student literacy. We found that increased frequency in homework help resulted in lower literacy scores. When students are struggling, parents step in and try to help. This is consistent with prior research conducted in single countries. The cross-national consistency and robustness of this association is notable.

Table 4.

Science literacy OLS regression coefficients for parental involvement, by country.

Country	Parental involvement															Intercept		R ²	Number of Observations^a
Country	Eat Meal with Parent	SE		Talk with Parent	SE		Discuss Books, Films or TV	SE		Discuss Politics or Social Issues	SE		Parent Helps with Homework	SE		Intercept		R ²	Number of Observations^a
Australia	0.41	1.76		1.77	1.65		1.47	2.05		14.06	1.98	**	−8.18	1.77	**	228.8	**	0.25	2840
Austria	2.12	1.90		3.36	1.60	*	3.14	1.67		8.09	1.86	**	−14.94	1.30	**	285.1	**	0.32	2630
Belgium	8.43	2.18		2.04	1.28		3.61	1.30	**	6.00	1.24	**	−11.07	1.11	**	155.6	**	0.47	3690
Canada	6.69	0.94	**	3.13	0.94	**	5.00	0.92	**	8.46	1.09	**	−13.02	0.74	**	305.9	**	0.22	16,320
Czech Republic	5.52	1.97	**	2.38	1.99		4.18	1.57	**	10.53	1.45	**	−11.46	1.32	**	228.6	**	0.37	3050
Denmark	16.85	4.11	**	−1.07	2.44		7.44	1.82	**	11.93	1.86	**	−10.28	2.00	**	199.9	**	0.30	2335
Finland	3.44	1.73	*	8.01	2.39	**	1.19	1.54		12.20	1.85	**	−16.82	1.45	**	320.2	**	0.22	2710
France	8.98	2.78	**	−0.06	1.99		1.92	1.46		8.70	1.35	**	−8.47	1.19	**	193.7	**	0.47	2575
Germany	4.58	2.12	*	1.85	1.82		2.49	1.57		6.27	1.88	**	−9.75	1.49	**	245.8	**	0.40	2820
Greece	4.13	2.47		3.52	2.74		0.78	1.84		10.09	2.14	**	−11.90	1.34	**	207.6	**	0.25	2575
Hungary	4.67	1.87	**	0.25	3.82		3.87	1.57	**	8.28	1.55	**	−11.50	1.55	**	185.9	**	0.36	2780
Ireland	6.73	2.37	**	3.07	2.15		6.03	1.60	**	7.00	1.46	**	−13.51	1.39	**	224.4	**	0.25	2120
Italy	13.15	4.12	**	−4.19	2.94		−0.46	1.72		10.40	1.30	**	−10.00	1.30	**	121.7	*	0.23	2750
New Zealand	1.55	1.56		4.96	1.86	**	−4.00	1.87	*	16.65	1.56	**	−10.43	1.96	**	103.1	**	0.31	2015
Norway	11.80	3.13	**	0.39	2.38		3.24	1.78		13.67	1.75	**	−12.98	1.76	**	175.8	**	0.24	2260
Portugal	1.28	2.30		1.53	2.18		3.18	1.43	*	8.00	1.26	**	−5.64	0.98	**	272.0	**	0.49	2530
Russian Federation	2.96	3.08		6.04	1.95	**	3.75	1.01	**	7.45	1.08	**	−9.83	1.14	**	247.5	**	0.18	3710
Spain	4.28	1.67	**	−0.26	1.61		1.79	1.75		7.57	1.57	**	−9.12	1.27	**	224.4	**	0.37	3370
Sweden	6.79	3.28	*	3.08	1.91	**	4.50	1.64	**	11.34	1.56	**	−16.35	1.70	**	293.2	**	0.23	2425
United Kingdom	3.70	1.45	**	6.93	2.29	**	7.29	1.63	**	6.23	1.66	**	−17.95	1.35	**	194.3	**	0.34	5125
United States	1.47	2.11		3.75	2.73		6.08	2.08	**	4.68	1.77	**	−14.19	1.76	**	290.1	**	0.32	2125

** = p .01, * = p .05.

Number of observations represents the unweighted average.

