Sage Journals: Discover world-class research

Abstract

Studies highlighting the importance of parental involvement in schooling have multiplied over the past years. However, well-validated tools assessing the different dimensions of parental involvement are lacking, especially when addressing young students’ perception of their parents’ involvement. This study offers a preliminary validation of the Student-Rated Parental School Involvement Questionnaire (SR-PSIQ); factor structure, measurement invariance, and predictive validity were assessed. Data collected on four measurement occasions from 923 French-Canadian primary school students was used. Results favored a four-factor model (parental expectations, parent–child communication, homework supervision, and school-based involvement). The SR-PSIQ was invariant across time, student gender, parental immigration status, and socioeconomic status. Regarding predictive validity, all dimensions of parental involvement were associated with later student engagement. Overall, the SR-PSIQ is a brief, valid, and reliable instrument that can easily be used by researchers or partitioners who want to understand how parents are involved in their child’s schooling.

Keywords

parental involvement factor structure measurement invariance student engagement

Introduction

Defining Parental Involvement

Parental involvement usually refers to school-related parent–child interactions and parent-school interactions (Hill et al., 2004). While it is well accepted that parental involvement is a multidimensional construct, the number of dimensions and their definition is an open question. Some studies only distinguish home-based and school-based dimensions (Chowa et al., 2013; Walker et al., 2005); parental expectations are sometimes added as a third dimension (Hill & Tyson, 2009). Most conceptualizations, however, break down home-based and school-based involvement into four to six dimensions (Epstein, 1995; Fan & Chen, 2001; Jeynes, 2005; Tardif-Grenier & Archambault, 2016).

Home-based involvement generally includes dimensions such as parental expectations, parent–child communication, and homework supervision (Castro et al., 2015; Fan & Chen, 2001; Jeynes, 2005). Parents who hold high expectations set high standards of success for their children, encourage intellectual pursuit, communicate the usefulness of school, and emphasize high academic aspirations. Parent–child communication refers to discussions that parents have with their child about their school experiences, successes, or difficulties. Parents offering high quality communication demonstrate their interest and emotional support to their child’s schooling. Lastly, homework supervision includes practices such as direct instruction, checking homework, encouragement, and organizing homework space and time.

School-based involvement refers to three types of practices that may be considered separately but are most often combined in a general dimension (Chowa et al., 2013; Fan & Chen, 2001; Epstein, 1995; Hill & Tyson, 2009). These encompass informal and spontaneous communication or contacts between parents and the school staff, more formal contacts such as participation in organized school activities or volunteering, and involvement in school decision-making and administrative processes.

Perceiving Parental Involvement

When addressing parental involvement in school, parents’ and teachers’ perspectives are important (Daniel et al., 2016). Parents are surely good informants to report on their own practices at home and at school, such as how often they contact teachers or their expectations of success for their child. However, student and parent perceptions do not always align. Parents tend to rate themselves as more involved than what their child perceives (DePlanty et al., 2007; Thomas et al., 2020). Teachers’ perspective may also be informative but remains limited as they can only form an imprecise impression of parents’ home-based involvement. For these reasons, it is critical to assess students’ perception. Grolnick and Slowiaczek (1994) even argued that students’ subjective experience of parental involvement matters more than the actual practices. For parental involvement to be influential on school outcomes, students need to perceive the support they receive at home, the implicit or explicit messages their parents send them about the importance of school, or the concrete links parents develop with the school. When compared to parent reports, student reports have the greatest contribution to students schooling experience (e.g., DePlanty et al., 2007; Ice & Hoover-Dempsey, 2011), reinforcing the idea that it is not what parents do that is most important, but what the child perceives. Thus, student reports of parental involvement represent the best way to access students’ subjective experience. Furthermore, student reports are especially important in research settings where parents are sometimes not available or able to participate, allowing to study populations that are more difficult to reach (e.g., immigrant parents, parents in disadvantaged areas) and thereby contributing to the advancement of knowledge regarding parental involvement.

Many factors may impact the way parents get involved in their child’s schooling, and ultimately, the way these practices are experienced by children. Namely, parents may be involved differently according to the family cultural background or the child’s characteristics (Antony-Newman, 2019). For instance, many immigrant parents tend to hold higher expectations of success and be more involved in schoolwork at home, while non-immigrant parents tend to be more involved in school (Turney & Kao, 2009; Villiger et al., 2014). Families with low socioeconomic status (SES) face several barriers such as lack of time and resources that may hinder parental involvement practices (Camacho-Thompson et al., 2016). Studies indicated that parents with low SES were less involved in their child’s schooling (Camacho-Thompson et al., 2016; Li et al., 2020). Still, some parental practices, such as teacher-parent communication and home-based practices may be more frequent and easier to implant for these parents (Malone, 2017; Tan, 2019).

Parental involvement may also change over the years; with younger children, parents are more directly involved in activities (e.g., reading, homework supervision), whereas with teenagers, parents tend to provide an environment that enables school success (e.g., encouragement, discussions, support; Boonk et al., 2018). Gender is also important to consider as parents may be involved differently. For instance, parents might offer more emotional support and supervision to girls, and with boys, they communicate more with teachers (Carter & Wojtkiewicz, 2000; Kristjánsson & Sigfúsdóttir, 2009). Therefore, parental involvement may translate differently among different groups and over the years.

Regardless of how and for whom parental involvement is assessed, its benefits on school achievement and adjustment are well established (Barger et al., 2019; Castro et al., 2015). Recently, studies have underlined the importance of parental involvement in student behavioral, affective, and cognitive engagement, which are important predictors of school success (Fredricks et al., 2004). Hence, when parents are involved, students tend to follow instructions better, make more efforts towards school tasks and homework (behavioral engagement), use more self-regulation strategies to organize schoolwork, environment, and time (cognitive engagement), value, like, and have more interest in school (affective engagement; Fan & Williams, 2010; Tardif-Grenier & Archambault, 2017; Wang & Sheikh-Khalil, 2014).

Measuring Parental School Involvement

To better understand how parental involvement is assessed across studies, we briefly reviewed journal articles referenced in Psychinfo since 2000 addressing parental involvement in school. The screening process identified 250 studies (see online supplement, Table S1). These studies assessed parental involvement in several countries, ethnic groups, and grade levels. Parents (43% of the identified studies), school staff (8%) and students (37%) reported parental involvement, and some studies relied on more than one respondent (11%), mostly parents and teachers. When student perception was assessed, it was mainly among high school students (73%) or mixed samples (9%); few studies relied on younger children perspective (primary level; 18%). Measures of parental involvement were very disparate, and most did not go through a rigorous validation process. Namely, less than half of the studies relied on measures that tested factor validity using factor analysis directly in the study (34%) or in another study (11%), a third only evaluated internal consistency (32%) and too many studies (18%) did not address any psychometric qualities. Very few studies estimated measurement invariance (n = 3), that is, the degree to which items are interpreted in the same way across different groups or measurement occasions. Moreover, despite the multidimensional nature of parental involvement, many studies considered a limited number of dimensions (two dimensions or less; 57%). In sum, even if many studies have addressed parental school involvement, rigorously validated measures assessing multiple dimensions of parental involvement are lacking, particularly those resting on the perception of younger children.

