Measurement Invariance of the Regulation of Emotions in Parenting Scale (REPS): Psychometric Validation Across Ethnoracial Groups

Abstract

Parental emotion regulation is crucial for parent–child interactions and child psychological outcomes. However, limited research has examined whether the Regulation of Emotions in Parenting Scale (REPS) functions equivalently across ethnoracial groups, raising concerns about measurement bias. This study evaluated the psychometric properties of the REPS, including measurement invariance, reliability, and differential validity, across an ethnically diverse sample of n = 1,408 parents. Using multigroup confirmatory factor analysis, we tested configural, metric, and scalar invariance. While configural and metric invariance were supported, full scalar invariance was not. A partial scalar invariance model, allowing three item intercepts to vary, showed acceptable fit. McDonald’s omega coefficients indicated strong internal reliability across all subscales and racial groups. Multiple regression analyses tested differential validity and found no significant interaction effects, supporting consistent predictive validity. These findings confirm the REPS as a reliable tool for diverse populations. Future research should explore REPS applicability across sociocultural contexts.

Keywords

parental emotion regulation measurement invariance racial and ethnic differences REPS psychometric validation

Introduction

Parenting is a fulfilling yet complex experience that brings both moments of joy and significant challenges, requiring parents to navigate and manage a broad spectrum of emotions (Dix, 1991). Parental emotion regulation plays a fundamental role in shaping parent–child interactions and, consequently, child development (Rutherford et al., 2015). The way parents manage their emotions in the context of parenting can influence how children develop emotional regulation skills, cope with stress, and respond to social challenges (Morris et al., 2017).

Despite the significant role of parental emotion regulation in child development, research on this topic has often focused on emotion regulation as a broad psychological process rather than one uniquely shaped by the specific challenges of parenting. While emotion regulation is crucial across various aspects of life, the caregiving context presents distinct demands that require parents to manage their own emotions while also guiding their child’s emotional development (Rodriguez & Shaffer, 2021; Rutherford et al., 2015). General emotion regulation abilities may not automatically translate to effective emotion regulation in parenting situations, particularly in high-stress interactions with children. Similarly, difficulties with emotion regulation in daily life do not always indicate problematic parenting behaviors or negative child outcomes. Some parents who struggle with general emotion regulation may still successfully regulate their emotions within parent–child interactions, providing warmth and emotional support. The unique and interactive nature of emotion regulation in parenting highlights the need for assessment tools that specifically capture how parents navigate emotional regulation within the caregiving context. Understanding emotion regulation in parenting requires a framework that accounts for how parenting may influence emotion regulation, rather than relying solely on generalized emotion regulation constructs. Notably, research indicates that certain emotion regulation strategies, such as rumination, may function differently in parenting contexts compared to general emotional experiences. For instance, recent evidence suggests that parenting-specific rumination might have paradoxical implications; while generally considered maladaptive, rumination in the context of parenting has been associated not only with increased parental distress but also with heightened awareness and intention to improve future parent–child interactions (Rodriguez et al., 2025). Such findings highlight that the emotional dynamics specific to parenting interactions could alter the typically negative interpretations of rumination, underscoring the need for context-specific examination of emotion regulation constructs. Therefore, unlike general ER, parenting-specific emotion regulation involves simultaneously managing one’s emotions when interacting with a child, potentially within emotionally charged and culturally dependent interactions. This emotional responsibility likely requires parents to employ distinct regulation strategies that may differ fundamentally from those used in non-parental contexts.

Previous research highlights considerable ethnoracial variation in emotion regulation processes, underscoring the necessity of considering culture when examining these constructs (Weiss et al., 2022). A recent systematic review found that 71% of studies investigating ethnoracial differences in emotion regulation identified meaningful group differences; however, the majority of these studies assessed emotion regulation broadly, without explicitly considering how context-specific demands, such as parenting, may differently shape emotion regulation across diverse racial and ethnic groups (Weiss et al., 2022). Moreover, observed ethnoracial differences in emotion regulation scores can result from actual variation in emotion regulatory processes or from measurement bias stemming from culturally influenced interpretations of questionnaire items or response styles (Rodriguez et al., 2023). Without testing for measurement invariance—which statistically assesses whether a measure functions equivalently across groups—it remains unclear whether these ethnoracial differences reflect true psychological differences or artifacts due to culturally specific responding (Meredith, 1993; Putnick & Bornstein, 2016). Given that parenting contexts involve culturally embedded norms, expectations, and emotional practices, assessing measurement invariance in parenting-specific emotion regulation measures like the Regulation of Emotions in Parenting Scale (REPS) is particularly essential. Establishing measurement invariance ensures that researchers and clinicians interpret group differences accurately, thereby avoiding erroneous conclusions that could unintentionally reinforce cultural stereotypes or lead to inequitable recommendations for intervention and policy. Cultural contexts profoundly shape emotional expression, regulation strategies, and parenting behaviors (Weiss et al., 2022; Rodriguez et al., 2025). For instance, cultural values such as familismo, respect, or collectivism may influence how emotions are managed within family interactions, potentially leading to systematic differences in emotion regulation across ethnoracial groups. Therefore, culturally responsive measures and psychometric validations, such as measurement invariance testing, become imperative. Outside of interpretability, there have been calls for increasing ethnoracial representation in parenting research (Rodriguez et al., 2023) but doing so requires responsive methods as well as methodological rigor. Without establishing measurement invariance, parenting interventions based on flawed group comparisons risk exacerbating disparities rather than alleviating them.

The primary objective of this study was to evaluate the psychometric properties of the REPS in a large, ethnically diverse sample. Specifically, we conducted a multigroup confirmatory factor analysis (CFA) to assess the measurement invariance of the REPS across racial and ethnic groups. Measurement invariance testing followed a stepwise approach, beginning with configural invariance, which examines whether the same factor structure holds across groups (Meredith, 1993). Establishing configural invariance ensures that all groups conceptualize the underlying constructs in a similar manner. Next, we tested metric invariance, which assesses whether factor loadings are equivalent across groups. If metric invariance is supported, it suggests that the associations between individual items and their underlying latent constructs remain consistent across different racial and ethnic identities. Finally, we tested for scalar invariance, which examines whether item intercepts are equivalent across groups. Scalar invariance is critical for ensuring that group differences in observed scores reflect true differences in latent constructs rather than systematic response biases. Given that full scalar invariance was not supported, we examined partial scalar invariance, allowing for minor adjustments to item intercepts to achieve an acceptable model fit.

Beyond measurement invariance, we also assessed the internal reliability of the REPS subscales using McDonald’s omega, which provides a more robust estimate of reliability compared with Cronbach’s alpha (McDonald, 2013; Revelle & Zinbarg, 2009). Strong reliability coefficients across all racial and ethnic groups would further validate the consistency of the REPS as a measurement tool. In addition, we examined differential validity by investigating whether the predictive associations between REPS subscales and child psychopathology outcomes varied by ethnoracial identity. Specifically, we conducted multiple regression analyses with interaction terms between the REPS subscales and ethnoracial identity to test whether the strength of associations between parental emotion regulation strategies and child psychological outcomes differed across groups. Establishing differential validity is critical for ensuring that the REPS functions similarly across diverse populations and that it accurately captures parental emotion regulation processes regardless of racial or ethnic background (Aguinis & Gottfredson, 2010). Evaluating differential validity ensures that the REPS can accurately capture parental emotion regulation strategies’ associations with child outcomes universally rather than only within specific ethnoracial contexts. Confirming that these predictive relationships hold across diverse groups strengthens the measure’s utility for both researchers and clinicians working with ethnoracially diverse populations.