Note: Models include controls for student gender, grade, immigrant status, family structure, parental occupation, parent’s education, and number of books in the home.

It is worth exploring, if not tentatively, the patterns that emerged from this study and attempt to typify them. To make sense of the variation, the countries were grouped qualitatively and four distinct types emerged. The first grouping (cultural communication beneficiaries) highlights the countries (Canada, Czech Republic, Denmark, Sweden, United Kingdom, United States) in which there was a significant association between all three domains of literacy and parental involvement in the form of cultural communication within the home. These countries were unique in that while the association between cultural communication and literacy was consistent, it was not the case for the measures of social communication. The second group (reading literacy beneficiaries) consists of countries (Australia, Austria, Finland, Ireland, Norway) where there was an association between all measures of social and cultural communication for reading literacy only. The salient feature of the third group (political/social issues discussion beneficiaries) was the association between discussing political or social issues with parents and the positive impact it had on reading, mathematics, and science literacy. The countries in this group included Belgium, Greece, Hungary, Italy, New Zealand, Portugal, Russian Federation, and Spain. Finally, the fourth group consists of only two countries (France and Germany). The common feature here was the consistent association between eating a meal with a parent/s and all three literacy domains. For these two countries, however, the associations between literacy and the other measure of social communication (talking with parent/s) and cultural communication were mixed (Table 5).

Table 5.

A typology based on the social and cultural communication findings by group name, country, and features associated with each grouping.

Grouping	Country	Features
Cultural communication beneficiaries	Canada	Countries in which students benefit educationally from cultural communication in the home. The benefits are across all three literacy domains (reading, mathematics, and science). However, the educational benefits of social communication in the home are mixed within each country.
	Czech Republic
	Denmark
	Sweden
	United Kingdom
	United States
Reading literacy beneficiaries	Australia	Countries where all social and cultural communication parental involvement activities benefit students in the domain of reading literacy. However, the association between social and cultural communication across the other domains of mathematics and science literacy are not consistent within countries.
	Austria
	Finland
	Ireland
	Norway
Political/social issues discussion beneficiaries	Belgium	Countries in which students benefit educationally from discussions in the home concerning political and social issues. This is the case across all three literacy domains (reading, mathematics, science). However, the association between literacy and the social communication and the other cultural communication measure of discussions concerning books, films, and TV are mixed.
	Greece
	Hungary
	Italy
	New Zealand
	Portugal
	Russian Federation
	Spain
Eating together beneficiaries	France	Countries in which students benefit educationally from eating a meal together with their parent/s. This is the case across all three literacy domains (reading, mathematics, science). However, there was no consistent within country association between cultural communication and talking with parents in terms of the three literacy domains.
Eating together beneficiaries	Germany

Discussion

The results of our study provide some evidence to suggest that parental involvement plays a role in child well-being and fosters good educational outcomes. Additionally, our results suggest that the association between parental involvement and student literacy cuts across national borders and is found across various forms of parental involvement and multiple measures of student literacy. More specifically, we observe that across the 21 countries included in our study, in general, increased social and cultural communication with parents is associated with higher levels of student literacy, net of family background characteristics. This affirms our first hypothesis. Moreover, students appear to benefit from varied and multiple types of social and cultural communication. Eating meals with parents and talking with parents about a range of issues including more serious topics such as politics and social issues, as well as less serious ones such as popular culture appear to be beneficial.

Consistent with our other hypotheses, our findings also suggest that the association varies across type of parental involvement and by country. However, not all types of parental involvement appear to be associated with higher levels of student literacy. Specifically, we observe further support for, and cross-national evidence of, the ‘reactive hypothesis’: parents provide supervision and assistance with homework in response to poor school performance. In all countries we discover that increases in the frequency with which students receive parental assistance with homework is negatively and significantly associated with student literacy. This suggests that parental involvement in schoolwork (i.e. helping with homework) may signal problems with a student’s school performance. As such, parental involvement with homework may be acting as a proxy for parental monitoring (e.g. ensuring that homework is completed and correct) in times of need.