Study Objectives

The aim of the present study was to undertake the preliminary validation of the Student-Rated Parental School Involvement Questionnaire (SR-PSIQ). Regarding its factor structure, two (home-based and school-based involvement), three (parental expectation, home, and school involvement), four (parental expectation, parent–child communication, homework supervision, in-school involvement), and five-factor models (subdividing in-school involvement into communication and school participation) were compared. Moreover, to ensure that the measurement characteristics of the SR-PSIQ are invariant across time and groups (i.e., student gender, parental immigration status, and SES), measurement invariance was assessed across four measurement occasions spanning over a 2-year period. To better understand how the parental involvement dimensions changed in time or varied according to sociodemographic characteristics, mean differences of the invariant models are reported. Lastly, in addressing the role of perceived parental involvement on student schooling experience, the predictive validity of the SR-PSIQ was evaluated by estimating the longitudinal associations with student behavioral, cognitive, and affective engagement.

Method

Participants and Procedure

The sample consisted of 923 French-Canadian (Montreal, Quebec) students in grades 3 to 6 from five primary schools (30 classrooms) located in disadvantaged and multiethnic urban areas. These students participated in a larger project aiming to study factors associated with student engagement in school (Archambault et al., 2015). Students completed an online questionnaire in the school computer laboratory at the beginning and the end of the school year over two consecutive years (Fall 2012 to Spring 2014), resulting in four measurement occasions (T1-T4). A time sequential design was used. Hence, in the second year, new Grade 3 students were added to the sample while students who were in Grade 6 left, transitioning to secondary school (57.1% participated at all time points during the 2-year period; T1 n = 711; T2 n = 705; T3 n = 613; T4 n = 628). At T1 (year 1), students’ mean age was 9.80 years (SD = 1.28) and at T3 (year 2) 9.79 years (SD = 1.29). The sample included 52.8% of girls (T1 = 53.9%; T3 = 52.2%); 76.2% of students had immigrant parents (i.e., both parents born outside of Canada; T1 = 75.5%; T3 = 77.5%). Most of the participants lived in low-income families (<$50,000/year; 76%). This research was approved by the University Institutional Review Board. Parents gave their active consent for their child’s participation in the study.

Measures

Parental Involvement

Students reported their perception of their parents’ involvement in school using 20 items. These items were first developed based on an extensive literature review of existing tools and theoretical frameworks of parental involvement. Then, two experts of parental involvement evaluated the relevance of the proposed items. Finally, items were presented to six families to ensure readability and understanding. This was done in an interview format, for each question participants were asked to express their understanding of the item. While some items were reworded slightly, no items were removed. These items assessed four dimensions: (a) parental expectation, (b) parent–child communication, (c) homework supervision, and (d) school-based involvement using a four-point response scale (see Table 1). Higher scores indicated higher involvement.

Table 1.

Items of the Student-Rated Parental School Involvement Questionnaire.

Response scale: 1 = totally false; 2 = somewhat false; 3 = somewhat true; 4 = totally true

Parental Expectations

1. For my parents, it is important that I succeed at school

2. For my parents, it is important that I have good grades in school

3. For my parents, it is important that I go to college

4. For my parents, it is important that I am among the best in my class

5. My parents tell me that school is important. (W)

Response scale: 1 = never; 2 = rarely 3 = sometimes 4 = often

Parent–Child Communication

6. My parents talk to me about what I am learning in school

7. My parents talk to me about my grades

8. My parents talk to me about what I’m going through at school with my teacher or my friends

9. My parents talk to me about the difficulties I am experiencing at school

Homework Support

10. My parents help me when I don’t understand something in my homework

11. My parents ask me if I have done my homework. (W)

12. My parents help me plan my time and organize my homework

13. At home, my parents give me a quiet space to do my homework

14. My parents praise me when my homework is well done

School-Based Involvement

15. My parents speak with my teacher (for example, on the phone, at school daycare)

16. My parents attend school activities that I am involved in (for example sports, plays)

17. My parents go to parent meetings at school. (W)

18. My parents go to school to get my report card. (W)

19. My parents write notes to my teacher in my agenda

20. My parents participate in school committees (for example, parents’ committees)

Note. Translated items from French. Withdrawn items are denoted by (W).

Student Engagement

Students reported their behavioral (6 items, e.g. “I follow my teacher’s instructions during French/reading activities”), cognitive (6 items, e.g. “I check my math work to make sure there are no errors”), and affective engagement (two scales that were subject-related: language arts [3 items] and mathematics [3 items]; e.g. “What we learn in mathematics is interesting”) using the School Engagement Dimensions Scale (Archambault & Vandenbossche-Makombo, 2014). All items were measured on a five-point response scale, from 1 (not at all or almost never) to 5 (a lot or almost always). For each dimension of engagement, an average score was calculated; higher scores indicated higher engagement. All scales showed adequate reliability (α = .667–.800).

Grouping Variables

Students reported their gender (coded as girl = 0; boy = 1), and their parents’ immigration status (coded as immigrant = 0; non-immigrant = 1). SES was based on parents’ report of their family income on five-point responses scale (1 = less than $29,999, 2 = 30,000–49,999$, 3 = 50,000–79,999$, 4 = 80,000–119,999$, 5 = more than $120,000). To assess group invariance across levels of SES, two categories were created according to Canada’s low-income cut-off for the years of the study (2012–2014; Statistics Canada, 2022): low (<$50,000 = 1) and average/high (≥$50,000 = 2).

Statistical Analyses

Model Estimation

All measurement models were estimated using Mplus 8.5 (Muthén & Muthén, 2017). Exploratory factor analysis (EFA) was used to explore the factor structure of the SR-PSIQ. Models from 2 to 5 factors were estimated. Given the categorical nature of the items, robust weighted least square estimator (WLSMV) was used. Under WLSMV, when no covariates are used, missing data is handled by pairwise deletion. This approach performs well when data is missing completely at random (MCAR; Asparouhov & Muthén, 2010). According to Littles MCAR test, ran in IMB SPSS (Version 26), the present data was MCAR (χ² (1475) = 1499.38, p = .323). Pairwise deletion should therefore lead to unbiased parameter estimates. Theta parametrization was used, which allowed testing for invariance of unique and latent factor variances (Muthén & Asparouhov, 2002). Geomin oblique rotation was used, based on previous studies, with an epsilon value of .5¹ (Morin et al., 2013). To account for the nested structure of the data (students within classrooms), the Type = Complex option was used with classroom membership as cluster variable, correcting standard errors and chi-square tests of model fit (Muthén & Muthén, 2017).

