Psychometric properties and measurement invariance of the short form of grit scale in Korean adolescents

Sung-Woo Bae; Ji Geun Kim; Byung-Sun Park; Keungeun Lee; Jungkyu Park

doi:10.1371/journal.pone.0296795

Abstract

This study aimed to identify the factor structure of the Korean version of the Short Grit Scale (Grit-S) and examine its cross-sectional and longitudinal measurement invariance (MI). Data from the Korean Children and Youth Panel Survey 2018 were analyzed, which included two cohorts, comprising 2,327 and 2,325 fourth-year elementary and first-year middle school students, respectively. It was found that the two-factor model fit the data well for the elementary and middle school samples. The results of the cross-sectional MI tests across genders indicated that the full threshold and loading invariance were also supported for the elementary school sample, and the partial threshold and loading invariance were supported for the middle school sample. The analyses of the longitudinal MI revealed that the partial threshold and loading invariance were supported for both samples. The reliability analysis revealed satisfactory McDonald’s Omega values for both samples at each time point and moderate stability coefficients over time. Based on these findings, it was concluded that the Korean version of the Grit-S demonstrated satisfactory psychometric properties and exhibited MI across gender and time in Korean adolescents.

Citation: Bae S-W, Kim JG, Park B-S, Lee K, Park J (2024) Psychometric properties and measurement invariance of the short form of grit scale in Korean adolescents. PLoS ONE 19(1): e0296795. https://doi.org/10.1371/journal.pone.0296795

Editor: Syed Amir Ashraf, University of Ha’il, SAUDI ARABIA

Received: November 1, 2023; Accepted: December 19, 2023; Published: January 19, 2024

Copyright: © 2024 Bae et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: The data underlying this study’s findings can be found here: https://www.nypi.re.kr/archive/mps. To access data, click on "Korean Children and Youth Panel Survey" under the Panel Survey Overview section.

Funding: The author(s) received no specific funding for this work.

Competing interests: The authors declare that they have no competing interest.

Introduction

Grit is a psychological concept proposed by Duckworth et al. [1], which they defined as a personal, noncognitive attribute that reflects a person’s passion and perseverance for achieving long-term goals. Grit has two dimensions: consistency of interests (CI), which refers to the extent to which people can maintain a focused passion for a specific interest or goal over a long time, and perseverance of effort (PE), which refers to people’s inclination to uphold commitment and sustain exertion when confronted with challenges and adversity. Previous research has demonstrated that grit could be a measure for assessing positive educational, professional, and personal outcomes [2–4], and has been positively associated with school-related performances such as academic achievement, academic engagement, and school motivation [1, 5, 6]. People with high grit exhibit greater work retention and engagement and have fewer career changes (e.g.: [1, 6]). Grit also has associations with positive personal outcomes, such as psychological well-being, good mental health, and finding meaning in life [7–9].

To appropriately measure grit, Duckworth et al. [1] initially developed a 12-item Grit Scale, after which Duckworth and Quinn [10] improved the overall psychometric properties of the scale and validated an 8-item Short Grit Scale (Grit-S). Confirmatory factor analyses showed that the Grit-S fit data better than the Grit-O, and the predictive validity had not decreased even though there were fewer items, that is, the Grit-S was a more efficient grit measure. Consequently, the Grit-S was translated into many languages, including Korean [11], Japanese [12], Turkish [13, 14], German [15], Polish [16], Italian [17], Spanish [18, 19], Czech [20, 21], Chinese [22–24], Arabic [25], and Portuguese [26].

Although Grit-S has been the most frequently used grit measure, there have been concerns that the factor structure of Grit-S is inconsistent for different cultural groups [27]. The original factor structure of Grit-S, based on confirmatory factor analysis was a higher-order structure with two first-order factors, PE and CI [10]; however, factorial structural differences were found in collectivistic cultures. For example, the hierarchical Grit-S structure was not replicated in Filipino samples [28], and two items did not load significantly on the CI factor in Turkish samples [13]. Disabato et al. [27] analyzed the grit factor structure using culturally diverse samples from 109 countries and found that both CI and PE had low correlations in collectivistic cultures, and the CI and PE factor loadings on the general grit factor had variability even in collectivistic cultures. Those findings suggested that the Western individualistic concept of grit may be perceived differently in collectivistic social contexts [3, 29].