Consistent with Bronfenbrenner’s ecological systems theory, we expected to see variation in the association between parental involvement and student literacy across the countries that we examined. Bronfenbrenner’s model suggests that children learn and develop within multiple nested systems starting from proximate systems such as family and schools and progressing to more distal systems such as the economy, government, cultural values and norms, and public policy. While we observe very little cross-national variation in the association between some types of parental involvement such as homework assistance or political discussions, the results also indicate low to moderate variation in the overall association between parental involvement and student literacy across countries, consistent with our hypothesis and Bronfenbrenner’s model. Many countries have policies at either the national or school level to encourage parental involvement so it is not surprising that many parents engage is regular interactions with their child and that, in turn, this has an impact on educational outcomes. However, it is interesting to note that there was a clustering of countries into four distinct groups. In six of the countries students appear to benefit educationally from parents who engage them in cultural communication in the home. Students in five countries benefitted in terms of reading literacy from both social and cultural communication in the home. In the third cluster, which consisted of eight countries, the salient feature was the educational benefit of discussions in the home around political and social issues. Finally, in two countries, eating a meal together benefitted students across all literacy measures. Our results suggest that variations across countries in school systems and more distal institutions that shape public policy as well as culture and values likely moderate the association between parental involvement and student literacy. Thus, while students generally appear to benefit from cultural and social communication with parents, our observations also suggest that the benefits parental involvement offer may be shaped to some degree by the educational system, public policies aimed at supporting families and schooling, the economy and social cultural norms and values. For example, it may be the case that in non-standardized school systems such as Canada, Czech Republic, Denmark, Sweden, United Kingdom, and the United States, stratification is greater.⁸ In turn, the impact of family practices and resources are more prominent in the educational achievement of students in those countries. This explanation is supported by the results of this research that indicates that for this cluster of countries cultural communication is educationally beneficial (see cultural communication beneficiary cluster). It should be noted that not all non-standardized countries were part of this cluster (i.e. Belgium). However, for reading literacy there was a clear association between cultural communication and reading literacy for the Belgium students. However, this was not a consistent finding for Belgium across the other literacy measures of mathematics and science.

In conclusion, it is important to note that many forms of parental involvement are beneficial to students and these results hold across many different countries and contexts. There is, of course, no silver bullet when it comes to education and the promotion of higher levels of educational outcomes. However, this research does emphasize that there are multiple pathways to higher literacy in terms of parent-child interaction and support. One clear message from this study of 21 countries is that parent involvement is an important aspect of a child’s education irrespective of the country in which they live. While different types of parental involvement are more salient in promoting better educational outcomes in some countries over others, there is clear evidence that regular social and cultural communication between a parent and their child reaps educational benefits.

For those individuals who are proximal to the child (i.e. parents), speaking and generally spending time with them will have a positive impact on their achievement irrespective of the cultural or country level context. It does not mean that the more distal factors do not matter, because they do. This research demonstrated cross-national variation and, as such, country level factors have a role to play in moderating the association between parental involvement and student literacy.

Significance and limitations

Our study is consistent with and builds upon past research that has examined the ways in which parental involvement affects student outcomes. In addition, it expands upon past work by taking a cross-national perspective and exploring the association between parental involvement and student literacy. This study advances previous work by Hampden-Thompson (2009) in one notable way. In previous work, the major focus was on differences in literacy by family structure. Parental involvement in Hampden-Thompson (2009) was explored as a moderator of the association between family structure and literacy achievement and not as a main effect. Furthermore, Buchmann and Dalton (2002) focused specifically on the effects of parent attitudes on students’ aspirations in 12 countries using the Third International Mathematics and Science Study (TIMSS) 1995 data. They did not consider the impact of these aspirations on student performance. However, their work is notable in that it classified the 12 countries by institutional context. As indicated above, future comparative work in the area of parental involvement and student achievement should consider country-level predictors, in a more extensive way to Park’s (2008) approach, in order to help explain further the patterns found in this study.