To determine the number of factors to retain, we used parallel analysis, the recommended method (Timmerman & Lorenzo-Seva, 2011). Since parallel analysis is not available in Mplus for categorical items, the FACTOR program (Version 10.10.03; Lorenzo-Seva & Ferrando, 2006) was used applying Horn’s method with default settings. Theoretical consideration was also an important part of the decision to choose the optimal factor structure. Distribution of items into their respective factor had to be coherent with a priori theoretical expectations. To retain items, based on the traditional criteria suggested by Comrey and Lee (1992), all items had to show a standardized loading of at least .32 on their a priori target factor, while also keeping cross loadings to a minimum. To evaluate scale reliability, McDonald’s Omega coefficients were computed (McNeish, 2018).

Measurement Invariance Analysis

Measurement invariance was assessed in a six-step hierarchical model across time, gender, parental immigration status, and SES. Exploratory structural equation modeling (ESEM) was used for invariance models. ESEM has been proposed to overcome limitations regarding CFA (i.e., zero cross loadings and inflated correlations between factors) and provides all parameters used in SEM, allowing likelihood testing without going through CFA (Asparouhov & Muthén, 2009; Marsh et al., 2009). Because different groups (according to gender or parents’ immigration status) did not contain data on all values of categorical items, data was recoded on a three-point response scale, collapsing categories one and two (see Table 1). All invariance analyses were done with the recoded data.

Measurement invariance was based on Meredith’s (1993) taxonomy with specifications applying to categorical data (see Liu et al., 2017; Mplus syntax for all models are provided in the online supplement). We tested the adequacy of increasingly restrictive models: configural invariance (constraining the number of factors to equality across groups/time), weak invariance (constraining the number of factors and factor loadings to equality across groups/time), strong invariance (constraining number of factors, factor loadings, and item thresholds to equality across groups/time) and strict invariance (constraining number of factors, factor loadings, item thresholds, and item uniquenesses to equality across groups/time). In addition, variance-covariance and latent means invariance were also tested where variances, cross-sectional covariances, and means were also fixed to equality across time points and groups (Marsh et al., 2009; Morin et al., 2013). As suggested for longitudinal data, lagged uniquenesses were included in the longitudinal models (Morin et al., 2013). Latent means differences between groups (gender, immigration status, and SES), and time points are reported.

Model Fit

The overall fit of models was first evaluated using the adjusted Chi-square test of model fit for WLSMV (Muthén & Muthén, 2017). Since the chi-square test tends to be overly sensitive to sample size, model fit was also based on other recommended fit indices: Comparative Fit Index (CFI), Tucker–Lewis Index (TLI), and Root Mean Squared Error of Approximation (RMSEA; West et al., 2012). Common cut-offs for CFI and TLI (>.95), and for RMSEA (<.06), indicated good fit of models (Hu & Bentler, 1999).

Comparison of nested invariance models relied on different fit indices. First, differences in relative fit indices were used; ΔCFI ≤ −.01 and ΔRMSEA ≤.015 between models indicated that invariance hypothesis should not be rejected (Chen, 2007; Cheung & Rensvold, 2002). Difference in TLI was also considered with the same guidelines as for CFI (invariance was supported if ΔTLI ≤ −.01; Morin et al., 2013). The fit of different invariance models was also estimated using chi-square test. Under WLSMV, the difference in χ² between two nested models does not follow a χ² distribution (Muthén & Muthén, 2017); the corrected χ² difference provided in Mplus DIFFTEST was used.

Predictive Validity

Predictive validity was assessed through structural equation modeling. Using ESEM, T2 school engagement dimensions were regressed on each T1 latent parental involvement factor while controlling for the initial level of engagement (T1). Missing data was handled through full information maximum likelihood. All cross-sectional correlations between control variables at T1 and parental involvement factors were estimated.

Results

Factor Structure

The results from the parallel analysis suggested that a 4-factor model should be retained at every time point. Exploratory factor analyses from two to five factors were estimated at all time points. As indicated in Table 2, the 2-factor model did not fit data, while the results of the 3-factor models were mixed. The patterns of factor loadings in these models were hard to interpret and many items did not correspond to their respective a priori target factor. The 4-and 5-factor models presented good fit at every time point. However, in accordance with parallel analysis results, the patterns of factor loadings in the 4-factor model were the most conceptually coherent. The fifth factor was simply an incoherent breakdown of a factor (school-based involvement). Finally, after careful examination of the 4-factor models, four items were eliminated due to low factor loading or inconsistency across the different time points. Moreover, retaining these items led to convergence problems in subsequent models. Thus, the final retained 4-factor solution counted four items for each factor.

Table 2.

Goodness-of-Fit Statistics for the Exploratory Factor Analyses (T1-T4).

Data	Model	χ² (df)	CFI	TLI	RMSEA [90% CI]
T1	2-factor	269.79 (151)^***	.934	.917	.033 [.027, .040]
	3-factor	201.58 (133)^***	.962	.946	.027 [.019, .034]
	4-factor	171.15 (116)^***	.969	.950	.026 [.017, .034]
	5-factor	127.05 (100)^*	.985	.972	.020 [.006, .029]
T2	2-factor	440.96 (151)^***	.914	.892	.052 [.047, .058]
	3-factor	299.83 (133)^***	.951	.929	.042 [.036, .049]
	4-factor	199.96 (116)^***	.975	.959	.032 [.024, .039]
	5-factor	149.26 (100)^***	.985	.972	.026 [.017, .035]
T3	2-factor	362.73 (151)^***	.918	.897	.048 [.042, .054]
	3-factor	243.17 (133)^***	.957	.939	.037 [.029, .044]
	4-factor	164.62 (116)^**	.981	.969	.026 [.016, .035]
	5-factor	112.24 (100)	.995	.991	.014 [.000, .026]
T4	2-factor	362.13 (151)^***	.933	.916	.047 [.041, .053]
	3-factor	232.74 (133)^***	.968	.955	.035 [.027, .042]
	4-factor	178.13 (116)^***	.980	.968	.029 [.020, .037]
	5-factor	135.89 (100)^**	.989	.978	.024 [.012, .034]

Note. n = 710 (T1); n = 703 (T2); n = 613 (T3); n = 627 (T4).

*p < .05, **p ≤ .01, and ***p ≤ .001.

The standardized factor loadings of the final model are reported in Table 3. According to item content, the four factors represented (a) parental expectation, (b) parent–child communication, (c) homework supervision, and (d) parent school-based involvement. All items showed strong enough loading on their a priori target factor (λ = .363–.912; M = .645). Apart from one high cross loading (λ = .352) on the parent–child communication factor, cross loadings were low (λ = |.001–.234|, M = |.086|). Correlations between factors were also low to moderate (Table 4; r = .067–.440, M = .266), suggesting adequate discriminant validity between factors.