By integrating these analyses, we aimed to provide a comprehensive psychometric evaluation of the REPS in a diverse population. Establishing measurement invariance, assessing reliability, and evaluating differential validity are crucial steps in ensuring that the REPS is a valid and reliable tool for studying parental emotion regulation across different racial and ethnic groups. The findings have important implications for both research and clinical practice, as they determine whether the REPS can be used confidently to compare parental emotion regulation strategies across diverse populations and assess their impact on child psychological well-being.

Method

Ethical Approval

Ethical approval was obtained from the University of Illinois Urbana-Champaign Institutional Review Board.

Participant Eligibility

Eligible parents for the interviews and survey were at least 18 years of age, parenting at least one child aged 3 to 17, and living with their target child at least 50% of the time. For the purposes of the study, if the parent reported having more than one child, the parent was asked to enter the name, age, and sex of each child. If the parent had more than one child, they were instructed to choose the one with the most recent birthday.

Recruitment

The survey was distributed by QualtricsXM, a data collection and management company. Available participants in Qualtrics panels are nationally representative samples of adults who volunteer to participate in surveys online in exchange for monetary compensation. Qualtrics contacts their prescreened participant panels to assess eligibility and emails the survey to participants to complete via a unique link, then disburses compensation as a participation incentive. Recruitment ran from January to February 2024. In total, N = 1,509 eligible respondents completed the survey. For quality purposes, only eligible participants who correctly and appropriately answered attention and quality checks embedded throughout the survey were included in the final dataset. Specifically, the survey was piloted in n = 150 participants initially and an average was taken of their survey completion time. Participants whose survey completion time was faster than 1.5 standard deviations from the pilot average were excluded.

Measures

Sociodemographic Characteristics

Participants were asked to complete a sociodemographic questionnaire about their age, gender, race, ethnicity, educational attainment, relationship status, employment status, and household income. Parents were also asked about the age of their child.

Emotion Regulation in Parenting

Participants completed the REPS (Rodriguez & Shaffer, 2021), an 18-item self-report measure assessing parents’ use of specific emotion regulation strategies during interactions with their child. Parents rate the frequency of their responses using a 5-point Likert-type scale ranging from 1 (never) to 5 (always). The REPS encompasses three distinct subscales: Adaptive emotion regulation strategies (10 items), reflecting effective emotional awareness and management (e.g., “When I’m upset with my child, I put the situation into perspective before I react”); Suppression (4 items), indicating the tendency to minimize or avoid emotional experiences after negative interactions (e.g., “When I lose my temper with my child, I don’t think about it much after it’s over”); and Rumination (4 items), capturing repetitive negative thinking related to parenting interactions (e.g., “I think about my mistakes or failures as a parent”). Thus, in subsequent analyses, the REPS was modeled as a three-factor structure with 10 Adaptive Strategies items, 4 Suppression items, and 4 Rumination items, specified as correlated latent variables. In its initial validation study, the REPS demonstrated strong psychometric properties, including measurement invariance by parent sex, confirming its utility across mothers and fathers (Rodriguez & Shaffer, 2021).

Child Symptoms of Psychopathology

Parents completed the Strengths and Difficulties Questionnaire (SDQ; Goodman, 1997), a 25-item behavioral screening questionnaire that assesses conduct problems, hyperactivity-inattention, emotional symptoms, peer problems, and prosocial behavior. The SDQ uses a 3-point Likert-type scale ranging from 0 (not true) to 2 (certainly true) to describe how much an item describes the target child’s behavior. For parents of target children ages 3 to 4, reliability was acceptable in this sample (ω = .79). For parents of target children ages 5 to 11 and 12 to 17, reliability was acceptable in both groups (ω = .83).

Data Analysis

Preliminary Analyses

First, we computed descriptive statistics to summarize the sample’s demographic characteristics. We then conducted preliminary diagnostics for normality, missing data, and outliers to ensure preliminary assumptions were met.

Measurement Invariance

We evaluated measurement invariance of the three-factor model of the REPS across ethnoracial groups using multigroup CFA. We compared the three-factor structure by ethnoracial identity to assess whether the scale functions similarly across these groups. In line with best practices for ordinal data, we initially attempted to estimate the model using the Weighted Least Squares (WLSMV) estimator with items specified as ordered. However, this approach yielded convergence warnings and identification issues, likely due to the high parameterization required for ordinal estimation in combination with the relatively small sample size of the Asian parent subgroup (n ≈ 190).

Despite this limitation, we retained Asian participants in the analysis, given their persistent underrepresentation in parenting research and the importance of evaluating measurement equivalence across diverse racial and ethnic groups. To balance estimation feasibility with inclusivity, we proceeded with ML estimation. This decision is supported by prior simulation work showing that ML performs adequately for ordinal indicators with five or more categories and sufficient sample sizes (Rhemtulla et al., 2012).

Measurement invariance was evaluated by sequentially constraining parameters (configural, metric, scalar) and examining changes in model fit in the lavaan package (Rosseel, 2012). Model fit was assessed using χ² difference tests, comparative fit index (CFI), Tucker–Lewis index (TLI), root mean square error of approximation (RMSEA), and standardized root mean square residual (SRMR; Hu & Bentler, 1999). We followed the standard stepwise approach (Meredith, 1993), testing configural, metric, scalar, and partial scalar invariance models. The model specified three correlated latent constructs: Adaptive Strategies (10 items), Suppression (4 items), and Rumination (4 items), consistent with the theoretical structure of the REPS (Figure 1).

Reliability

We then computed McDonald’s omega reliability coefficient to assess the reliability of the REPS subscales across ethnoracial groups (McDonald, 2013).

Differential Validity

Next, to examine differential validity, we used multiple regression analyses to test whether the associations between each REPS subscale and child psychopathology outcomes varied by parent ethnoracial identity. Specifically, we fit a series of multiple regression models for each SDQ outcome (emotional symptoms, conduct problems, hyperactivity, peer problems, prosocial behavior, and total difficulties), using REPS latent factor scores as predictors. These scores were extracted from the final partial scalar invariance model to account for minor measurement non-invariance. Each model included three main predictors (latent Adaptive Strategies, Suppression, and Rumination), three categorical ethnoracial group indicators (Asian, Black, Hispanic; White was the reference group), and the interaction terms between each REPS subscale and each ethnoracial group.

To test for differential validity, models were fit using lm() function to fit a linear regression model, which follows the standard ordinary least squares method. To account for violations of homoscedasticity without altering the estimated coefficient, we calculated heteroscedasticity-robust standard errors using the Huber-White (HC1) robust variance estimator. Given that our data did not meet normality assumptions, we calculated heteroscedasticity-robust standard errors from the “sandwich” package (version 3.1-1; Zeileis, 2004; Zeileis et al., 2020). The robust standard errors were applied in hypothesis testing using the coeftest() function from the “lmtest” package (Zeileis & Hothorn, 2002), which provides corrected t-values and p-values.