There are several important limitations of the study worth noting. First, the cross-sectional design of the PISA data results is the most significant limitation. As such, the results of this study are fundamentally associative in nature and cannot support causal inferences. Second, the effect sizes for the results are small. Having said this, the small effect sizes are generally consistent with prior research on the association between parental involvement and child outcomes. Third, it is important to note that because the PISA assessment tests were designed to measure everyday knowledge and cross-curricular learning rather than a student’s mastery of school-based curricula, the pattern of association observed here may not extend to studies that examine student academic achievement or school-based learning. Nonetheless, we believe that the use of student literacy as measured by PISA is an advantage for comparative studies as it is designed to produce a cross-national measure of the educational yield. Fourth, the sample for this study is students who were 15 years of age and close to the end of their compulsory education. It is unclear whether the results of this study could be generalized to younger students. Indeed, previous research found that differences in the relationship between parental involvement and child well-being vary by child’s age. It is important to recognize that systems of education are bounded by countries’ national cultures (Purves, 1987) and this may have implications for the interpretation of not only achievement cross-nationally but also the association between parental involvement measures and student literacy. Last, this study did not explore country-specific factors that may potentially moderate the association between parental involvement and student literacy. While stated as a limitation, it is also an indication of potential further research in this area. The variability in the association between parental involvement and student literacy cross-nationally is a strong indication that specific country-level factors may be salient in understanding the relationship between parental involvement and child outcomes.

Footnotes

Funding

Research for this project was made possible by the generous support of the National Institute of Child Health and Human Development through the Family and Child Well-being Research Network Grant IU01 HD37558-01.

Notes

References

Abd-El-Fattah

(2006) Effects of family background and parental involvement on Egyptian adolescents’ academic achievement and school disengagement: a structural equation modelling analysis. Social Psychology of Education 9(2): 139–157.

Alexander

Entwisle

(1988) Achievement in the First Two Years of School: Patterns and Processes (Monograph Vol. 46, No. 188). Ann Arbor, MI: Society for Research in Child Development.

Baker

Stevenson

(1986) Mothers’ strategies for children’s school achievement: Managing the transition to high school. Sociology of Education 59: 156–166.

Barnard

(2004) Parent involvement in elementary school and educational attainment. Children and Youth Services Review 26(1): 39–62.

Bronfenbrenner

(1979) The Ecology of Human Development. Cambridge, MA: Harvard University Press.

Bronfenbrenner

(1989) Ecological systems theory. In: Vasta

(ed.) Annals of Child Development: Vol. 6. Six Theories of Child Development: Revised Formulations and Current Issues. Greenwich, CT: JAI Press, pp. 187–249.

Buchmann

(2002) Measuring family background in international studies of education: Conceptual issues and methodological challenges. In: Porter

Gamoran

(eds) Methodological Advances in Cross-National Surveys of Educational Achievement. Washington, DC: National Academy Press.

Buchmann

Dalton

(2002) Interpersonal influences and educational aspirations in 12 countries: The importance of institutional context. Sociology of Education 75: 99–122.

Coleman

Campbell

Hobson

McPartland

Mood

Weinfeld

York

(1966) Equality of Educational Opportunity. Washington, DC: US Government Printing Office.

10.

Coleman

Hoffer

Kilgore

(1982) High School Achievement: Public, Catholic, and Private Schools Compared. New York: Basic Books.

11.

Epstein

(1987) Parent involvement: What research says to administrators. Education and Urban Society 19: 277–294.

12.

Epstein

(1988) Homework practices, achievement, and behaviors of elementary school students. Report no. 26, Center for Research on Elementary and Middle Schools, Johns Hopkins University.

13.

Epstein

(1991) Effects on student achievement of teachers’ practices of parental involvement. Advances in Reading/Language Research 5: 261–276.