Table 3.

Standardized Factor Loadings ( λ) and Uniquenesses (δ) From the Longitudinal Invariant Model.

Items	Parental expectations	Parent–child communication	Homework supervision	School-based involvement	Uniquenesses
Items	λ (SE)	λ (SE)	λ (E)	λ (SE)	δ
1.	.892 (.023)^***	−.003 (.026)	.072 (.022)^***	−.010 (.024)	.182 (.036)^***
2.	.912 (.019)^***	−.009 (.020)	−.002 (.024)	.010 (.020)	.171 (.035)^***
3.	.624 (.035)^***	.174 (.039)^***	−.044 (.040)	−.030 (.034)	.544 (.046)^***
4.	.693 (.024)^***	.153 (.039)^***	−.169 (.051)^***	.073 (.034)^*	.462 (.028)^***
6.	.102 (.024)^***	.602 (.031)^***	.216 (.039)^***	.060 (.032)	.389 (.021)^***
7.	.231 (.038)^***	.506 (.039)^***	.160 (.031)^***	−.009 (.025)	.527 (.031)^***
8.	−.034 (.018)	.767 (.024)^***	.067 (.016)^***	.094 (.026)^***	.309 (.019)^***
9.	.006 (.018)	.742 (.030)^***	.061 (.022)^**	.028 (.021)	.386 (.033)^***
10.	−.049 (.021)^*	.091 (.031)^**	.689 (.031)^***	−.002 (.029)	.473 (.037)^***
12.	.097 (.028)^***	.081 (.024)^***	.663 (.033)^***	.167 (.026)^***	.365 (.030)^***
13.	.061 (.039)	.189 (.038)^***	.481 (.038)^***	.064 (.028)^**	.601 (.030)^***
14.	.058 (.035)	.179 (.039)^***	.565 (.041)^***	.090 (.031)^**	.488 (.037)^***
15.	.005 (.022)	.352 (.036)^***	−.165 (.039)^***	.363 (.045)^***	.708 (.033)^***
16.	−.033 (.032)	−.016 (.043)	.234 (.052)^***	.490 (.030)^***	.648 (.027)^***
19.	.011 (.033)	.214 (.043)^***	−.113 (.053)^*	.437 (.031)^***	.730 (.024)^***
20.	.001 (.012)	.004 (.009)	.035 (.013)^**	.849 (.032)^***	.257 (.052)^***

Note. Boldface values are the target items for each factor.

*p < .05, **p < .01, and ***p ≤ .001.

Table 4.

Longitudinally Invariant Standardized Factor Correlations.

Variables	1	2	3	4
1. Parental Expectations	−
2. Parent–child Communication	.251^***	−
3. Homework Supervision	.159^***	.444^***	−
4. School-Based Involvement	.067^**	.373^***	.301^***	−

Note. Standardized factor correlations from the mean invariant longitudinal model.

**p < .01, ***p < .001.

Reliability

Omega coefficients calculated from standardized longitudinal invariant loadings and uniquenesses indicated that the SR-PSIQ scales are reliable. Except for the school-based involvement scale that had a slightly lower omega coefficient (ω = .661), the parental expectation (ω = .878), parent–child communication (ω = .810), and the homework support (ω = .749) scales showed good reliability.

Measurement Invariance

Longitudinal Invariance

We first tested invariance across measurement occasions (T1 to T4). Goodness-of-fit indices are reported in Table 5. The configural invariance model (fixed number of factors) showed good fit. All fit indices also indicated good fit of subsequent models when adding equality constraints on loadings, thresholds, and uniquenesses also showed adequate. These constraints did not worsen model fit, as indicated by very small changes in CFI, TLI, and RMSEA as well as non-significant DIFFTESTs. Therefore, strict longitudinal invariance of the measurement model of the SR-PSIQ was supported.

Table 5.

Goodness-of-Fit Statistics for Longitudinal and Group Invariance Models.

Invariance models	χ² (df)	CFI	TLI	RMSEA [90% CI]	Δχ² (Δdf) DIFFTEST	ΔCFI	ΔTLI	ΔRMSEA
Longitudinal invariance (n = 920)
Configural	1672.47 (1592)	.987	.983	.007 [.000, .011]
Weak	1816.68 (1736)	.987	.985	.007 [.000, .011]	163.12 (144)	.000	.002	.000
Strong	1853.01 (1772)	.987	.985	.007 [.000, .011]	42.07 (36)	.000	.000	.000
Strict	1903.67 (1820)	.986	.985	.007 [.000, .011]	60.78 (48)	−.001	.000	.000
Var-covar	1947.58 (1852)	.985	.983	.007 [.000, .011]	59.71 (32)^**	−.001	−.002	.000
Latent means	1965.03 (1864)^*	.984	.982	.008 [.000, .011]	82.36 (44)^***	−.001	−.001	.001
Gender invariance - T4 (n = 627; 328 girls)
Configural	148.13 (124)	.993	.986	.025 [.000, .039]
Weak	187.04 (172)	.995	.994	.017 [.000, .031]	49.15 (48)	.002	.008	−.008
Strong	196.13 (184)	.996	.995	.015 [.000, .029]	9.71 (12)	.001	.001	−.002
Strict	212.83 (200)	.996	.995	.014 [.000, .029]	18.72 (16)	.000	.000	−.001
Var-covar	214.16 (210)	.999	.999	.008 [.000, .026]	7.78 (10)	.003	.004	−.006
Latent means	224.73 (214)	.997	.996	.013 [.000, .027]	7.77 (4)	−.002	−.003	.005
Parental immigration status invariance -T4 (n = 627; 484 immigrants)
Configural	147.81 (124)	.994	.988	.025 [.000, .039]
Weak	195.43 (172)	.994	.991	.021 [.000, .034]	56.14 (48)	.000	.003	−.004
Strong	208.56 (184)	.993	.991	.021 [.000, .033]	16.18 (12)	−.001	.000	.000
Strict	220.58 (200)	.994	.993	.018 [.000, .031]	14.65 (16)	.001	.002	−.003
Var-covar	230.06 (210)	.995	.994	.017 [.000, .030]	12.37 (10)	.001	.001	−.001
Latent means	271.30 (214)^**	.985	.983	.029 [.017, .039]	23.80 (4)^***	−.010	−.011	.012
SES invariance - T4 (n = 362; 274 low SES)
Configural	150.07 (124)	.985	.971	.034 [.000, .052]
Weak	191.13 (172)	.989	.985	.025 [.000, .043]	50.83 (48)	.004	.014	−.009
Strong	205.80 (184)	.987	.984	.026 [.000, .043]	18.99 (12)	−.002	−.001	.001
Strict	226.56 (200)	.985	.982	.027 [.000, .043]	26.98 (16)^*	−.002	−.002	.001
Var-covar	256.20 (210)^*	.973	.969	.035 [.016, .049]	27.15 (10)^**	−.012	−.013	.008
Latent means^a	249.28 (204)^*	.974	.969	.035 [.016, .049]	19.75 (4)^***	−.011	−.013	.008

Notes. CFI = comparative fit index; TLI = Tucker–Lewis index; RMSEA = root mean square error of approximation; CI = confidence interval; Var-covar = variance-covariance. Δχ², ΔCFI, ΔTLI, and ΔRMSEA are relative to the preceding invariant model.