In addition to cultural differences, the grit factor structure may vary in different age groups. For example, Muenks et al. [30] compared the model fit indices for several educational levels and found that the bifactor model had a better fit for college students and that the two-correlated factor model was more appropriate for high school students. These findings suggested that people at different developmental stages may perceive grit differently. Unfortunately, most previous research has focused on validating the psychometric properties of grit using a wide range of samples, such as high school students [22, 31], college students [12, 13, 17, 18, 21, 26], and adult populations [16, 19, 24]. However, as the Grit-S factor structure in early adolescence has not yet been explored, the first objective of this study was to explore the grit factor structure in elementary and middle school students and the second objective was to investigate the Grit-S factor structure using samples from South Korea, a collectivist culture. Therefore, the findings of this study could shed light on the generalizability of the Western individualistic grit concept to a collectivist society.

A third objective was to assess the cross-sectional and longitudinal measurement invariance (MI) of Grit-S. Previous cross-sectional MI of Grit-S studies that have examined the associations between gender and grit have had mixed findings. For example, some studies have found that females have higher levels of grit than men [6, 10, 32, 33], while others found no gender differences [1, 26, 34]. These contradictory findings suggested that more research was needed to better understand the relationship between grit and gender and that establishing cross-sectional measurement invariance (CMI) was essential for accurately interpreting any observed mean differences between gender. In addition, several previous studies that examined MI for Grit-S across genders reported varying results across different age groups. For example, scalar invariance was found in middle school students [35], partial strict invariance in high school students [31], strict invariance in college students [26], and partial scalar or strict invariance in adult populations [16, 24]. However, due to the lack of gender-specific MI research with early adolescent samples, it remains unclear whether the Grit-S measures the same constructs in girls and boys in this age group. Therefore, this study also examined the MI of the Grit-S across genders in a sample of elementary and middle school students.

Given the widespread research interest in understanding the growth and changes in grit over time [36–39], it is also important to establish longitudinal measurement invariance (LMI) of the Grit-S. If there is evidence that shows that Grit-S measures the same construct with consistent structural validity across several time points, then observed changes in grit over time could be interpreted as genuine changes rather than changes in the construct’s structure or measurement. However, there is little research examining the LMI on Grit-S. To the best of our knowledge, only two cross-cultural studies have been conducted [23, 40], which makes it difficult to generalize. Furthermore, as these studies utilized analytical methods that did not account for the ordinal nature of Grit-S, this study examined the LMI of Grit-S by conducting a multigroup confirmatory factor analysis (CFA) using the unweighted least squares (UWLS) method following the guideline by Wu & Estabrook [41] to account for the ordered categorical nature of the Grit-S measurement.

As further research concerning the psychometric properties and measurement invariance of the Grit-S in early adolescents and collectivistic cultures is warranted, this study aimed to (a) explore the factor structure of the Grit-S in Korean elementary and middle school students, (b) assess the CMI to determine whether the Grit-S measures the same attribute across genders, and (c) examine the LMI at three time points over a one-year interval to assess the adequacy of the Grit-S for longitudinal comparisons.

Materials and methods

Participants

This study was conducted using data from the Korean Children and Youth Panel Survey 2018 (KCYPS 2018), which is a nationally representative longitudinal dataset collected by the National Youth Policy Institute using a stratified sampling approach [42]. The participant data were extracted using multistage cluster sampling. In total, 16 administrative districts were stratified, and schools were randomly selected from each district using proportionate probability sampling based on the population rate. Finally, one class from each school was randomly selected, and the data were collected annually from face-to-face interviews. The KCYPS 2018 consisted of two cohorts that included fourth-year elementary school students and first-year middle school students.