Interaction terms were included between each REPS subscale (latent factor scores) and parent ethnoracial identity to examine whether associations between parental emotion regulation and child psychopathology outcomes varied across ethnoracial groups. The ethnoracial group included White parents as the reference group. To balance parsimony and comprehensiveness of results, we included several interactions per outcome (e.g., emotional symptoms and conduct problems) within a single model. For each SDQ outcome, we fit an omnibus regression model that included the main effects of the REPS subscales, main effects of race/ethnicity, and all interaction terms. This approach allowed us to comprehensively test for moderation while limiting the number of models and reducing the risk of Type I error. No interaction terms reached statistical significance after applying false discovery rate (FDR) corrections, indicating no evidence of differential predictive validity across groups. As such, no post hoc pairwise comparisons were conducted.

All analyses were conducted using RStudio 2023.12.1.

Results

Sociodemographic Characteristics

Sociodemographic characteristics are provided in Table 1. Participants included n = 1,408 parents (M_age = 39.74, SD = 9.70), with 60.8% identifying as female and 39.2% as male. The sample was 33.3% White, 27.7% Black, 25.5% Hispanic, and 13.5% Asian. Most parents were married (55.7%), while 22.7% had never married, 11.9% were in a committed relationship, 8.5% were separated or divorced, and 1.4% were widowed. In terms of education, 17.3% held a postgraduate degree, 25.9% had a bachelor’s degree, 13.6% had an associate’s degree, 15.8% had some college experience, 25.4% completed high school or a General Education Development (GED) program, and 2.1% had not completed high school. Regarding employment, 65.9% worked full-time, 13.3% worked part-time, and 20.7% were not employed outside the home. Household income varied, with 46.9% earning below $70,000% and 19.8% earning over $110,000 annually. Financial difficulties were reported by 18.0% of participants as severe, 27.1% as moderate, 26.9% as minor, and 28.1% reported no difficulty. The mean age of the target child was 10.60 years (SD = 4.37).

Table 1.

Sociodemographic Characteristics (N = 1,408).

Variable	N (%)/M (SD)
Parents
Age	39.74 (9.70)
Sex
Female	856 (60.80%)
Male	552 (39.20%)
Ethnoracial identity
White	469 (33.31%)
Asian	190 (13.49%)
Black	390 (27.70%)
Hispanic	359 (25.50%)
Relationship status
Never married	319 (22.66%)
In a committed relationship	167 (11.86%)
Separated or divorced	119 (8.45%)
Married	784 (55.68%)
Widowed	19 (1.35%)
Education
Postgraduate degree	244 (17.33%)
Bachelor’s degree	364 (25.85%)
Associate’s degree	192 (13.64%)
Partial college	222 (15.77%)
High school or GED	357 (25.36%)
Partial high school	29 (2.06%)
Employment
Not employed outside the home	292 (20.74%)
Work part-time	187 (13.28%)
Work full-time	929 (65.98%)
Household yearly income
Less than $10,000	86 (6.1%)
$10,000–$30,000	176 (12.5%)
$30,000–$50,000	228 (16.2%)
$50,000–$70,000	280 (19.9%)
$70,000–$90,000	210 (14.9%)
$90,000–$110,000	149 (10.6%)
Over $110,000	279 (19.8%)
Financial difficulty
A great deal of difficulty	253 (18.0%)
Some difficulty	382 (27.1%)
Only a little difficulty	378 (26.9%)
No difficulty at all	395 (28.1%)
Target child
Age	10.60 (4.37)

Note. Values represent the number of participants (N) and corresponding percentages (%) for categorical variables, and means (M) with standard deviations (SD) for continuous variables.

Configural Invariance

As summarized in Table 2, we first tested the configural three-factor model (Adaptive Strategies, Suppression, and Rumination), which allowed all parameters to vary freely across groups while maintaining the same factor structure. The configural three-factor model, with Adaptive Strategies, Suppression, and Rumination specified as correlated latent constructs, fit the data well, χ²(572) = 1,103.4, CFI = 0.951, TLI = 0.949, RMSEA = 0.048, and SRMR = 0.059. These results suggest that the basic factor structure of the REPS is consistent across racial groups.

Table 2.

Model Fit Indices for Measurement Invariance Across Racial Groups.

Model	χ² (df)	Δχ² (Δdf)	p	RMSEA	CFI	TLI	AIC	BIC
Configural Invariance	988.3 (508)			0.045	0.952	0.95	66,240	67,542
Metric Invariance	1,048.0 (553)	59.67 (45)	.07	0.03	0.949	0.947	66,210	67,276
Scalar Invariance	1,138.6 (598)	29.23 (9)	< .0016	0.079	0.94	0.935	66,211	67,040
Partial Scalar Invariance	1,109.4 (589)	61.43 (36)	.005	0.045	0.948	0.946	66,199	67,076

Note. χ² = chi-square; df = degrees of freedom; Δχ² = chi-square difference test; Δdf = change in degrees of freedom; RMSEA = root mean square error of approximation; CFI = comparative fit index; TLI = Tucker–Lewis index; AIC = Akaike information criterion; BIC = Bayesian information criterion. The configural model allowed all parameters to vary freely across groups, while the metric model constrained factor loadings to be equal. The scalar model further constrained item intercepts, resulting in a significant decline in model fit. The partial scalar model freed the intercepts of three items (REPS15, REPS17, and REPS18) to improve fit. Partial scalar invariance was supported.

To support this, we present standardized factor loadings and item intercepts by group for the configural model in Supplemental Tables 1 and 5, respectively. While some variation in magnitude is evident across groups (e.g., REPS02 factor loading ranges from 0.53 to 0.75), all items loaded moderately to strongly onto their intended factors.

Metric Invariance

Next, we tested metric invariance by constraining factor loadings to be equal across groups. The model fit was acceptable, χ²(589) = 1,120.5, with CFI = 0.949, TLI = 0.947, RMSEA = 0.049, and SRMR = 0.060. A chi-square difference test indicated that the constraint did not significantly worsen model fit compared with the configural model, Δχ²(17) = 17.1, p = .450, and the change in CFI (ΔCFI = 0.002) was within the acceptable threshold (ΔCFI < 0.01; Cheung & Rensvold, 2002). These results support metric invariance, indicating that the factor loadings are equivalent across racial groups.

Inspection of Supplemental Tables 1 (configural) and 2 (metric) shows minimal changes in factor loadings across models. For instance, REPS08 shows stable loadings for Asian parents (0.70 vs. 0.69) and Hispanic parents (0.64 vs. 0.67), supporting statistical invariance findings. Intercepts remain freely estimated in this model (see Supplemental Table 6).

Scalar Invariance

We then tested scalar invariance by constraining item intercepts to be equal across groups in addition to the previously constrained factor loadings. The model fit worsened compared with the metric model, χ²(598) = 1,138.6, CFI = 0.945, TLI = 0.943, RMSEA = 0.051, and SRMR = 0.063. A chi-square difference test indicated a significant decrease in model fit, Δχ²(9) = 29.2, p = .0006, and ΔCFI = 0.004, suggesting that full scalar invariance did not hold.