^aVariances and covariances were not constrained to equality, fit is compared to strict invariance model.

*p ≤ .05, **p < .01, and ***p ≤ .001.

In the next steps, the variance-covariance and mean invariance models were tested adding equality constraints on the latent factors cross-sectional correlations, variances and means.² All indices indicated good fit of models. Although DIFFTESTs were significant, according to the very small changes in CFI, TLI and RMSEA, the added constraints did not worsen model fit.

We interpreted mean differences between time points at the variance-covariance step (Table 6). For identification purposes, means were fixed to 0 for a referent time point (T4). Mean differences mainly concerned the parent–child communication factor; compared to T4, other time points showed higher means. Other differences were also seen on the homework supervision factor at T1 and T3, and the parental expectation factor at T1, these means were significantly higher than T4 means.

Table 6.

Standardized Latent Factor Mean Differences Between Measurement Occasions and Sociodemographic Characteristics.

Latent factors/Invariance models	T1	T2	T3	T4^a
Latent factors/Invariance models	M_diff (SE)	M_diff (SE)	M_diff (SE)	M_diff (SE)
P. expectations
Longitudinal	0.132 (0.062)^*	0.021 (0.050)	−0.039 (0.039)	—
Gender	0.327 (0.114)^**	0.414 (0.099)^***	0.042 (0.114)	0.225 (0.106)^*
Imm. Status	0.074 (0.192)	−0.909 (0.092)^***	−0.783 (0.124)^***	−0.681 (0.131)^***
SES	—	−0.303 (0.172)	−0.575 (0.202)^**	−0.765 (0.198)^***
P-C comm.
Longitudinal	0.146 (0.054)^**	0.095 (0.046)^*	0.163 (0.030)^***	—
Gender	−0.004 (0.099)	0.110 (0.100)	0.077 (0.092)	0.023 (0.107)
Imm. Status	−0.581 (0.127)^***	0.052 (0.095)	−0.006 (0.139)	−0.127 (0.138)
SES	—	0.049 (0.141)	−0.123 (0.160)	0.028 (0.119)
Homework sup.
Longitudinal	0.155 (0.076)^*	−0.016 (0.062)	0.094 (0.046)^*	—
Gender	−0.207 (0.104)^*	−0.064 (0.092)	−0.106 (0.118)	−0.086 (0.109)
Imm. Status	−0.203 (0.109)	0.150 (0.142)	0.041 (0.126)	0.194 (0.163)
SES	—	−0.248 (0.154)	0.114 (0.185)	0.079 (0.153)
School-based inv.
Longitudinal	0.023 (0.083)	0.061 (0.077)	0.051 (0.050)	—
Gender	0.256 (0.186)	0.400 (0.139)^**	0.099 (0.118)	0.264 (0.100)^**
Imm. Status	0.152 (0.124)	0.184 (0.163)	0.042 (0.156)	0.345 (0.133)^*
SES	—	0.157 (0.173)	−0.023 (0.192)	1.049 (0.488)*

Notes. Standardized mean differences of the invariant models (variance-covariance model). Latent means are fixed to zero for identification purpose for the referent measurement occasion, T4 in longitudinal invariance models, and the referent group for group invariance models: Girls = 0; Immigrant parents = 0; low SES = 0. Because both categories of SES did not contain all values at T1, mean differences at T1 were not computed for SES. P = parent; Imm. = immigration status; SES = socioeconomical status; C = child; comm. = communication; sup. = supervision; inv. = involvement.

^aT4 mean differences for the SES invariances models are from the strict invariance model.

*p < .05, **p < .01, and ***p ≤ .001.

Group Invariance

Gender

Measurement invariance across genders was tested at every time point. For brevity, only T4 results are reported in Table 5 (see online supplement Table S2 for full results). All models showed good fit and the added equality constraints did not lead to a significant decrease in model fit. Therefore, factor loadings, thresholds, uniquenesses, latent variances, covariances, and means were invariant across genders for the SR-PSIQ. Except at T1, for which uniqueness invariance was not fully attained, the invariance patterns were similar at other time points.

Despite the apparent invariance of means, some differences between genders can be pointed out. Results presented in Table 6 indicate that differences between boys and girls for the SR-PSIQ scales mainly concerned the parental expectation and the school-based involvement factors; boys perceived higher parental expectations at most time points and higher school-based involvement at T2 and T4. Girls also perceived more homework supervision at T1.

Parental Immigration Status

We next tested if the measurement characteristics were also invariant according to parents’ immigration status. The invariance patterns were very similar at all time points. T4 fit indices are presented in Table 5 (see online supplement Table S3 for full results). The configural invariance model showed good fit. According to all fit indices and DIFFTESTs, there was no significant change from configural through variance-covariance models in all fit indices. Therefore, the structure, loadings, thresholds, uniquenesses, and latent variances and covariances appeared invariant across parents’ immigrant status.

When constraining means to equality, changes in fit indices and significant DIFFTESTs indicated that mean invariance could not be assumed. Latent mean differences according to parents’ immigration status are reported in Table 6. Significant differences concerned the parental expectation factor at T2, T3 and T4 for which immigrant parents were perceived as maintaining higher expectations than non-immigrant parents. At T1, parent–child communication was rated higher by children of immigrant parents. Students also reported higher parent school-based participation among non-immigrant parents at T4.

Family SES

We also tested if the SR-PSIQ measurement properties were invariant according to family SES. T4 fit indices are reported in Table 5 (see online supplement Table S3 for full results). Configural through strict invariance models showed good fit and the added equality constrains (loadings, intercepts, uniquenesses) did not worsen the fit. Adding equality constraints on variances and covariances significantly worsened fit according to DIFFTEST as well as decreases in CFI and TLI slightly above the suggested guidelines, indicating that variances and covariances were not fully invariant across SES. Therefore, mean invariance model was compared to the strict invariance model. Constraining means to equality also decreased model fit suggesting that T4 means were not invariant across SES. According to invariance models at other time points, variances, covariances and means are invariant at T2 and T3.

Latent mean differences are reported in Table 6. Students in low SES families perceived their parents as having higher school expectations than students from average/high SES at T3 and T4. At T4, students from higher SES families perceived their parents as having higher levels of school-based involvement.

