The degree to which mothers feel competent in their parenting role has important implications for their mental health, and their children’s well-being and development (Coleman & Karraker, 1998; Jones & Prinz, 2005; Coleman & Karraker, 1998). Mothers who derive more value from the parenting role report less depression and parenting stress (Crnic & Greenberg, 1990; Rogers & Matthews, 2004). Accordingly, self-appraisal of parenting competence is a targeted mechanism in a variety of parenting-based interventions (Gilmore & Cuskelly, 2009). Efforts to deliver these interventions effectively will be enhanced by the availability of reliable and valid measures of parenting self-esteem and feelings of competence.

Given cultural differences in parenting practices and cross-cultural differences in the degree of importance placed on the parental role (Bornstein & Cheah, 2006), it is important to validate measures of parenting self-esteem among culturally and ethnically diverse samples. Members of the second-fastest growing ethnic minority group in the U.S. (Pew Research Center, 2017), Mexican-origin mothers (e.g., first-generation Mexican immigrants and women with Mexican heritage who are living in the U.S.) may straddle two cultures, each with potentially different expectations for the maternal role. The traditional Mexican cultural values of familism and marianismo place high value on motherhood, emphasizing the importance of putting children and family members above one’s own needs (Castillo & Cano, 2007). Embedded in these values is the expectation that Mexican mothers will be successful and self-sacrificial mothers (Connelly et al., 2013). In contrast, European-American values reflect egalitarian beliefs about parenting and expectations that women can successfully maintain dual roles as parents and providers (Thornton & Young-DeMarco, 2001). European-American mothers also tend to express high levels of confidence in their parenting role relative to immigrant mothers in the U.S. (Bornstein & Cote, 2004). Thus, self-appraisal of one’s own parenting may be especially relevant to the well-being of Mexican-origin mothers in the U.S.

According to recent reviews (Jones & Prinz, 2005; Hurley et al., 2014), one of the most commonly used and psychometrically acceptable measures of parenting cognitions and perceived competence is the Parenting Sense of Competence Scale (PSOC; see Table 1; Gibaud-Wallston & Wandersman, 1978; Johnston & Mash, 1989). Derived from Wallston & Wandersman’s (1978) measure, the PSOC is comprised of items aimed at measuring parenting self-esteem, the degree to which parents feel competent in and satisfied by the parenting role. Johnston and Mash (1989) first documented the PSOC’s factor structure, concluding that an interpretable two-factor structure existed among their sample of Canadian parents (297 mothers and 215 fathers of children ages 4 to 9 years). Their commonly used, two-facture structure is comprised of parenting satisfaction, the degree to which individuals derive value from the parenting role, and parenting self-efficacy, the degree to which individuals believe they are capable of successful parenting. Ohan et al. (2000) replicated the two-factor structure (satisfaction and self-efficacy) among a Canadian sample of 110 mothers and 110 fathers (children ages 5 to 12 years), suggesting that the underlying structure of the PSOC remains stable across parental gender. Gilmore and Cuskelly (2009) evaluated factor structure stability among an Australian community sample of 586 mothers and 615 fathers (children ages 0 to 18 years), finding evidence of a three-factor solution (satisfaction, self-efficacy, and interest). The “parenting interest” factor was comprised of two items and contributed to a smaller proportion of overall variance accounted for than the first two factors among both mothers (8.4%) and fathers (8.5%). Others found similar evidence suggestive of a third factor reflecting parenting interest (Rogers & Matthews, 2004). Gilmore and Cuskelly (2009) also found preliminary evidence of a potential fourth factor (control), but given differential item loading for mothers and fathers, this factor was removed.