As shown in Supplemental Table 7, constraining all intercepts to be equal across groups eliminated the variability observed in the configural and metric models (e.g., REPS15 fixed at 3.71 despite group-specific estimates ranging from 3.68 to 3.92 in Supplemental Table 5). These uniform constraints likely contributed to the misfit observed.

Partial Scalar Invariance

Given the significant decrement in fit, we examined modification indices to identify items contributing to misfit. Based on modification indices above 10, we freed the intercepts of three items (REPS15, REPS17, and REPS18) while keeping all other intercepts constrained. As shown in Supplemental Table 8, intercepts for these three items once again vary across groups, reflecting patterns in the configural model (Supplemental Table 5). For example, REPS15 ranges from 3.63 (Black parents) to 3.75 (White parents), whereas all were constrained to 3.71 in the scalar model. These differences highlight the value of partial invariance testing. Specifically, parents identifying as Black (Group 2), Hispanic (Group 3), and White (Group 6) reported systematically different levels of agreement on these guilt and self-reflection-related items compared with the reference group. For instance, Black parents endorsed significantly lower intercepts on items such as “After losing my temper around my child, I feel guilty” (REPS15) and “No matter how much I do, I feel guilty about not doing enough for my child” (REPS17), whereas White parents showed slightly higher agreement levels on similar items. As a result, these differences precluded achieving full scalar invariance and necessitated the adoption of a partial scalar invariance model to facilitate valid cross-group latent mean comparisons.

The revised partial scalar model showed improved fit, χ²(589) = 1,109.4, CFI = 0.948, TLI = 0.946, RMSEA = 0.050, and SRMR = 0.060. A chi-square difference test comparing the full scalar and partial scalar models showed a significant improvement in fit, Δχ²(9) = 29.2, p = .0006, suggesting that relaxing constraints on these three items allowed the model to achieve acceptable scalar invariance. These results indicate that partial scalar invariance is supported, meaning that latent mean comparisons across racial groups are still valid while accounting for minor non-invariance in the intercepts of REPS15 (“After losing my temper around my child, I feel guilty”), REPS17 (“No matter how much I do, I feel guilty about not doing enough for my child”), and REPS18 (“I think about my interactions with my child, wishing it had gone better”).

Together, Supplemental Tables 1–8 provide a full accounting of model parameters (factor loadings and intercepts) across configural, metric, scalar, and partial scalar models.

Correlations Among Latent Factors

To further evaluate the construct validity of the REPS, we examined correlations among the three latent factors (Supplemental Table 9). Across groups, Adaptive Strategies was negatively correlated with Suppression (r range = –.20 to –.14), although this association was nonsignificant for White parents (r = .06). In contrast, Adaptive Strategies was consistently positively correlated with Rumination (r range = .29 to .36). Finally, Suppression and Rumination were weakly to moderately positively correlated (r range = .10 to .18), with the correlation reaching significance in three of the four groups. These inter-factor correlations support the theoretical distinctiveness of the three subscales while also highlighting their expected overlap.

Magnitude of Measurement Non-Invariance

To quantify the practical impact of noninvariance, we calculated fit index changes between the metric and scalar models. The change in comparative fit index (ΔCFI = .0045), root mean square error of approximation (ΔRMSEA = .0003), and standardized root mean square residual (ΔSRMR = .0013) were all well below recommended thresholds for meaningful noninvariance (e.g., ΔCFI > .01; ΔRMSEA > .015; ΔSRMR > .015; Cheung & Rensvold, 2002), indicating that the deviation from full scalar invariance was negligible.

To further evaluate the magnitude of measurement bias associated with non-invariant items, we computed dMACS values (Nye & Drasgow, 2011) based on group-specific intercepts, loadings, and residual variances from the partial scalar invariance model. Most comparisons yielded dMACS values below 0.20, indicating trivial to small bias (see Supplemental Table 10). However, two comparisons slightly exceeded that threshold: REPS17 for White versus Hispanic parents (dMACS = —0.412) and REPS18 for White versus Black parents (dMACS = —0.330), suggesting moderate item-level bias in these isolated cases. These findings further underscore the value of using a partial scalar invariance approach.

Reliability

The REPS subscales showed acceptable to excellent McDonald’s omega reliability coefficient across Adaptive Strategies, Suppression, and Rumination, demonstrating robust internal reliability for all ethnoracial groups of parents, as indicated in Table 3.

Table 3.

McDonald’s Omega Coefficients by Group.

Group	Adaptive strategies	Suppression	Rumination
Black	0.88	0.84	0.72
Hispanic	0.88	0.85	0.83
White	0.89	0.89	0.74
Asian	0.91	0.83	0.83

Note. McDonald’s omega was calculated for three latent factors within each group.

Differential Validity

Using latent factor scores from the final partial scalar invariance model to account for minor measurement noninvariance, we tested whether the predictive associations between REPS subscales and child psychopathology varied by parent ethnoracial identity. Across all models, there was no evidence of differential validity: none of the interaction terms between REPS subscales and ethnoracial identity were statistically significant after correcting for multiple comparisons. This suggests that the relationship between parenting-specific emotion regulation and child outcomes was consistent across racial and ethnic groups. Detailed model results are presented in Table 4.

Table 4.

Regression Results Predicting Child Psychopathology From REPS Latent Factors and Race Interactions.