Predictive Validity

To assess predictive validity, all dimensions of student engagement (T2) were regressed on the four scales of the SR-PSIQ (T1), while controlling for engagement at T1. The standardized regression coefficients between latent factors and student engagement variables are reported in Table 7. Parent–child communication and homework supervision were positively associated with all dimensions of engagement, even though some associations were marginal. Parental expectations were only linked marginally to behavioral engagement and affective engagement in language arts. School-based involvement was associated significantly or marginally with lower behavioral, cognitive, and affective-language arts-engagement.

Table 7.

Standardized Regression Coefficients Between T1 ESEM Latent Factors and T2 School Engagement Dimensions.

T1 Predictors	Behavioral (T2)	Cognitive (T2)	Affective-Math (T2)	Affective-Language Arts (T2)
T1 Predictors	β (SE)	β (SE)	β (SE)	β (SE)
Behavioral Engagement	.281 (.054)^***	.064 (.060)	−.132 (.068)^†	−.127 (.070)^†
Cognitive Engagement	.030 (.035)	.322 (.046)^***	−.026 (.032)	.042 (.042)
Affective-Math	.024 (.034)	−.032 (.028)	.545 (.029)^***	−.021 (.028)
Affective-Language Arts	.071 (.043)^†	.014 (.046)	.003 (.038)	.457 (.041)^***
Parental Expectations	.086 (.048)^†	.061 (.044)	.037 (.039)	.061 (.032)^†
P-C Communication	.168 (.091)^†	.222 (.079)^**	.167 (.085)^*	.159 (.090)^†
Homework Supervision	.161 (.073)^*	.118 (.060)^*	.147 (.057)^**	.178 (.071)^*
School-Based Involvement	−.291 (.085)^***	−.221 (.089)^*	−.149 (.090)^†	−.141 (.096)

Note. n = 740.

†p < .10, *p < .05, **p < .01, and ***p ≤ .001.

Discussion

This study is one of the rare studies that tested factorial validity, measurement invariance, and predictive validity of students’ perception of parental involvement among a sample of young elementary school children. Our results support the multidimensionality of parental involvement. Hence, specific dimensions of parental involvement should be assessed separately, rather than general school involvement. While parental school-based involvement is measured through a single factor, home-based involvement includes three dimensions: parental expectations, parent–child communication, and homework supervision. Measuring different aspects of parental involvement, these dimensions show good discriminant validity. Furthermore, the measurement characteristics of the SR-PSIQ are invariant across time and among sociodemographic characteristics; strict invariance is met in all models. This allows us to compare means between groups and time knowing that these differences are not biased by measurement artifacts. Mean invariance is met for gender and time, that is, even if slight differences may arise when comparing specific means, means are relatively stable across time and gender. As for SES and immigration status, mean invariance did not hold suggesting that means differ significantly between groups.

According to mean differences, in line with other studies (e.g., Villiger et al., 2014), results show that parents who were born in a foreign country, maintain higher expectation for their child’s schooling. Immigrant parents also tend to communicate more with their child but are less school-based involved than non-immigrant parents. These results support the idea that immigrant parents may encounter certain barriers regarding their participation in school. As such, it might be easier for them to be involved in their child schooling at home (Camacho-Thompson et al., 2016). In contrast to many studies (e.g., Li et al., 2020), our study further indicates that parents of low SES are perceived by their children as having higher expectation for them regarding school success and attainment, compared to parents from average or high SES. This is potentially explained by the fact that in many low SES communities, children with an immigrant background have highly educated, but underemployed parents (Adversario, 2021). As in our study, these parents also have higher expectations for their children, which is often not reflected by their current SES.

This study also shows that the SR-PSIQ has adequate predictive validity as most scales are related to later school engagement. Students report higher levels of engagement when their parents supervise their homework and communicate more with them, behaviorally, cognitively, and affectively. However, parental expectations are only associated with later behavioral and affective engagement, and to a lesser extent. This result is somewhat surprising as this dimension of parental involvement has been positioned as one of the most important to academic success (Castro et al., 2015). However, practices such as communicating or helping with homework are more concrete and likely to change during the school year than having high expectations. Considering the short time span between measurement occasions, direct actions may be more likely to bring about changes in the level of engagement, especially in a sample of primary students.

Contrary to what might be expected, school-based involvement is associated with lower student engagement. While most studies evidence positive links between school-based participation and school outcomes, our results agree with the few studies that indicate a negative link (Barger et al., 2019; Boonk et al., 2018). Different reasons can explain this result. For instance, if students perceive these contacts as parent or teacher control, more frequent contact with school may hinder student engagement (Vasquez et al., 2016). It is also possible that these contacts with school are more frequent among students with decreasing levels of engagement, that is, those who presented more behavior problems or school difficulties (Archambault & Dupéré, 2017). Evidence shows that teachers are more prone to contact parents when the student is having difficulty than when he is functioning well in class, especially among boys and immigrant students (Zimmermann & Keynton, 2021). Further studies could examine these hypotheses.

Limitations and Future Research

While the SR-PSIQ seems to be a valid instrument that can be used among different populations, the present sample came from a highly disadvantaged urban area. Even if factorial structure withholds, other differences could occur among populations with higher SES. Namely, the negative link between school-based involvement and engagement may be an artifact of parents’ low SES. These parents might be more numerous to have experienced school more negatively and, therefore, to perceive family-school interactions as less welcoming. Future studies could confirm this hypothesis using the SR-PISQ among different populations.

Given that the SR-PSIQ reflects students’ perception of parental involvement, certain aspects are not taken into consideration. For instance, parents’ feeling of welcome at school and school’s openness to parental involvement were not assessed as these aspects of involvement may be difficult for a child to judge, particularly at the primary level. Nevertheless, these dimensions remain important to fully understand parental involvement, and well developed and validated tools assessing parents or teachers points of view are available if these aspects are of one’s interest (e.g., Dawson & Wymbs, 2016; Tardif-Grenier & Archambault, 2016).

Finally, predictive validity in this study has been limited to school engagement. Future studies could examine the predictive value of the SR-PSIQ on other indicators of school success. For instance, it would be relevant to examine the instrument’s capacity to predict achievement, a well-known outcome of parental involvement, or other aspects of school adjustment that have been less examined, such as teacher-student interactions or externalized behaviors.

Supplemental Material

Supplemental Material - Validation of the Student-Rated Parental School Involvement Questionnaire: Factorial Validity and Invariance Across Time and Sociodemographic Characteristics

Supplemental Material for Validation of the Student-Rated Parental School Involvement Questionnaire: Factorial Validity and Invariance Across Time and Sociodemographic Characteristics by Julie Goulet, Isabelle Archambault, Julien Morizot, Elizabeth Olivier and Kristel Tardif-Grenier in Journal of Psychoeducational Assessment

Footnotes

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) disclosed receipt of the following financial support for the research,authorship,and/or publication of this article: This research received funding from the Fonds de Recherche Québécois -Société et Culture (FRQ-SC) from the grant id (2012-RP-145548,269573) to the second author and the doctoral fellowships from FRQ-SC to the first author.