This study was approved by the Institutional Review Board of Kyungpook National University (IRB No. 2023–0170) and an informed consent was waived because the data were obtained from a public data depository which is freely available online. S1 Table presents the demographic characteristics of the sample participants.

The study examined the baseline data for both cohorts for CMI based on gender and separately analyzed the data for LMI from the baseline (2018) to the third survey year (2020) for each cohort. A total of 2,607 participants responded at baseline in 2018 for the elementary school sample; however, 170 (6.5%) and 110 (4.2%) participants dropped out in 2019 and 2020, making respective totals of 2,437 and 2,327 responses. Complete responses were received from 2,590 participants for the middle school sample in 2018; however, 152 (5.9%) and 113 (4.4%) participants dropped out in 2019 and 2020, making respective totals of 2,437 and 2,325 responses.

Measures

The Grit-S [10] is a short version of the full Grit-O developed by Duckworth et al. [1]. The Grit-S evaluates trait-level perseverance and passion for long-term goals using two factors, CI and PE, each of which comprises four items. In the Korean version, each item of the self-reported Grit-S is rated using a four-point Likert-type scale ranging from 1 = not at all like me to 4 = very much like me.

Kim and Hwang [11] reported on the measurement translation procedures and measurement properties, including the reliability and validity, of the Korean Grit-S version. The authors translated and back-translated the scale and made the necessary corrections and revisions to derive the final Korean version, the reliability of which was assessed using Cronbach’s α (.714). The scale was also found to have concurrent validity, which confirmed its positive and significant correlations with academic motivation (r = .51) and self-control (r = .61).

Data analysis

This study had four main objectives: (a) to explore the factor structure of the Grit-S using Korean elementary and middle school samples, (b) to evaluate the CMI of the Grit-S across genders, (c) to examine the LMI of the Grit-S over time, and (d) to investigate the reliability of the Grit-S scale. To achieve the first objective, exploratory factor analysis (EFA) was first conducted to explore the Grit-S factor structure without assuming any prior factor structure knowledge. Then CFA was conducted to determine whether the solution obtained from the EFA was retained. Due to the ordinal nature of the items, the EFA and CFA were conducted using the UWLS estimator with a polychoric correlation matrix.

To ensure the robustness of the findings, the elementary and middle school samples were randomly divided into two subsamples at time 1 (elementary school sample 1: N = 1,163; sample 2: N = 1,164; middle school sample 1: N = 1,162; sample 2: N = 1,163), after which the EFA and CFA were conducted on each subsample. Due to the inter-factor correlations, geomin rotation was used for the EFA to facilitate interpretability of solutions. Furthermore, parallel analysis was employed to determine the optimal number of factors to be extracted, which has been recognized as a suitable method for evaluating dimensionality in ordinal data with polychoric correlations [43].

A graphical technique called exploratory graph analysis (EGA) was adopted to validate the Grit-S factor structure’s EFA [44]. EGA is a network-based method to determine the dimensional structure of given items. Specifically, the Gaussian graphical model with polychoric correlations was adopted for the given items to establish a network, which was estimated using graphical least absolute shrinkage (GLASSO) with an extended Bayesian information criteria model selection operator owing to its capability to handle ordinal data [45]. The Louvain algorithm was applied to determine the optimal dimensions of the eight items, which provided information on which items belonged to each dimension [46]. The derived dimensions enabled a comparison of the EFA solution. A bootstrapping procedure was employed to assess the items’ stability and dimensional stability. Specifically, to assess whether the solution previously obtained was maintained, the item community assignments were estimated using 1,000 bootstrapped samples to determine which item belonged to each item community and the number of dimensions. The procedures were performed using the EGAnet package and the network visualization qgraph package [47].