Predictor	B (Unstandardized)	SE	t	p	FDR p
Emotional Symptoms
(Intercept)	2.67	0.11	24.27	<.001	<.001
Asian	−0.89	0.18	−4.83	<.001	<.001
Black	−0.22	0.17	−1.33	.184	.448
Hispanic	0.10	0.17	0.56	.578	.870
Latent Adaptive Strategies	−0.60	0.14	−4.44	<.001	<.001
Latent Suppression	1.03	0.14	7.46	<.001	<.001
Latent Rumination	0.55	0.13	4.43	<.001	<.001
Asian × Latent Adaptive Strategies	−0.14	0.22	−0.64	.525	.844
Black × Latent Adaptive Strategies	−0.13	0.20	−0.62	.534	.844
Hispanic × Latent Adaptive Strategies	−0.18	0.21	−0.86	.389	.747
Asian × Latent Suppression	−0.51	0.27	−1.93	.054	.172
Black × Latent Suppression	−0.21	0.20	−1.03	.305	.660
Hispanic × Latent Suppression	−0.46	0.21	−2.17	.031	.105
Asian × Latent Rumination	−0.05	0.23	−0.20	.843	.949
Black × Latent Rumination	0.13	0.19	0.72	.473	.840
Hispanic × Latent Rumination	0.12	0.19	0.66	.510	.844
Conduct
(Intercept)	2.30	0.10	22.73	<.001	<.001
Asian	−0.95	0.15	−6.27	<.001	<.001
Black	−0.18	0.15	−1.19	.233	.547
Hispanic	−0.08	0.15	−0.53	.595	.879
Latent Adaptive Strategies	−0.67	0.12	−5.51	<.001	<.001
Latent Suppression	0.89	0.12	7.41	<.001	<.001
Latent Rumination	0.45	0.12	3.74	<.001	.001
Asian × Latent Adaptive Strategies	0.10	0.16	0.65	.517	.844
Black × Latent Adaptive Strategies	0.09	0.18	0.50	.616	.896
Hispanic × Latent Adaptive Strategies	−0.04	0.18	−0.23	.818	.949
Asian × Latent Suppression	−0.43	0.20	−2.19	.029	.101
Black × Latent Suppression	−0.26	0.19	−1.37	.171	.443
Hispanic × Latent Suppression	−0.28	0.18	−1.52	.129	.353
Asian × Latent Rumination	−0.18	0.17	−1.04	.297	.660
Black × Latent Rumination	−0.04	0.18	−0.20	.842	.949
Hispanic × Latent Rumination	−0.09	0.17	−0.55	.580	.870
Hyperactivity
(Intercept)	3.93	0.11	35.56	<.001	<.001
Asian	−0.73	0.21	−3.38	.001	.003
Black	−0.36	0.16	−2.25	.025	.092
Hispanic	−0.12	0.17	−0.70	.483	.844
Latent Adaptive Strategies	−0.84	0.14	−6.07	<.001	<.001
Latent Suppression	0.04	0.12	0.30	.768	.949
Latent Rumination	0.74	0.16	4.70	<.001	<.001
Asian × Latent Adaptive Strategies	0.19	0.24	0.78	.437	.806
Black × Latent Adaptive Strategies	0.14	0.19	0.73	.467	.840
Hispanic × Latent Adaptive Strategies	0.20	0.21	0.94	.345	.705
Asian × Latent Suppression	0.17	0.28	0.62	.537	.844
Black × Latent Suppression	0.12	0.18	0.67	.504	.844
Hispanic × Latent Suppression	0.01	0.19	0.03	.975	.989
Asian × Latent Rumination	−0.10	0.28	−0.34	.732	.925
Black × Latent Rumination	−0.08	0.21	−0.36	.719	.920
Hispanic × Latent Rumination	−0.05	0.22	−0.22	.823	.949
Peer problems
(Intercept)	2.63	0.09	28.85	<.001	<.001
Asian	−0.34	0.15	−2.28	.023	.088
Black	−0.12	0.13	−0.91	.363	.726
Hispanic	−0.01	0.14	−0.07	.941	.987
Latent Adaptive Strategies	−0.52	0.11	−4.61	<.001	<.001
Latent Suppression	0.67	0.11	6.26	<.001	<.001
Latent Rumination	0.24	0.12	2.09	.037	.121
Asian × Latent Adaptive Strategies	−0.01	0.18	−0.03	.979	.989
Black × Latent Adaptive Strategies	−0.03	0.15	−0.22	.830	.949
Hispanic × Latent Adaptive Strategies	−0.13	0.16	−0.78	.434	.806
Asian × Latent Suppression	−0.33	0.20	−1.67	.095	.284
Black × Latent Suppression	−0.03	0.15	−0.17	.867	.957
Hispanic × Latent Suppression	−0.21	0.16	−1.35	.176	.445
Asian × Latent Rumination	0.17	0.19	0.89	.372	.728
Black × Latent Rumination	0.02	0.16	0.15	.883	.957
Hispanic × Latent Rumination	0.06	0.16	0.38	.702	.911
Prosocial
(Intercept)	7.72	0.09	86.31	<.001	<.001
Asian	0.07	0.18	0.39	.695	.911
Black	−0.05	0.14	−0.39	.697	.911
Hispanic	−0.02	0.14	−0.11	.915	.976
Latent Adaptive Strategies	0.87	0.12	6.98	<.001	<.001
Latent Suppression	−0.01	0.09	−0.06	.954	.987
Latent Rumination	−0.12	0.11	−1.02	.307	.660
Asian × Latent Adaptive Strategies	−0.09	0.21	−0.44	.662	.911
Black × Latent Adaptive Strategies	−0.04	0.19	−0.19	.850	.949
Hispanic × Latent Adaptive Strategies	0.09	0.19	0.44	.657	.911
Asian × Latent Suppression	−0.33	0.22	−1.48	.140	.373
Black × Latent Suppression	−0.07	0.15	−0.46	.646	.911
Hispanic × Latent Suppression	−0.15	0.14	−1.02	.309	.660
Asian × Latent Rumination	0.05	0.21	0.25	.800	.949
Black × Latent Rumination	0.21	0.16	1.32	.187	.448
Hispanic × Latent Rumination	0.16	0.16	0.98	.326	.680
Total psychopathology
(Intercept)	11.49	0.33	34.89	<.001	<.001
Asian	−2.88	0.55	−5.28	<.001	<.001
Black	−0.91	0.49	−1.88	.061	.188
Hispanic	−0.13	0.49	−0.27	.788	.949
Latent Adaptive Strategies	−2.65	0.40	−6.59	<.001	<.001
Latent Suppression	2.62	0.40	6.49	<.001	<.001
Latent Rumination	2.03	0.41	4.97	<.001	<.001
Asian × Latent Adaptive Strategies	0.25	0.63	0.39	.696	.911
Black × Latent Adaptive Strategies	0.09	0.60	0.14	.887	.957
Hispanic × Latent Adaptive Strategies	−0.11	0.61	−0.19	.850	.949
Asian × Latent Suppression	−1.19	0.74	−1.61	.108	.315
Black × Latent Suppression	−0.36	0.60	−0.59	.552	.855
Hispanic × Latent Suppression	−0.94	0.61	−1.53	.126	.353
Asian × Latent Rumination	−0.26	0.66	−0.39	.698	.911
Black × Latent Rumination	−0.01	0.58	−0.01	.989	.989
Hispanic × Latent Rumination	−0.03	0.58	−0.06	.956	.987

Note. B = unstandardized regression coefficient. SE = standard error. FDR p = p-value adjusted for multiple comparisons using the false discovery rate (FDR) correction. Reference group is White.

Discussion

In this study, we examined the measurement invariance, reliability, and differential validity of the REPS in an ethnoracially diverse national sample of parents. The findings of this study provide robust evidence for the reliability and applicability of the REPS across diverse ethnoracial groups. Our measurement invariance analysis indicates that the REPS maintains a consistent factor structure across groups, supporting its use in racially and ethnically diverse populations. Establishing configural invariance suggests that all racial and ethnic groups conceptualize the REPS constructs similarly, providing a solid foundation for cross-group comparisons (Meredith, 1993). Furthermore, achieving metric invariance suggests that the associations between observed items and their respective latent factors remain stable across groups, indicating that parental emotion regulation strategies, as measured by the REPS, are interpreted consistently across different racial and ethnic identities (Putnick & Bornstein, 2016).

However, the absence of full scalar invariance, evidenced by a significant decline in model fit when item intercepts were constrained across groups, suggests that parents from different racial and ethnic backgrounds may systematically differ in their responses to certain REPS items (Cheung & Rensvold, 2002). This finding highlights the complexity of cross-cultural comparisons in psychological research, particularly in the domain of parenting and emotion regulation. This variation could stem from cultural differences in emotional expression, emotion regulation strategies, or parenting norms (Matsumoto et al., 2008; Rodriguez et al., 2025; Weiss et al., 2022). Specifically, three items related to rumination—REPS15 (“After losing my temper around my child, I feel guilty”), REPS17 (“No matter how much I do, I feel guilty about not doing enough for my child”), and REPS18 (“I think about my interactions with my child, wishing it had gone better”)—exhibited measurement non-invariance, meaning that parents from different groups interpreted these rumination items differently, even when their underlying levels of emotion regulation were similar. These differences suggest that guilt-related items may be particularly susceptible to cultural variation, possibly due to distinct socialization processes surrounding parental responsibility and self-reflection.