Ethical Approval

All procedures performed in this study were approved by the Institutional Review Board of the University of Montreal as well as all the school boards ethical committees.

Informed Consent

Active informed consent was obtained from all individual participants included in the study. Parents also gave their active consent for their child to participate.

ORCID iD

Julie Goulet

Supplemental Material

Supplemental material for this article is available online.

Notes

References

Adversario

(2021). Adult immigrant journeys: Occupational downgrading and continuously evolving selves. Adult Learning, 32(1), 5–16. https://doi.org/10.1177/1045159520952562

Antony-Newman

(2019). Parental involvement of immigrant parents: A meta-synthesis. Educational Review, 71(3), 362–381. https://doi.org/10.1080/00131911.2017.1423278

Archambault

Dupéré

(2017). Joint trajectories of behavioral, affective, and cognitive engagement in elementary school. The Journal of Educational Research, 110(2), 188–198. https://doi.org/10.1080/00220671.2015.1060931

Archambault

Tardif-Grenier

Dupéré

Janosz

Mc Andrew

Pagani

Parent

Oliver

Richer

M.-J.

Kurdi

(2015). Étude comparative de l’engagement scolaire des élèves de milieux défavorisés issus ou non de l’immigration: Contributions de l’environnement scolaire et des pratiques enseignantes[comparative study of school engagement among immigrant and non-immigrant students from disadvantaged backgrounds: Contributions of the school environment and teaching practices]. Rapport soumis au Fonds de Recherche Québécois société et culture. Fonds de recherche du Québec - Société et culture. https://frq.gouv.qc.ca/app/uploads/2021/08/prs_archambaulti_rapport_eleves-immigration-ou-non.pdf

Archambault

Vandenbossche-Makombo

(2014). Validation of the scale of the dimensions of school engagement among primary school students. Canadian Journal of Behavioural Science, 46(2), 275–288. http://doi.org/10.1037/a0031951

Asparouhov

Muthén

(2009). Exploratory structural equation modeling. Structural Equation Modeling: A Multidisciplinary Journal, 16(3), 397–438. https://doi.org/10.1080/10705510903008204

Asparouhov

Muthén

(2010). Weighted least squares estimation with missing data. Technical Report. http://www.statmodel.com/download/GstrucMissingRevision.pdf

Barger

M. M.

Kim

E. M.

Kuncel

N. R.

Pomerantz

E. M.

(2019). The relation between parents’ involvement in children’s schooling and children’s adjustment: A meta-analysis. Psychological Bulletin, 145(9), 855–890. https://doi.org/10.1037/bul0000201

Boonk

Gijselaers

H. J. M.

Ritzen

Brand-Gruwel

(2018). A review of the relationship between parental involvement indicators and academic achievement. Educational Research Review, 24(June), 10–30. https://doi.org/10.1016/j.edurev.2018.02.001

10.

Camacho-Thompson

D. E.

Gillen-O’Neel

Gonzales

N. A.

Fuligni

A. J.

(2016). Financial strain, major family life events, and parental academic involvement during adolescence. Journal of Youth and Adolescence, 45(6), 1065–1074. https://doi.org/10.1007/s10964-016-0443-0

11.

Carter

Wojtkiewicz

(2000). Parental involvement with adolescents’ education: Do daughters or sons get more help? Adolescence, 35(137), 29–44.

12.

Castro

Expósito-Casas

López-Martín

Lizasoain

Navarro-Asencio

Gaviria

J. L.

(2015). Parental involvement on student academic achievement: A meta-analysis. Educational Research Review, 14(February), 33–46. https://doi.org/10.1016/j.edurev.2015.01.002

13.

Chen

F. F.

(2007). Sensitivity of goodness of fit indexes to lack of measurement invariance. Structural Equation Modeling: A Multidisciplinary Journal, 14(3), 464–504. https://doi.org/10.1080/10705510701301834

14.

Cheung

G. W.

Rensvold

R. B.

(2002). Evaluating goodness-of-fit indexes for testing measurement invariance. Structural Equation Modeling: A Multidisciplinary Journal, 9(2), 233–255. https://doi.org/10.1207/S15328007SEM0902_5

15.

Chowa

G. A. N.

Masa

R. D.

Tucker

(2013). The effects of parental involvement on academic performance of Ghanaian youth: Testing measurement and relationships using structural equation modeling. Children and Youth Services Review, 35(12), 2020–2030. https://doi.org/10.1016/j.childyouth.2013.09.009

16.

Comrey

A. L.

Lee

H. B.

(1992). A first course in factor analysis (2nd ed.). Psychology Press. https://doi.org/10.4324/9781315827506

17.

Daniel

G. R.

Wang

Berthelsen

(2016). Early school-based parent involvement, children’s self-regulated learning and academic achievement: An Australian longitudinal study. Early Childhood Research Quarterly, 36(3rd Quarter), 168–177. https://doi.org/10.1016/j.ecresq.2015.12.016

18.

Dawson

A. E.

Wymbs

B. T.

(2016). Validity and utility of the parent — teacher relationship scale—II. Journal of Psychoeducational Assessment, 34(8), 751–764. https://doi.org/10.1177/0734282915627027

19.

DePlanty

Coulter-Kern

Duchane

K. A.

(2007). Perceptions of parent involvement in academic achievement. The Journal of Educational Research, 100(6), 361–368. https://doi.org/10.3200/JOER.100.6.361-368

20.

Epstein

J. L.

(1995). School/family/community partnerships: Caring for the children we share. The phi delta kappan, 76(9), 701–712. https://doi.org/10.1177/003172171009200326

21.

Fan

Williams

C. M.

(2010). The effects of parental involvement on students’ academic self-efficacy, engagement and intrinsic motivation. Educational Psychology, 30(1), 53–74. https://doi.org/10.1080/01443410903353302

22.

Fan

Chen

(2001). Parental involvement and students’ academic achievement: A meta-analysis. Educational Psychology Review, 13(1), 1–22. https://doi.org/10.1023/A:1009048817385

23.

Fredricks

J. A.

Blumenfeld

P. C.

Paris

A. H.

(2004). School engagement: Potential of the concept, state of the evidence. Review of Educational Research, 74(1), 59–109. https://doi.org/10.3102/00346543074001059

24.

Grolnick

W. S.

Slowiaczek

M. L.

(1994). Parents’ involvement in children’s schooling: A multidimensional conceptualization and motivational model. Child Development, 65(1), 237–252. https://doi:10.2307/1131378

25.

Hill

N. E.

Castellino

D. R.

Lansford

J. E.

Nowlin

Dodge

K. A.

Bates

J. E.

Pettit

G. S.

(2004). Parent academic involvement as related to school behavior, achievement, and aspirations: Demographic variations across adolescence. Child Development, 75(5), 1491–1509. https://doi:10.1111/j.1467-8624.2004.00753.x

26.