Several fit indices were employed for the CFA to compare the goodness-of-fit of the proposed model, including the comparative fit index (CFI), the root mean square error of approximation (RMSEA), and the standardized root mean square residual (SRMR). The incremental fit measures, including a CFI greater than 0.95, were considered a good fit, with absolute fit measures, such as the RMSEA and SRMR, close to 0.06 indicating a reasonable fit [48].

Recent methodological studies have found that the fit index cutoff proposed by Hu and Bentler [48] could not be universally applied as a strict rule [49, 50], particularly when the UWLS or diagonally weighted least squares estimator are utilized to analyze ordered categorical data. In such cases, the conventional cutoff values exhibited worse performances in distinguishing misspecified models that fit poorly and correctly specified models that fit well, which led to models with significant misfits being considered acceptable [51–53]. To overcome such problematic performances of cutoffs with the categorical fit indices, the evaluation of the CFA models was further supplemented by generating the cutoffs with a dynamic fit index (DFI), which considered the specific characteristics of the models being fitted to derive the appropriate cutoff values [54]. R package dynamic [55] was used to customize the cutoff values that accommodated the ordered or categorical responses.

Subsequently, a multigroup CFA using the UWLS estimator was conducted to examine the MI of the Grit-S across genders at the baseline following the guidelines by Wu & Estabrook [41]. Specifically, a series of nested models were sequentially compared by constraining the model parameters to be equal across genders at each invariance test. The configural model was first estimated by freely estimating all parameters, followed by the model with equality constraints on the thresholds across gender, suggesting equivalent proportions for each response category. Subsequently, the model was estimated with equal thresholds and loadings, indicating similar measure for the constructs across gender. This approach was chosen over the conventional approach (e.g.: [56]) because the conventional approach, for which a baseline model is first established and then the constraints on loadings and thresholds are sequentially imposed, relies on how the baseline threshold structure is identified in relation to the latent response variable scales, which can potentially lead to misleading conclusions in MI tests.

Similarly, the LMI was then evaluated over time by equally constraining the model parameters across times in a series of nested models; configural, thresholds, and both threshold and loading invariance models. The MI was considered to be achieved if no significant differences were found in the fit indices between the more constrained and less constrained models. To determine the significance in the model fit changes, the Satorra–Bentler scaled chi-square difference test was utilized, which provided an appropriate evaluation for the differences between the nested models with ordered categorical indicators [57]. This approach computed the scaled chi-square statistics by dividing the standard chi-square statistics using scaling factors, which resulted in a better approximation of the chi-square distribution. A significant chi-square difference (Δχ²) indicated that the more constrained model had a poorer fit than the less constrained model.

In addition to the scaled chi-square difference test, a difference test based on the RMSEA value, referred to as RMSEA_D, was also implemented [58] to assess the significance of the model fit changes. As the interpretation of this index was consistent with the conventional RMSEA value, a value more than .08 indicated that the more constrained model had a worse fit than the less constrained model due to the introduced constraints within the model. The lavaan [59] and semTools [60] packages in R were utilized to implement these procedures.

The Grit-S reliability was examined using internal consistency measures and stability coefficients. The internal consistency at each time point was evaluated based on McDonald’s ω, and the stability coefficient, which represented the correlation between the factors at two time points, was evaluated using a threshold and loading invariance model to assess the relative stability of the Grit-S. The analyses were performed using Mplus 8.4 [61], and the EGA was conducted using R 4.1.1 [62].

Results

Factor analysis and network estimation

Table 1 shows the descriptive statistics; means, standard deviations, skewness, and kurtosis; for the Grit-S at the three time points.

Download:

Table 1. Descriptive statistics of the Grit-S.

https://doi.org/10.1371/journal.pone.0296795.t001

Prior to EFA, the Solomon method was used to assess the equivalence of the subsamples in each school, and the community ratio index (CRI) was calculated as a measure for the similarities between the subsamples, with a high value (close to unity) indicating a high degree of equivalence. The CRI values for the middle school and elementary school samples were close to unity (CRI: elementary school = .987; middle school = .961), which suggested that these subsamples were representative of the same populations.