Black parents reported significantly lower intercepts (agreement levels) for guilt-related items such as REPS15 (“After losing my temper around my child, I feel guilty”) and REPS17 (“No matter how much I do, I feel guilty about not doing enough for my child”). Ethnoracial differences in self-reported guilt or self-reflective emotions likely reflect culturally distinct norms around parental responsibility and emotional expression (Markus & Kitayama, 1991; Mesquita et al., 2017). For instance, collectivist-oriented cultures may approach parental guilt differently due to stronger emphasis on relational interdependence and family harmony, whereas individualistic cultures may interpret guilt within the context of personal accountability or self-oriented standards. Our specific findings align with studies showing that African American parenting practices often emphasize emotional restraint, possibly reflecting a protective strategy to shield children from external racial biases or negative stereotypes (Dunbar et al., 2015; Nelson et al., 2013). The comparatively lower endorsement of guilt may reflect cultural norms emphasizing strength, resilience, and a different conceptualization of parental responsibility (Dunbar et al., 2015). Hispanic parents exhibited moderate levels of endorsement, higher than Black parents but slightly different compared with White parents. This pattern is consistent with cultural values of familism and respeto, which emphasize parental responsibility and collective family well-being but may also lead to nuanced expressions of guilt. Latino cultures often balance emotional expressiveness with hierarchical family dynamics, resulting in moderated emotional displays (Calzada et al., 2012; Halgunseth et al., 2006). White parents had slightly higher intercepts, especially on guilt and self-reflection-related items. This result corresponds with findings that individualistic Western cultures often emphasize personal accountability and internal standards, increasing the likelihood of self-reported guilt when parenting interactions do not align with personal ideals or expectations (Markus & Kitayama, 1991; Matsumoto et al., 2008; Mesquita et al., 2017). The systematic variability in item intercepts across groups is consistent with cross-cultural theories indicating that collectivist-oriented groups (e.g., many Asian and Hispanic cultures) conceptualize guilt in relational and interdependent terms, whereas individualist groups (predominantly White American cultures) interpret guilt through personal standards of accountability (Markus & Kitayama, 1991; Matsumoto et al., 2008; Mesquita et al., 2017).

To address this, we relaxed constraints on these items, achieving partial scalar invariance (Putnick & Bornstein, 2016). This adjustment ensures that comparisons of latent means across groups remain valid, while accounting for minor noninvariance in these items. To assess whether this non-invariance was practically meaningful, we examined changes in global model fit between the metric and scalar models. All differences were well below conventional thresholds for concern (ΔCFI = .0045, ΔRMSEA = .0003, ΔSRMR = .0013; Cheung & Rensvold, 2002), suggesting that the partial scalar invariance solution introduced minimal bias. The need for partial scalar invariance underscores the importance of considering cultural aspects of emotion regulation in the context of parenting, as these factors may significantly influence both parent and child behaviors (Gratz & Roemer, 2004). Future research should further investigate how culturally specific parenting beliefs and practices shape responses to self-reported measures, as well as explore alternative ways to capture emotion regulation constructs that minimize cultural bias. We also evaluated the magnitude of item-level bias by calculating dMACS values (Nye & Drasgow, 2011) for each freed item across ethnoracial groups. Most comparisons indicated trivial to small bias (dMACS < 0.20), with only two instances exceeding this threshold: REPS17 for White versus Hispanic parents (dMACS = —0.412) and REPS18 for White vs. Black parents (dMACS = —0.330). These moderate group differences further underscore the cultural salience of guilt-related items and justify their treatment under a partial invariance framework.

The observed partial scalar noninvariance for guilt and reflection-oriented rumination items (REPS15, REPS17, REPS18) is particularly informative, as it aligns with recent studies highlighting the complex association between rumination in the parenting context and parenting behaviors (Rodriguez et al., 2025). Specifically, while rumination is typically conceptualized negatively, parents’ endorsement of these items may reflect both self-critical patterns and healthy introspective processes aimed at improving parenting interactions. Cultural factors likely shape these patterns: Black parents’ relatively lower endorsement might reflect cultural strengths in emotional resilience or protective self-concepts; Hispanic parents’ moderate endorsement could indicate a culturally normative emphasis on balanced emotional reflection due to familism; and White parents’ higher endorsement likely relates to culturally reinforced ideals of self-accountability and individualistic emotional expression (Markus & Kitayama, 1991; Mesquita et al., 2017). Future studies should further examine the adaptive and maladaptive dimensions of rumination in parenting across cultural contexts, potentially employing mixed methods designs to disentangle these complex relationships.

Many studies argue that scalar invariance is often unattainable due to cultural and contextual differences that shape how individuals interpret and respond to survey items (Van de Schoot et al., 2015). Scalar non-invariance in self-report measures is particularly common in psychological and parenting research, as individuals from different ethnoracial backgrounds may endorse Likert-type-scale items differently due to socialization, language nuances, or differences in self-perception (Vandenberg & Lance, 2000). In emotion regulation research, guilt and self-reflection—both key constructs measured by the REPS rumination subscale—are experienced and expressed differently across cultural groups (Mesquita et al., 2017). For instance, collectivist cultures may emphasize family harmony and interdependent relationships, leading parents to experience guilt differently than those from more individualistic cultures (Markus & Kitayama, 1991). Given these inherent cultural variations, strict scalar invariance should not always be expected, and achieving partial scalar invariance may be the most realistic and interpretable approach (Byrne et al., 1989). Therefore, given the growing calls for increasing the representation of ethnoracially diverse groups in parenting research (Rodriguez et al., 2023, 2025), these findings are promising.

The reliability analysis further supports the robustness of the REPS across diverse racial and ethnic groups. McDonald’s omega reliability coefficients for all REPS subscales ranged from acceptable to excellent, indicating strong internal consistency across populations (McDonald, 1999; Revelle & Zinbarg, 2009). Specifically, the Adaptive Strategies and Suppression subscales demonstrated consistently high reliability across all groups (ω ranging from .88 to .91 and .83 to .89, respectively), affirming their stability and usefulness in diverse parenting contexts. However, findings regarding the Rumination subscale warrant further discussion. Although reliability was acceptable (ω ranging from .72 to .83), it was notably lower for Black (ω = .72) and White parents (ω = .74) compared with Hispanic and Asian parents (ω = .83). This variability aligns with other findings of partial scalar invariance for Rumination-related items. Furthermore, Rodriguez et al. (2025) highlighted an unexpected association where rumination within the parenting context correlated positively with both supportive and unsupportive emotion socialization practices. Rodriguez et al. (2025) suggested that this unusual pattern might reflect nuances in the REPS items themselves, such as parents’ feelings of guilt as well as desires to improve interactions, which may inadvertently capture both adaptive reflection and maladaptive rumination. Future research should therefore examine whether differences in the interpretation or cultural experiences of rumination among racial and ethnic groups could explain the observed variability in reliability, as well as the contradictory associations found in parenting outcomes. Overall, these findings affirm the REPS as psychometrically sound across groups yet highlight the complexity of rumination as a parental emotion regulation strategy, underscoring the need for further measure refinement and validation in diverse groups of parents.