Table 1 PSOC Item Loadings onto EFA Two-Factor Solution at 12 monthsa
Full size table

Although evidence for a two-factor structure (satisfaction, self-efficacy) of the PSOC has been reported across multiple studies, the majority of prior work is derived from samples of Canadian (Johnston & Mash, 1989; Ohan et al., 2000) and Australian (Gilmore & Cuskelly, 2009) parents. Few psychometric evaluations of the PSOC have been conducted among non-English-speaking or ethnically diverse parents. Of those studies, there exists variability in the number of factors and underlying factor constructs that emerge. Ngai et al. (2007) found evidence supporting the two-factor structure (satisfaction and self-efficacy) of the Chinese-translated PSOC among a sample of Chinese mothers. Among a sample of Thai fathers, researchers found additional evidence in support of the two-factor structure, valuing/comfort (satisfaction) and skill/knowledge (self-efficacy) (Suwansujarid et al., 2013). Nunes et al. identified a three-factor structure (dissatisfaction, self-efficacy, and controllability) of the Portuguese-translated PSOC among at-risk Portuguese mothers (Nunes et al., 2016). Among a sample of 259 at-risk Spanish mothers, Menéndez Álvarez-Dardet et al. (2011) propose a shortened 10-item PSOC assessing two subscales: self-efficacy and controllability. Another study with 600 Spanish mothers concluded the original two-factor structure, satisfaction and efficacy, to be the best fit to the data (Oltra‐Benavent et al., 2020). Haack et al. (2011) demonstrated good internal consistency and convergent validity for the original parenting self-efficacy subscale (7 items; from Johnston & Mash, 1989) of a Spanish-translated PSOC among a sample of Spanish-speaking, Latino parents in the United States (85% of the sample identified as Mexican); the authors did not evaluate the psychometric properties of the full PSOC or other subscales.

Among a sample of Mexican American mothers, we had the following study aims: (1) evaluate the factor structure of the PSOC at 12 months postpartum, (2) provide validity evidence by comparing the current factor structure to existing research, and (3) compare scores on the PSOC to relevant maternal (age, parity, country of birth, language use, depressive symptoms, parenting stress) and child (temperament) characteristics as a preliminary test of validity. Relevant maternal and child characteristics were selected for tests of validity consistent with prior psychometric work on the PSOC, which supports a negative association between maternal depressive symptoms and parenting sense of competence during infancy (Ngai et al., 2007), a negative association between maternal parenting stress and parenting sense of competence (Nunes et al., 2016), and a negative association between child behavior problems and parenting sense of competence (Johnston & Mash, 1989). No specific hypotheses were made surrounding maternal sociodemographic characteristics, given significant variability in direction and significance of effects across existing psychometric work with ethnically and linguistically diverse samples.

Method

Participants

The sample included 205 mothers (176 Spanish-speaking, 29 English-speaking) participating in the Las Madres Nuevas study of low-income, Mexican American mothers and children. Women were recruited from a hospital-based prenatal clinic. Eligibility criteria included: (1) self-identification as Mexican or Mexican American, (2) English or Spanish fluency, (3) at least 18 years of age, (4) low-income status (defined as family income below $25,000 or eligibility for Medicaid or Federal Emergency Services coverage for the childbirth), and (5) anticipated delivery of a singleton birth with no prenatal evidence of health or developmental problems. The Arizona State University IRB and the Maricopa Integrated Health System IRB approved all study procedures prior to recruitment.

At the time of enrollment, women were between 18 and 42 years of age (M = 27.8, SD = 6.48) and had completed, on average, 10 years of education. Most women were born in Mexico (86.0%) and spoke Spanish as their primary language (82.0%). The modal family income was $10,001–$15,000 for an average household of four people. Most women were married or living with a romantic partner (77%). The number of other biological children ranged from zero to nine (22.2% of the women were first-time mothers) at the time of enrollment.

Procedures

Data for the current study were collected at a prenatal home visit and laboratory visit when children were 12 months old. During a prenatal care appointment, women were approached by a bilingual interviewer who explained the study and evaluated eligibility. The interviewer obtained informed consent at a prenatal home visit between 26–38 weeks gestation. Interviews were conducted in participants’ choice of Spanish (85%) or English (15%). Survey questions were read aloud and participants were given visual aids with written and graphic descriptions of item response formats. The 12-month visit was conducted in the laboratory at Arizona State University. Participants were compensated $50 and given small gifts for the baby (e.g., rattles, bibs) at the prenatal visit, and $100 at the laboratory visit.