In addition, the Kaiser–Meyer–Olkin (KMO) index was used to ensure the data appropriateness for the EFA on each subsample. All subsamples had values more than 0.6, which indicated that the data were suitable for performing the EFA (minimum KMO: elementary school = .763; middle school = .762).

The parallel analysis found that the two-factor model was the most appropriate for the elementary and middle school samples, which explains 43.2% and 41.7% of the total variance, respectively. Table 2 shows the standardized factor loadings and eigenvalues obtained from two-factor model. All individual items were loaded onto their corresponding constructs with loading values greater than .35. The estimated correlations between the two factors for the elementary and middle school samples were .47 and .40, respectively.

Download:

Table 2. Standardized factor loadings for two-factor model of the Grit-S at baseline.

https://doi.org/10.1371/journal.pone.0296795.t002

Fig 1 shows the estimated Grit-S network. For both samples, two item dimensions were identified using bootEGA, which was the same solution as the EFA. The results of bootstrap iterations allowed for an examination of the dimensional stability found that the two-dimension solution exhibited a 99.5% replication rate (995 out of 1,000 samples) in the elementary school sample and a 100.0% replication rate (1,000 out of 1,000 samples) in the middle school sample. All items loaded perfectly on their respective dimensions in both samples.

Download:

Fig 1. Estimated network models for the Grit-S at the first time in the elementary and middle school samples.

The panel on the left-hand side represents an elementary school’s network, and the panel on right-hand side represents a middle school’s network. In the estimated network, CI indicates consistency of interest factor and PE shows perseverance of the effect factor.

https://doi.org/10.1371/journal.pone.0296795.g001

The CFA revealed that the two-factor model selected using EFA and EGA was superior to the one‐factor model for both samples (elementary school: χ² = 103.5, df = 19, CFI = .975, SRMR = .050, RMSEA = .062; middle school: χ² = 121.3, df = 19, CFI = .968, SRMR = .054, RMSEA = .068), as shown in Table 3. Table 3 also reveals that the CFI value obtained from the two-factor model exceeded the strictest DFI cutoff levels (Level-1) in the elementary school sample even though the empirical values for SRMR and RMSEA were slightly higher than the cutoff values.

Download:

Table 3. Fit indices for the measurement models.

https://doi.org/10.1371/journal.pone.0296795.t003

All empirical fit indices deteriorated in comparison to the DFI cutoff in the middle school sample. Nevertheless, the two-factor model exhibited a relatively smaller discrepancy between the empirical fit measures and the dynamic fit measure cutoffs compared to the one-factor model. Based on these findings, the two-factor model was selected as the final model to adequately explain the Grit-S factor structure for both samples, and the following analyses were conducted using the two-factor model. S2 and S3 Tables present the CFA standardized factor loadings obtained for both samples.

Cross-Sectional measurement invariance

Table 4 displays the results of the CMI analyses across genders. The configural model showed an adequate fit to the data for both samples (elementary school: χ² = 483.670, df = 38, CFI = .947, SRMR = .051, RMSEA = .100; middle school: χ² = 564.517, df = 38, CFI = .927, SRMR = .056, RMSEA = .109). Subsequently, we established a thresholds invariance (TI) model by restricting the item thresholds to be equal across genders. The fit of the TI model did not deteriorate more than the cutoff values from the fit obtained using the configural model in the elementary school sample (Δχ² = 7.370, df = 8, p = .050; RMSEA_D. = 031), but the difference in the middle school sample was significant (Δχ² = 29.957, df = 8, p < .001; RMSEA_D. = 045). These results indicated that the threshold invariance across genders was supported only in the elementary school sample; however, we found a lack of full invariance for the thresholds in the middle school sample.