Importantly, our analysis of differential validity did not reveal significant interaction effects between parental ethnoracial identity and REPS subscales in predicting child psychopathology when using latent REPS scores. This suggests that the associations between parental emotion regulation strategies and child psychological outcomes do not vary significantly by racial or ethnic identity. In other words, the effects of parental emotion regulation on child emotional and behavioral outcomes appear consistent across ethnoracial groups, supporting the generalizability of the REPS in diverse family contexts. This lack of significant moderation is somewhat surprising, given prior literature consistently demonstrating ethnoracial differences in emotional expression, parenting practices, and emotion regulation norms (Matsumoto et al., 2008; Mesquita et al., 2017; Rodriguez et al., 2025). Studies frequently highlight distinct patterns in how parents from diverse cultural backgrounds socialize emotions, respond to stressors, and manage emotional displays (Dunbar et al., 2015; Halgunseth et al., 2006; Nelson et al., 2013). However, one plausible explanation for our findings is that core parental emotion regulation processes—such as the fundamental capacity to manage one’s emotional reactions to respond to a child’s needs—may transcend specific cultural norms and behaviors. While parenting values, discipline strategies, and socialization practices undoubtedly differ across groups, the essential function of parental emotion regulation in supporting child well-being may remain largely universal (Rutherford et al., 2015).

Indeed, previous research has indicated that adaptive emotion regulation strategies (e.g., cognitive reappraisal, emotional awareness) can universally protect against parenting stress and foster positive parent–child interactions, irrespective of cultural background (Gratz & Roemer, 2004; Morris et al., 2017). Conversely, maladaptive strategies (e.g., suppression and rumination) may universally heighten stress, negatively affecting parenting quality and, subsequently, child outcomes (Gross, 1998). Given that parents across diverse ethnoracial contexts routinely face parenting challenges and emotional demands, the ability to regulate emotional responses adaptively likely has consistent implications for parenting effectiveness and child development (Morris et al., 2017; Rutherford et al., 2015). Thus, despite documented cultural differences in emotional expression and parenting norms, the core process of parental emotion regulation may exhibit robust generalizability, accounting for the absence of significant moderation effects observed in our study.

Although our differential validity analysis found no significant interaction effects, indicating consistent predictive validity across ethnoracial groups, previous research suggests potential subtle nuances in how certain strategies like rumination relate to child outcomes. Given Rodriguez et al.’s (2025) unexpected findings—where rumination was positively associated with both supportive and unsupportive parenting behaviors—future research might explore whether cultural interpretations of rumination mediate its association with parenting quality and child well-being. It is possible that subtle cultural differences in emotional expression or expectations around emotional introspection and guilt could moderate these associations in ways our current analysis did not detect. Qualitative research could clarify whether these subtle cultural differences impact practical parenting outcomes, informing culturally tailored intervention strategies. Nevertheless, our findings warrant replication.

Strengths and Limitations

This study has several notable strengths. First, it uses a large and racially diverse sample, allowing for a comprehensive examination of measurement invariance across ethnoracial groups. The use of multigroup CFA strengthens the findings by systematically testing the factor structure, factor loadings, and item intercepts of the REPS across racial and ethnic groups (Meredith, 1993). In addition, by incorporating McDonald’s omega rather than relying solely on Cronbach’s alpha, we provide a more robust estimate of internal reliability (Revelle & Zinbarg, 2009). Finally, the study’s inclusion of differential validity analyses helps determine whether the REPS functions similarly across groups in predicting child psychopathology outcomes, further supporting its generalizability (Aguinis & Gottfredson, 2010).

However, some limitations should be acknowledged. First, the study relied exclusively on self-report measures, which may introduce social desirability and response biases. Future studies should incorporate multi-informant methods, such as reports from children or observational measures of parent–child interactions, to improve validity. Second, while partial scalar invariance was achieved, the presence of measurement noninvariance in specific items suggests that cultural differences in emotional expression and parenting norms may influence how parents interpret emotion regulation items (Matsumoto et al., 2008). Future research should use qualitative approaches to explore these differences in depth. Finally, while our sample was racially and ethnically diverse, future studies should consider within-group variability, as racial and ethnic identity encompasses broad cultural, socioeconomic, and contextual factors that may influence parenting behaviors (Rodriguez et al., 2023). Furthermore, given the unexpected complexity surrounding the Rumination subscale highlighted by Rodriguez et al. (2025), future research should consider refining this subscale or developing supplementary qualitative or observational assessments to better capture the adaptive versus maladaptive dimensions of rumination in diverse parenting contexts. Exploring these complexities may not only improve psychometric robustness but also enhance the clinical utility and cultural responsiveness of the REPS.

Future Research Directions

Given the findings of this study, several avenues for future research should be explored. First, future studies should investigate how sociocultural factors, such as acculturation, socioeconomic status, and cultural values, shape parental emotion regulation strategies. Second, incorporating longitudinal designs would allow researchers to examine the stability of parental emotion regulation strategies over time and how these strategies influence long-term child outcomes (Putnick & Bornstein, 2016). Third, expanding research to include observational methods may provide deeper insights into how parents regulate their emotions in real-time interactions with their children. Finally, future validation studies should examine whether the REPS performs equivalently in other diverse populations, such as multiracial families or families with different immigration backgrounds.

Conclusion

Overall, this study provides promising evidence that the REPS is a reliable and valid measure of parental emotion regulation across ethnoracial groups. The findings suggest that the REPS can be confidently used to compare parental emotion regulation strategies in diverse populations, provided that minor adjustments for partial scalar invariance are considered. In addition, the absence of differential validity effects indicates that the REPS functions similarly across groups in predicting child psychopathology, reinforcing its applicability in research and clinical settings.

By continuing to investigate the cultural dimensions of parental emotion regulation, researchers and practitioners can develop more culturally informed approaches to understanding and supporting parental emotion regulation in diverse communities. These findings contribute to a growing body of literature emphasizing the importance of culturally responsive measurement tools in parenting and child development research.

Supplemental Material

sj-docx-1-asm-10.1177_10731911251390321 – Supplemental material for Measurement Invariance of the Regulation of Emotions in Parenting Scale (REPS): Psychometric Validation Across Ethnoracial Groups

Supplemental material, sj-docx-1-asm-10.1177_10731911251390321 for Measurement Invariance of the Regulation of Emotions in Parenting Scale (REPS): Psychometric Validation Across Ethnoracial Groups by Violeta J. Rodriguez in Assessment

Footnotes

Declaration of Conflicting Interests

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

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: V.J.R.’s work on this study was supported by a DP5 Early Independence Award from the National Institutes of Health Office of the Director (DP-5OD036508).

Consent to Participate

All participants provided informed consent before beginning the survey, in accordance with the guidelines approved by the relevant Institutional Review Board.

Consent for Publication

Not applicable.

ORCID iD

Violeta J. Rodriguez

Data Availability Statement

This study’s design and its analysis were not preregistered. All data, analysis code, and research materials are available upon request from the first author.

Supplemental Material

Supplemental material for this article is available online.