For administration in Spanish, all measures not already available in Spanish were translated and back-translated by certified translators fluent in English and Spanish, with discrepancies resolved by the principal investigators and lead translators, following standard procedures (Behling & Law, 2000). Spanish-translated measures were adapted and field tested to ensure cultural sensitivity and fit for our population. Measures that had previously been validated in Spanish were also reviewed for local relevance and cultural sensitivity for our population.

Measures

Parenting sense of competence scale

At the 12-month visit, mothers completed the 17-item Parenting Sense of Competence Scale (PSOC). Items are responded to on a 6-point scale from “strongly disagree” to “strongly agree”. Higher scores indicate higher sense of competence. One item (Item 17—“Being a good mother/father is a reward in itself.”) was removed preceding data collection due to prior research indicating a failure of that item to load onto any factor (Johnston & Mash, 1989; Suwansujarid et al., 2013). Items reflect general cognitions about parenting, rather than feelings about parenting a specific child. Most mothers completed the PSOC in Spanish (n = 176) rather than English (n = 29).

Edinburgh perinatal/postnatal depression scale

At the prenatal visit, women completed the 10-item Edinburgh Perinatal/Postnatal Depression Scale (EPDS; Cox et al., 1987; Cronbach’s α = 0.80). Higher scores indicate more depressive symptoms. The EPDS has been previously validated in both English and Spanish (Garcia-Esteve et al., 2003). A score of 13 or higher is considered clinically significant (Cox et al., 1987).

Center for epidemiological studies depression scale

At the 12-month visit, mothers completed the 20-item Center for Epidemiologic Studies Depression scale (CES-D; Radloff, 1977; Cronbach’s α = 0.88). Higher scores indicate more depressive symptoms. A score of 16 or more is considered clinically significant. Among ethnic minority women, a score of 24 or higher is considered very high risk (Heilemann et al., 2004). The CES-D has been previously validated in Spanish among Mexican American individuals (Roberts et al., 1990).

Parenting daily hassles scale

At the 12-month visit, mothers reported their parenting stress using the 20-item Parenting Daily Hassles Scale (PDH; Crnic & Greenberg, 1990; Cronbach’s α = 0.95). Respondents rate the frequency and intensity with which they face common parenting situations, such as “the kids get dirty several times a day requiring several changes of clothes,” on a scale from 1 (none of the time; not at all a hassle) to 5 (an extreme amount of the time; an extreme hassle). Higher scores indicate higher parenting stress. The current study used a sum score of the intensity ratings.

Infant behavior questionnaire

Mothers reported on their infants’ temperament at 12 months using the 91-item Infant Behavior Questionnaire—Revised, Short Form (Gartstein & Rothbart, 2003; Putnam et al., 2014). Respondents rate how often during the last two weeks their baby exhibited certain behaviors, such as “cry or fuss before going to sleep for naps” on a scale from 1 (never) to 7 (always). Three overarching constructs are derived: positive affectivity/surgency (Cronbach’s α = 0.69), negative affectivity (Cronbach’s α = 0.61), and regulatory capacity (Cronbach’s α = 0.45). Only the positive affectivity/surgency and negative affectivity constructs were used.

Miscellaneous

Demographic characteristics (maternal age; maternal country of birth coded 1 = United States, 2 = Mexico; language preference, coded 1 = English, 2 = Spanish; number of other children) were reported by mothers at the prenatal visit.

Data Analysis

All analyses were conducted in Mplus v. 8. (Muthén & Muthén, 1998–2017). Given substantial heterogeneity in factor structures across existing psychometric work on the PSOC among ethnically and linguistically diverse populations of parents, exploratory factor analysis (EFA) was first completed to examine the underlying factor structure among this population, given EFA is considered a more conservative approach (Hurley et al., 2014). The EFA was conducted using robust maximum likelihood (ML) factor analysis and oblique geomin rotation to allow for correlations among factors, consistent with existing literature. Robust normal theory ML is considered acceptable for categorical responses when there are 5 or more categories (Rhemtulla et al., 2012). Consistent with best practice recommendations, examination of the scree plot and parallel analysis were used to determine the number of factors to be tested using EFA. Item content was also used to help determine the final recommended factor solution. Confirmatory factor analysis (CFA) was conducted to evaluate model fit statistics [root mean square error of approximation (RMSEA) ≤ 0.08, comparative fit index (CFI) ≥ 0.95, and standardized root mean square residual (SRMR) < 0.08; Hu & Bentler, 1999)] of the factor solution suggested by the EFA. Robust ML was also used for the CFA.