Download:

Table 4. Measurement invariance for the Grit-S across gender at baseline.

https://doi.org/10.1371/journal.pone.0296795.t004

Several studies have found that as partial invariance is sufficient for achieving MI if the proportion of the noninvariant items is small in a scale [63, 64], it can be accommodated to proceed the subsequent group comparisons [65–67]. The partial threshold invariance (PTI) model was formed by relaxing the equality constraint on the second thresholds for items 7 and 8, which exhibited the two largest modification index values. Table 4 demonstrates that the difference in fit measures between the configural and PTI model was nonsignificant (Δχ² = 6.5003, df = 6, p = .370; RMSEA_D. = 022), that is, the PTI was supported for the middle school sample.

The threshold and loading invariance (T&LI) models were established by imposing constraints on both thresholds and factor loadings to be equal across genders. The difference in the fit index between the threshold invariance model (full and partial) and T&LI models was significant for the middle school sample (Δχ² = 39.561, df = 6, p < .001; RMSEA_D = .057). However, the difference between the two models was nonsignificant for the elementary school sample (Δχ² = 6.878, df = 6, p = .332; RMSEA_D = .032). These results indicated that the full threshold and loading invariance across genders was supported only in the elementary school sample, but was not maintained in the middle school sample.

Therefore, the partial thresholds and loading invariance (PT&LI) model was built by relaxing the equality constraint of factor loadings on items 1 (New ideas and projects sometimes distract me from previous ones) and 5 (I often set a goal but later choose to pursue a different one), both of which had the two largest loading differences between genders in the configural model. Table 4 reveals that the difference in fit measures between the PTI and PT&LI models was nonsignificant (Δχ² = 14.561, df = 4, p = .057; RMSEA_D = .044), which indicates that partial thresholds and loading invariance was supported for the middle school sample.

Longitudinal measurement invariance

Table 5 lists the results of the LMI. First, model fit was separately assessed at each time point. For both samples, all models displayed a satisfactory fit with the CFI, SRMR, and RMSEA values, which enabled to perform a further test of the LMI.

Download:

Table 5. Longitudinal measurement invariance for the Grit-S over time.

https://doi.org/10.1371/journal.pone.0296795.t005

Table 5 also indicates that the differences between the configural and TI models were statistically significant for both samples (elementary school: Δχ² = 37.667, df = 16, p = .002; RMSEA_D = .031; middle school: Δχ² = 29.328, df = 16, p = .022; RMSEA_D = .025), which suggested that the full threshold invariance was not supported for both samples. The PTI model was constructed by releasing the equality constraint over time for the second thresholds of item 1 at time 2 and 3 and item 4 at time 3 in the elementary school sample and the second thresholds of item 7 in the middle school sample. No significant differences were found between the configural and PTI models for both samples (elementary school: Δχ² = 21.149, df = 13, p = .070; RMSEA_D = .023; middle school: Δχ² = 24.138, df = 15, p = .063; RMSEA_D = .022), which indicated that the threshold invariance was partially supported for both samples.

Table 5 also shows that the differences in the fit index between the PTI and T&LI models were significant for both school samples (elementary school: Δχ² = 26.220, df = 12, p = .009; RMSEA_D = .038; middle school: Δχ² = 29.472, df = 12, p < .001; RMSEA_D = .046). This finding implied that the factor loadings did not remain invariant over time for both samples. Therefore, the PT&LI models were constructed by freely estimating the factor loadings for item 3 at time 2 in the elementary school sample and for item 5 at time 2 in the middle school sample.

It was found that the differences in the fit index between the partial metric and partial scalar models were nonsignificant for both samples (elementary school: Δχ² = 15.478, df = 11, p = .161; RMSEA_D = .027; middle school: Δχ² = 17.359, df = 11, p = .098; RMSEA_D = .032), which indicated that the partial T&LI was supported for both samples.

In summary, these findings suggested that the Grit-S exhibited LMI over time for the configuration, and partial threshold and loading in both school samples. S4 and S5 Tables show the standardized factor loadings for the longitudinal invariance model.