Hill

N. E.

Tyson

D. F.

(2009). Parental involvement in middle school: A meta-analytic assessment of the strategies that promote achievement. Developmental Psychology, 45(3), 740–763. https://doi.org/10.1037/a0015362

27.

L. T.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55. https://doi.org/10.1080/10705519909540118

28.

Ice

C. L.

Hoover-Dempsey

K. V.

(2011). Linking parental motivations for involvement and student proximal achievement outcomes in homeschooling and public schooling settings. Education and Urban Society, 43(3), 339–369. https://doi.org/10.1177/0013124510380418

29.

Jeynes

W. H.

(2005). A meta-analysis of the relation of parental involvement to urban elementary school student academic achievement. Urban Education, 40(3), 237–269. https://doi.org/10.1177/0042085905274540

30.

Kristjánsson

Á. L.

Sigfúsdóttir

I. D.

(2009). The role of parental support, parental monitoring, and time spent with parents in adolescent academic achievement in Iceland: A structural model of gender differences. Scandinavian Journal of Educational Research, 53(5), 481–496. https://doi.org/10.1080/00313830903180786

31.

Yang

Wang

Jia

(2020). Family socioeconomic status and home-based parental involvement: A mediation analysis of parental attitudes and expectations. Children and Youth Services Review, 116 (September), 105111. https://doi.org/10.1016/j.childyouth.2020.105111

32.

Liu

Millsap

R. E.

West

S. G.

Tein

J.-Y.

Tanaka

Grimm

K. J.

(2017). Testing measurement invariance in longitudinal data with ordered-categorical measures. Psychological Methods, 22(3), 486–506. https://doi.org/10.1037/met0000075

33.

Lorenzo-Seva

Ferrando

P. J.

(2006). Factor: A computer program to fit the exploratory factor analysis model. Behavior Research Methods, 38(1), 88–91. https://doi.org/10.3758/bf03192753

34.

Malone

(2017). Socioeconomic Status: A Potential Challenge for Parental Involvement in Schools. The Delta Kappa Gamma Bulletin International Journal for Professional Educators, 83(3), 58–62.

35.

Marsh

H. W.

Muthén

Asparouhov

Lüdtke

Robitzsch

Morin

A. J. S.

Trautwein

(2009). Exploratory structural equation modeling, integrating CFA and EFA: Application to students’ evaluations of university teaching. Structural Equation Modeling: A Multidisciplinary Journal, 16(3), 439–476. https://doi.org/10.1080/10705510903008220

36.

McNeish

(2018). Thanks coefficient alpha, we’ll take it from here. Psychological Methods, 23(3), 412–433. https://doi.org/10.1037/met0000144

37.

Meredith

(1993). Measurement invariance, factor analysis and factorial invariance. Psychometrika, 58(4), 525–543. https://doi.org/10.1007/BF02294825

38.

Morin

A. J. S.

Marsh

H. W.

Nagengast

(2013). Exploratory structural equation modeling. In Hancock

G. R.

Mueller

R. O.

(Eds.). Structural equation modeling: A second course (2nd ed.) (pp. 395–436).Information Age.

39.

Muthén

Asparouhov

(2002). Latent variable analysis with categorical outcomes: Multiple-group and growth modeling in Mplus. Technical Report Mplus Web Notes: No. 4. http://www.statmodel.com/download/webnotes/CatMGLong.pdf

40.

Muthén

L. K.

Muthén

B. O

. (2017). Mplus user’s guide, 8th ed. Muthén and Muthén.

41.

Statistics Canada. (2022). Table 11-10-0241-01: Low income cut-offs (LICOs) before and after tax by community size and family size, in current dollars. Government of Canada. https://doi.org/10.25318/1110024101-eng

42.

Tan

C. Y.

(2019). Involvement practices, socioeconomic status, and student science achievement: Insights from a typology of home and school involvement patterns. American Educational Research Journal, 56(3), 899–924. https://doi.org/10.3102/0002831218807146

43.

Tardif-Grenier

Archambault

(2016). Validation of the questionnaire sur l’implication parentale dans le suivi scolaire (QIPSS) with a sample of elementary student’s parents living in multiethnic and disadvantaged areas. European Review of Applied Psychology, 66(3), 139–150. https://doi.org/10.1016/j.erap.2016.04.007

44.

Tardif-Grenier

Archambault

(2017). Implication parentale et réussite éducative: Différences selon la région de naissance du parent [parental involvement and educational success: Differences according to parent’s region of birth]. Revue des Sciences de l’Éducation, 43(1), 209–247. https://doi.org/10.7202/1042079ar

45.

Thomas

Muls

De Backer

Lombaerts

(2020). Middle school student and parent perceptions of parental involvement: Unravelling the associations with school achievement and wellbeing. Educational Studies, 46(4), 404–421. https://doi.org/10.1080/03055698.2019.1590182

46.

Timmerman

M. E.

Lorenzo-Seva

(2011). Dimensionality assessment of ordered polytomous items with parallel analysis. Psychological Methods, 16(2), 209–220. https://doi.org/10.1037/a0023353

47.

Turney

Kao

(2009). Barriers to school involvement: Are immigrant parents disadvantaged? The Journal of Educational Research, 102(4), 257–271. https://doi.org/10.3200/JOER.102.4.257-271

48.

Vasquez

A. C.

Patall

E. A.

Fong

C. J.

Corrigan

A. S.

Pine

(2016). Parent autonomy support, academic achievement, and psychosocial functioning: A meta-analysis of research. Educational Psychology Review, 28(3), 605–644. https://doi.org/10.1007/s10648-015-9329-z

49.

Villiger

Wandeler

Niggli

(2014). Explaining differences in reading motivation between immigrant and native students: The role of parental involvement. International Journal of Educational Research, 64, 12–25. https://doi.org/10.1016/j.ijer.2013.10.004

50.

Walker

J. M. T.

Wilkins

A. S.

Dallaire

J. R.

Sandler

H. M.

Hoover-Dempsey

K. V.

(2005). Parental involvement: Model revision through scale development. The Elementary School Journal, 106(2), 85–104. https://doi.org/10.1086/499193

51.

Wang

M. T.

Sheikh-Khalil

(2014). Does parental involvement matter for student achievement and mental health in high school? Child Development, 85(2), 610–625. https://doi.org/10.1111/cdev.12153

52.

West

S. G.

Taylor

A. B.

(2012). Model fit and model selection in structural equation modeling. In Hoyle

R. H.

(Ed.). Handbook of structural equation modeling (pp. 209–231). The Guilford Press.

53.

Zimmermann

C. R.

Keynton

(2021). When to call home? The intersection of race and gender and teacher communication with parents in first grade. Race Ethnicity and Education, 1–20. https://doi.org/10.1080/13613324.2021.1890567

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.79 MB

0.00 MB