Missing Data

Because funding was not initially available for a 12-month lab visit, only 266 of 322 families were eligible for the 12-month lab visit. Of those eligible, 206 (77%) families completed the 12-month interview. One case was missing on the PSOC measure, such that the total sample size for all analyses was 205 mothers.

Results

Preliminary Analyses

Descriptive statistics for individual items comprising the PSOC are shown in Table 2. Several items were negatively skewed (skewness < −1; items 1, 10, 11, 13, and 15), and one was positively skewed (skewness > 1; item 12). Intra-item correlations reflected the traditional separation of the overall PSOC scale into satisfaction and self-efficacy subscales.

Table 2 Descriptive Statistics for 12-month PSOC Itemsa
Full size table

Primary Analyses

The scree plot for PSOC items suggested evaluation of one-, two-, three-, and four-factor models (see Fig. 1). The first four sample eigenvalues were 3.85, 2.72, 1.33, and 1.20. Comparison of parallel analysis and the scree plot suggested evaluation of a one-, two-, and three-factor solution (see Fig. 1). The first four eigenvalues from parallel analysis were 1.52, 1.39, 1.31, and 1.24. Factors are considered significant if the associated eigenvalue is bigger than that derived from the random data (Horn, 1965), and thus one-, two-, and three-factor solutions were evaluated. In the three-factor model, the third factor (significant loadings for items 14 and 16) was similar to the parenting interest factor previously documented in the literature, but not an exact replication of that factor (which consists of items 12 and 14). However, the third factor contributed minimal additional variance accounted for in the overall PSOC scale (5%) and consisted of two items only. Items loading onto the first and second factors aligned with prior psychometric evaluations of the PSOC. Therefore, the two-factor model was considered to be the most promising).

Fig. 1
figure 1

Scree Plot and Parallel Analysis for 12-month PSOC

Full size image
Fig. 2
figure 2

Confirmatory factor analysis of two-factor model of PSOC1. 1Solid lines are statistically significant; dashed lines are non-significant (standardized estimates are shown for both paths); Error terms and correlated residuals not shown

Full size image

Within the two-factor solution, the first factor (parenting satisfaction) consisted of 8 unique items assessing enjoyment in the parenting role, with loadings ranging from 0.32 to 0.75 (see Table 1). Factor 2 (parenting self-efficacy) consisted of 7 unique items assessing feelings of competence in the parenting role. Item loadings onto factor 2 ranged from 0.35 to 0.76 (see Table 1). Item 14 loaded significantly onto both factors, though neither loading was large (0.38 loading on the parenting satisfaction factor; 0.20 loading on the parenting self-efficacy factor). The correlation between the parenting satisfaction and parenting self-efficacy factors was not significant (r = −0.07), contrary to other work (e.g., Johnston & Mash, 1989; Ohan et al., 2000; Menéndez Álvarez-Dardet et al., 2011).

Scale scores for the parenting satisfaction subscale and parenting self-efficacy subscale were calculated by summing the relevant items identified in the EFA. Given that Item 14 loaded more strongly onto the satisfaction factor and is traditionally conceptualized to be part of the satisfaction subscale, it was included in the satisfaction subscale (Cronbach’s α = 0.81). The parenting self-efficacy subscore had internal reliability of Cronbach’s α = 0.71.

The concluded two-factor model solution from the EFA was then tested in a CFAFootnote 1. Nine items (2, 3, 4, 5, 8, 9, 12, 14, and 16) assessed the construct of parenting satisfaction. The error terms for Item 14 and Item 16 were allowed to covary given their loadings onto a potential third factor in the EFA. Seven items (1, 6, 7, 10, 11, 13, and 15) assessed the construct of parenting self-efficacy. The two-factor model had an acceptable fit to the data according to RMSEA and SRMR fit indices (CFI = 0.829, RMSEA = 0.077, SRMR = 0.075). Modification indices suggested that there remained residual covariance among subsets of the items. Seven such residuals were added (Item pairs 14/16, 14/7, 14/12, 15/8, 10/8, 6/3, and 13/3) and the resulting model had better fit (CFI = 0.92, RMSEA = 0.055, SRMR = 0.065; see Fig. 2) with minimal impact on the factor loadings (loadings with and without the modification indices were correlated over 0.99). Parenting satisfaction was not significantly correlated with parenting self-efficacy (r = −0.098, p = 0.32).