References

Aguinis

Gottfredson

R. K.

(2010). Best-practice recommendations for estimating interaction effects using moderated multiple regression. Journal of Organizational Behavior, 31(6), 776–786. https://doi.org/10.1002/job.686

Byrne

B. M.

Shavelson

R. J.

Muthén

(1989). Testing for the equivalence of factor covariance and mean structures: The issue of partial measurement invariance. Psychological Bulletin, 105(3), 456–466. https://doi.org/10.1037/0033-2909.105.3.456

Calzada

E. J.

Huang

K. Y.

Anicama

Fernandez

Brotman

L. M.

(2012). Test of a cultural framework of parenting with Latino families of young children. Cultural Diversity and Ethnic Minority Psychology, 18(3), 285–296. https://doi.org/10.1037/a0028694

Cheung

G. W.

Rensvold

R. B.

(2002). Evaluating goodness-of-fit indexes for testing measurement invariance. Structural Equation Modeling, 9(2), 233–255.

Dix

(1991). The affective organization of parenting: Adaptive and maladaptive processes. Psychological Bulletin, 110(1), 3–25. https://doi.org/10.1037/0033-2909.110.1.3

Dunbar

A. S.

Leerkes

E. M.

Coard

S. I.

Supple

A. J.

Calkins

(2015). An integrative conceptual model of parental racial/ethnic and emotion socialization and links to children’s social-emotional development among African American families. Child Development Perspectives, 9(1), 16–22. https://doi.org/10.1111/cdep.12104

Goodman

(1997). The Strengths and Difficulties Questionnaire: A research note. Journal of Child Psychology and Psychiatry, 38(5), 581–586. https://doi.org/10.1111/j.1469-7610.1997.tb01545.x

Gross

J. J.

(1998). Antecedent- and response-focused emotion regulation: Divergent consequences for experience, expression, and physiology. Journal of Personality and Social Psychology, 74(1), 224–237. https://doi.org/10.1037/0022-3514.74.1.224

Gratz

K. L.

Roemer

. (2004). Multidimensional assessment of emotion regulation and dysregulation: Development, factor structure, and initial validation of the difficulties in emotion regulation scale. Journal of Psychopathology and Behavioral Assessment, 26(1), 41–54.

10.

Halgunseth

L. C.

Ispa

J. M.

Rudy

(2006). Parental control in Latino families: An integrated review of the literature. Child Development, 77(5), 1282–1297. https://doi.org/10.1111/j.1467-8624.2006.00934.x

11.

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

12.

Markus

H. R.

Kitayama

(1991). Culture and the self: Implications for cognition, emotion, and motivation. Psychological Review, 98(2), 224–253. https://doi.org/10.1037/0033-295X.98.2.224

13.

Matsumoto

Yoo

S. H.

Fontaine

(2008). Mapping expressive differences around the world: The relationship between emotional display rules and individualism versus collectivism. Journal of Cross-Cultural Psychology, 39(1), 55–74. https://doi.org/10.1177/0022022107311854

14.

McDonald

R. P

. (1999). Test theory. A unified treatment. Mahwah, NJ: Erlbaum.

15.

McDonald

R. P.

(2013). Test theory: A unified treatment. Psychology Press.

16.

Meredith

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

17.

Mesquita

Boiger

De Leersnyder

(2017). The cultural construction of emotions. Current Opinion in Psychology, 17, 31–36. https://doi.org/10.1016/j.copsyc.2017.06.001

18.

Morris

A. S.

Criss

M. M.

Silk

J. S.

Houltberg

B. J.

(2017). The impact of parenting on emotion regulation during childhood and adolescence. Child Development Perspectives, 11(4), 233–238.

19.

Nelson

J. A.

Leerkes

E. M.

Perry

N. B.

O’Brien

Calkins

S. D.

Marcovitch

(2013). European-American and African-American mothers’ emotion socialization practices relate differently to their children’s academic and social-emotional competence. Social Development, 22(3), 485–498. https://doi.org/10.1111/j.1467-9507.2012.00673.x

20.

Nye

C. D.

Drasgow

(2011). Effect size indices for analyses of measurement equivalence: Understanding the practical importance of differences between groups. Journal of Applied Psychology, 96(5), 966–980.

21.

Putnick

D. L.

Bornstein

M. H.

(2016). Measurement invariance conventions and reporting: The state of the art and future directions for psychological research. Developmental Review, 41, 71–90.

22.

Revelle

Zinbarg

R. E.

(2009). Coefficients alpha, beta, omega, and the glb: Comments on Sijtsma. Psychometrika, 74(1), 145–154. https://doi.org/10.1007/s11336-008-9102-z

23.

Rhemtulla

Brosseau-Liard

P. É.

Savalei

(2012). When can categorical variables be treated as continuous? A comparison of robust continuous and categorical SEM estimation methods under suboptimal conditions. Psychological Methods, 17(3), 354–373.

24.

Rodriguez

Shaffer

(2021). Development and validation of the Regulation of Emotions in Parenting Scale (REPS). Journal of Family Psychology, 35(3), 330–340. https://doi.org/10.1037/fam0000782

25.

Rodriguez

V. J.

La Barrie

D. L.

Zegarac

M. C.

Shaffer

(2023). A systematic review of parenting scales measurement invariance/equivalence by race and ethnicity: Recommendations for inclusive parenting research. Assessment, 30(1), 22–36. https://doi.org/10.1177/10731911211038630

26.

Rodriguez

V. J.

Shaffer

A. E.

Parent

(2025). Revision of the Multidimensional Assessment of Parenting Scale (MAPS) in an ethnoracially diverse national sample. Journal of Family Psychology, 39(1), 107–120. https://doi.org/10.1037/fam0001271

27.

Rosseel

(2012). lavaan: An R package for structural equation modeling. Journal of Statistical Software, 48(2), 1–36. https://doi.org/10.18637/jss.v048.i02

28.

Rutherford

H. J. V.

Wallace

N. S.

Laurent

H. K.

Mayes

L. C.

(2015). Emotion regulation in parenthood. Developmental Review, 36, 1–14. https://doi.org/10.1016/j.dr.2014.12.008

29.

Vandenberg

R. J.

Lance

C. E

. (2000). A review and synthesis of the measurement invariance literature: Suggestions, practices, and recommendations for organizational research. Organizational Research Methods, 3(1), 4–70.

30.

Van de Schoot

Lugtig

Hox

(2015). A checklist for testing measurement invariance. European Journal of Developmental Psychology, 12(5), 541–560. https://doi.org/10.1080/17405629.2015.1046111

31.

Weiss

N. H.

Thomas

E. D.

Schick

M. R.

Reyes

M. E.

Contractor

A. A.

(2022). Racial and ethnic differences in emotion regulation: A systematic review. Journal of Clinical Psychology, 78(5), 785–808.

32.

Zeileis

(2004). Econometric computing with HC and HAC covariance matrix estimators. Journal of Statistical Software, 11(10), 1–17. https://doi.org/10.18637/jss.v011.i10

33.

Zeileis

Hothorn

(2002). Diagnostic checking in regression relationships. R News, 2(3), 7–10.

34.

Zeileis

Köll

Graham

(2020). Various versatile variances: An object-oriented implementation of clustered covariances in R. Journal of Statistical Software, 95(1), 1–36. https://doi.org/10.18637/jss.v095.i01

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.

0.00 MB

0.04 MB