14.

Epstein

(2001) School, Family, and Community Partnerships: Preparing Educators and Improving Schools. Boulder, CO: Westview Press.

15.

Fan

Chen

(2001) Parental involvement and students’ academic achievement: A meta-analysis. Educational Psychology Review 22: 1–22.

16.

Flouri

Buchanan

(2004) Early father’s and mother’s involvement and child’s later educational outcomes. British Journal of Educational Psychology 74: 141–153.

17.

Fryer

Levitt

(2004) Understanding the black–white test score gap in the first two years of school. The Review of Economics and Statistics 86(2): 447–464.

18.

Garnier

Stein

(1998) Values and the family: Risk and protective factors for adolescent problem behaviors. Youth and Society 30(1): 89–120.

19.

Graves

Wright

(2011) Parent involvement at school entry: A national examination of group differences and achievement. School Psychology International 32(1): 35–48.

20.

Guzman

Jekielek

(2004) Indicators of the social context of families: Family time. Office of the Assistant Secretary for Planning and Evaluation, US Department of Health and Human Services.

21.

Hampden-Thompson

(2009) Are two better than one? A comparative study of achievement gaps and family structure. Compare 39(4): 517–534.

22.

Hampden-Thompson

(2013) Family policy, family structure, and children’s educational achievement. Social Science Research 42(3): 804–817.

23.

Hampden-Thompson

Pong

S-L

(2005) Does family policy environment moderate the effect of single-parenthood on children’s academic achievement? A study of 14 European countries. Journal of Comparative Family Studies 36: 227–278.

24.

Hango

(2007) Parental investment in childhood and educational qualifications: Can greater parental involvement mediate the effects of socioeconomic disadvantage? Social Science Research 36: 1371–1390.

25.

S-C E

Willms

(1996) Effects of parental involvement on eighth-grade achievement. Sociology of Education 69: 126–141.

26.

Horn

West

(1992) A profile of parents of eight graders: statistical analysis report. NCES 92–488, National Center for Education Statistics.

27.

Hung

(2007) Family, schools and Taiwanese children’s outcomes. Educational Research 49(2): 115–125.

28.

Jencks

Smith

Acland

Bane

Cohen

Ginitis

Heyns

Michelson

(1972) Inequality: A Reassessment of the Effect of Family and Schooling in America. New York: Harper & Row.

29.

Jeynes

(2005a) Effects of parental involvement and family structure on the academic achievement of adolescents. Marriage & Family Review 37(3): 99–116.

30.

Jeynes

(2005b) A meta-analysis of the relation of parental involvement to urban elementary school student academic achievement. Urban Education 40(3): 237–269.

31.

Jeynes

(2007) The relationship between parental involvement and urban secondary school student academic achievement: A meta-analysis. Urban Education 42(1): 82–110.

32.

Keith

(1991) Parental involvement and achievement in high school. Advances in Reading/Language Research 5: 125–141.

33.

Lareau

(1987) Social class differences in family–school relationships: The importance of cultural capital. Sociology of Education 60: 73–85.

34.

Lemke

Sen

Pahlke

Miller

Williams

Kastberg

Jocelyn

(2004) International outcomes of learning in mathematics literacy and problem solving: PISA 2003. Results from the US perspective. NCES 2005–003, US Department of Education, National Center for Education Statistics, Washington, DC.

35.

Levels

Dronkers

Kraaykamp

(2008) Immigrant children’s educational achievement in western countries: Origin, destination, and community effects on mathematical performance. American Sociological Review 73(5): 835–853.

36.

McNeal

(1999) Parental involvement as social capital: Differential effectiveness on science achievement, truancy, and dropping out. Social Forces 78(1): 117–144.

37.

McNeal

(2012) Checking in or checking out? Investigating the parent involvement reactive hypothesis. The Journal of Educational Research 105(2): 79–89.

38.

Micklewright

Schnepf

(2006) Response bias in England in Pisa 2000 and 2003. Southampton Statistical Sciences Research Institute (S3RI), University of Southampton, UK.