Correlates of Parenting Competence

Maternal well-being

Prenatal depressive symptoms were negatively correlated with parenting satisfaction (r = −0.187, p < 0.01) and the parenting self-efficacy (r = −0.257, p < 0.001). Concurrent maternal depressive symptoms were negatively correlated with parenting satisfaction (r = −0.523, p < 0.001) and parenting self-efficacy (r = −0.286, p < 0.001). Concurrent parenting stress was negatively correlated with parenting satisfaction (r = −0.250, p < 0.001) and parenting self-efficacy (r = −0.361, p < 0.001).

Child temperament

Parenting satisfaction was positively correlated with infant positive affectivity/surgency (r = 0.155, p = 0.034), but not infant negativity (r = −0.104, p = 0.16). Parenting self-efficacy was negatively correlated with infant negativity (r = −0.172, p = 0.018), but not infant positive affectivity/surgency (r = 0.054, p = 0.46).

Demographic correlates

Parenting satisfaction was correlated with maternal country of birth (r = −0.147, p = 0.036) and maternal age (r = −0.139, p = 0.048), such that younger mothers and mothers born in the U.S. reported higher parenting satisfaction, but was not significantly correlated with interview language (r = −0.089, p = 0.20). Parenting self-efficacy was positively correlated with maternal age (r = 0.148, p = 0.035), such that older mothers reported higher parenting self-efficacy, but was not significantly correlated with interview language (r = 0.011, p = 0.87) or maternal country of birth (r = −0.003, p = 0.96). The number of other biological children in the home (parity) was negatively correlated with parenting satisfaction (r = −0.187, p < 0.01), such that more children in the home was associated with less parenting satisfaction, but not parenting self-efficacy (r = 0.016, p = 0.82).

Summary

In sum, a two-factor solution (parenting self-efficacy, parenting satisfaction) emerged as the best-fitting model of the PSOC administered at 12 months postpartum among a sample of low-income, predominantly Spanish-speaking, Mexican-origin mothers living in the United States. Although a third factor also emerged, it contributed less to the overall variance accounted for and consisted of two items only, and thus was excluded in the proposed factor structure. Parenting satisfaction and parenting self-efficacy were significantly associated with several mother and child characteristics.

Discussion

The current study evaluated the psychometric properties of the Parenting Sense of Competence (PSOC) scale among a sample of low-income, predominantly Spanish-speaking, Mexican-origin mothers living in the United States. Consistent with previous work (Johnston & Mash, 1989; Ohan et al., 2000; Ngai et al., 2007; Suwansujarid et al., 2013), we found support for a two-factor structure (satisfaction and self-efficacy) at 12 months postpartum. The parenting satisfaction and parenting self-efficacy subscores demonstrated acceptable internal consistency and convergent validity evidence, suggesting reliable and valid use of the PSOC scale among Spanish-speaking or Mexican American samples.

Unlike prior research, the satisfaction and self-efficacy subscales were not significantly correlated; thus, the current analyses do not support summing the two subscales to calculate overall feelings of parenting competence among predominantly Spanish-speaking, Mexican-origin mothers. Small to moderate positive correlations between the satisfaction and self-efficacy PSOC subscales have been reported in existing psychometric work on the PSOC scale (Gilmore & Cuskelly, 2009; Johnson & Mash, 1989, Ohan et al., 2000; Suwansujarid et al., 2013). Cultural differences may explain differences in the strength and significance of this correlation across ethnically and linguistically diverse samples. Embedded within traditional Mexican values, marianismo values encompass women’s self-sacrificial devotion to the maternal role and the centrality of the maternal role within Mexican culture. As such, Mexican American women and their families may be more likely to highly value pregnancy, the maternal role, and childrearing (Gress-Smith et al., 2013; Page, 2004). Among Latina mothers in the United States, increases in acculturation have been associated with less parenting self-efficacy (Ceballo & Hurd, 2008). Although the current study does not directly evaluate acculturation or alignment to traditional cultural values, mothers born in Mexico reported less parenting satisfaction. No differences in parenting satisfaction or parenting self-efficacy were noted based on maternal language preference. Given no other prior psychometric work on the PSOC among Mexican-origin mothers in the United States, future work may aim to confirm this unexpected nonsignificant correlation or directly evaluate alignment to traditional Mexican cultural values in relation to maternal report of parenting satisfaction and parenting self-efficacy.

Women who reported fewer prenatal depressive symptoms reported higher parenting satisfaction and parenting self-efficacy twelve months later. In line with prior work (Jones & Prinz, 2005), greater parenting satisfaction and parenting self-efficacy were associated with fewer concurrent depressive symptoms and less parenting stress. Greater parenting satisfaction was associated with fewer biological children in the home. Although having more children is frequently thought to promote parenting self-efficacy via greater experience, additional children may contribute to parenting stress, lessening the satisfaction mothers derive from their role. Interestingly, older mothers reported less parenting satisfaction, but greater parenting self-efficacy. Existing literature remains inconclusive regarding the effects of parental age on parental competence (Ngai et al., 2007). More infant positive affectivity was associated greater parenting satisfaction, while less infant negativity was associated with greater parenting self-efficacy. Mothers of less difficult children may consequentially feel more satisfied and efficacious in their maternal role (Coleman & Karraker, 2000); on the contrary, mothers who feel more satisfied and efficacious in their maternal role may report less difficulty in their infants’ temperament (Fox & Gelfand, 1994). Consistent with prior work, correlations between maternal parenting self-appraisals and child behavior were smaller (Ohan et al., 2000), while correlations between maternal parenting self-appraisals and concurrent mental health and parenting stress were in the moderate to large range (Ngai et al., 2007; Nunes et al., 2016).

This study makes an important contribution to existing literature by extending understanding of the psychometric properties of the PSOC scale to a community sample of low-income Mexican-origin mothers in the U.S. We demonstrated the validity of the PSOC scale among Mexican American mothers, the majority of whom are Spanish-speaking, and provided evidence for factor-structure stability of the Spanish-translated PSOC scale consistent with prior psychometric work among ethnically and linguistically diverse populations. However, there were several limitations. The results may not generalize to higher-income or more acculturated Mexican-origin mothers in the U.S, or to those with older children. Limited power, due to a small sample of English-speaking mothers (n = 29), precluded our ability to reliably test the Spanish-translated PSOC for bias using measurement invariance strategies. However, as noted earlier, the two-factor model of the PSOC fit adequately well among Spanish-speaking mothers compared to the full sample of Spanish- and English-speaking mothers. Future work may aim to test measurement invariance across the English PSOC and Spanish PSOC among Mexican American mothers. We only evaluated the factor structure of the PSOC scale among a sample of Mexican-origin mothers. Psychometric evaluation with Mexican-origin fathers would contribute to the literature. Lastly, sample size limited the ability to conduct a CFA on a hold-out sample.

In sum, the current study documents a two-factor structure of the PSOC scale and the use of satisfaction and self-efficacy subscales with low-income Mexican-origin mothers in the U.S., the majority of whom are Spanish-speaking. Women who felt more competent in the parenting role also reported fewer depressive symptoms and less parenting stress. Understanding the construct and correlates of parenting sense of competence among this group is critical given that low-SES Mexican-origin women in the U.S. have higher birth rates than ethnic majority women (Lopez, 2015) and experience more risks to well-being in the postpartum (Institute of Medicine, 2001). As Hurley et al. (2014) highlight, self-appraisal of parenting competence is a targeted mechanism across several parenting-based interventions; thus, psychometric evaluation of measures assessing parenting competence among diverse populations is critical to both research and practice. Our findings suggest the PSOC scale to be an adequate measure of parenting sense of competence among Spanish-speaking and low-income Mexican American mothers in the U.S.