39.

Miedel

Reynolds

(1999) Parental involvement in early intervention for disadvantaged children: Does it matter? Journal of School Psychology 24: 379–402.

40.

Milne

Myers

Rosenthal

Ginsburg

(1986) Single parents, working mothers, and the educational achievement of school children. Sociology of Education 59: 125–139.

41.

Moore

Whitney

Kinukawa

(2009) Exploring the Links between Family Strengths and Adolescent Outcomes. Child Trends Research Brief. Washington, DC: Child Trends.

42.

Muller

(1993) Parent involvement and academic achievement: An analysis of family resources available to the child. In: Schneider

Coleman

(eds) Parents, Their Children, and Schools. Boulder, CO: Westview Press, pp. 77–114.

43.

Muller

(1995) Maternal employment, parent involvement, and math achievement among adolescents. Journal of Marriage and the Family 57(1): 85–100.

44.

Mullis

Martin

Fierros

Goldberg

Stemler

(2000) Gender Differences in Achievement. Chestnut Hill, MA: Boston College.

45.

Organization for Economic Co-operation and Development (1999) Measuring Student Knowledge and Skills: A New Framework for Assessment. Paris: OECD.

46.

Organization for Economic Co-operation and Development (2000) Manual for the PISA 2000 Database. Paris: OECD.

47.

Organization for Economic Co-operation and Development (2001) Knowledge and Skills for Life: First Results from the OECD Programme for International Student Assessment. Paris: OECD.

48.

Organization for Economic Co-operation and Development (2009) Take the Test: Sample Questions from OECD’s Sample Assessments. Paris: OECD.

49.

Park

(2008) The varied educational effects of parent–child communication: A comparative study of 14 countries. Comparative Education Review 52(2): 219–243.

50.

Pong

S-L

Dronkers

Hampden-Thompson

(2003) Family policies and children’s school achievement in single- versus two-parent families. Journal of Marriage and Family 65: 681–699.

51.

Purves

(1987) IEA agenda for the future. International Review of Education 33: 104.

52.

Salazar

Schuldermann

Huynh

(2001) Canadian Filipino adolescents report on parental socialization for school involvement. Canadian Ethnic Studies 33(2): 52–77.

53.

Scaramella

Conger

Simons

Whitbeck

(1998) Predicting risk for pregnancy by late adolescence: A social contextual perspective. Developmental Psychology 34(6): 1233–1245.

54.

Schleicher

Tamassia

(2000) Measuring Student Knowledge and Skills: The PISA 2000 Assessment of Reading, Mathematical and Scientific Literacy. Education and Skills. Paris: OECD.

55.

Spera

(2005) A review of the relationship among parenting practices, parenting styles, and adolescent school achievement. Educational Psychology Review 22: 125–146.

56.

Steinberg

(1996) Beyond the Classroom: Why School Reform Has Failed and What Parents Need to Do. New York: Touchstone Books.

57.

Teachman

Paasch

Carver

(1996) Social capital and dropping out of school early. Journal of Marriage and the Family 58: 773–783.

58.

US Department of Education, National Center for Education Statistics ( 2001) Outcomes of learning: Results from the 2000 Programme for International Student Assessment of 15-year-olds in reading, math, and science literacy. NCES 2002–115, US Department of Education, Washington, DC.

59.

Westat (2000) WesVar 4.0 User’s Guide. Rockville, MD: Westat.

60.

White

(1982) The relation between socioeconomic status and academic achievement. Psychological Bulletin 913: 461–481.

61.

Young

Miller

Norton

Hill

(1995) The effect of parental supportive behaviors on life satisfaction of adolescent offspring. Journal of Marriage and the Family 57: 813–822.

62.

Zaslow

Gallagher

Hair

Egeland

Weinfield

Ogawa

Egeland

Tabors

De Temple

(2006) Longitudinal prediction of child outcomes from differing measures of parenting in a low-income sample. Developmental Psychology 42(1): 27–37.