Reliability analyses

The internal consistency evaluated using McDonald’s ω was acceptable (ω > 0.70) at all time points, with the ω coefficients respectively being .716, .700, and .712 at times 1, 2, and 3 in the elementary school sample, and .716, .700, and .712 at the same time points in the middle school sample.

The partial T&LI models were employed to examine the stability of the factors over time. The estimated factor correlations between times 1 and 2 were .403 and .474 for CI and PE in the elementary school sample and .519 and .562 in the middle school sample, all of which were statistically significant (p < .001). These findings suggested that the internal consistencies of the three factors were generally acceptable and that the factors demonstrated moderate stability over time.

Discussion

The major findings of this study were as follows. First, the CFA results indicated that the Korean version of the Grit-S had a two-factor structure in both the elementary and middle school samples. Second, the CMI analyses across genders indicated that the full and partial thresholds and the loading invariance were supported in both the elementary and middle school samples. Third, the LMI results revealed that the scale had partial T&LI in both samples. Finally, the scale reliability as measured by McDonald’s ω was acceptable, and the factors had moderate stability over time.

This study provides a more comprehensive understanding of the applicability of the Grit-S. First, the Grit-S was found to have a two-factor structure in both the Korean elementary and middle school samples. Because of the diverse Grit-S factor structures in previous studies, it was suggested that the Grit-S factor structure needed to be assessed in different cultures and age groups [27, 31]. Therefore, this study addressed these research gaps for early adolescent samples.

The results of this study were inconsistent with the Grit-S EFA results (i.e., the unidimensional model) from a study on early adolescents in the U.S. [68]. However, it has been suggested that the grit factor structure could differ across cultures due to the diverse values and beliefs in specific cultural groups [28, 29]. For example, individualist cultures highlight self-set goals whereas collectivist cultures tend to prioritize interpersonal harmony and the pursuit of group goals over individual goals [69, 70]. Due to a tendency to set goals that align with and adapt to significant others, students in collectivistic cultures may place less emphasis on consistency of interest than on persistence of effort in the two grit components [28], that is, people from collectivistic cultures may perceive grit as comprising two distinct but correlated constructs rather than seeing it as a combination of both CI and PE. Although further research is necessary, this study’s findings contribute to the ongoing discussion regarding cross-cultural differences in the grit structure.

Second, studies have agreed that a meaningful comparison of factors or observed means is possible when the scalar invariance level (factor loadings and intercepts) is supported [65, 66, 71, 72]. This verified that the Korean version of the scale had approximate scalar cross-sectional and longitudinal invariance in both school samples. Some studies have suggested that a meaningful comparison of the scale factor means is possible when partial invariance is supported [66, 71, 72] when the proportion of noninvariant items in the scale is small [63, 64]. As this study found sufficient evidence of configural, threshold, and loading invariance in the two MI aspects that corresponds to the scalar invariance, it was concluded that the Korean version of the Grit-S satisfied the CMI and the LMI in both samples. Therefore, as the study’s findings aligned with the reliability analyses, the Korean version of the Grit-S was deemed to have satisfactory psychometric properties for Korean adolescents. These results also indicated that the observed gender differences in the scale and the changes in the scale over time could be interpreted as actual differences and changes in the measured construct.

This study also highlighted relevant practical implications for future research. First, as the sample data size was sufficiently large and the data were collected using probability sampling, generalizability is possible to the Korean adolescent population. Second, the majority of previous studies tended to focus on separately reporting the CMI or LMI. To the best of our knowledge, this study is the first to simultaneously examine the CMI and LMI of the Grit-S based on an estimator that analyzed the ordered categorical data, which is an important contribution because it confirmed that the gender differences in the obtained scores reflect true gender differences and the increases in the observed scores over time were actual developmental changes in the given grit construct.

Regardless of the novel findings, this study had a few limitations. First, the samples were collected only from elementary and middle school settings. Therefore, future studies should examine adult samples to enhance the generalizability of the findings. Second, the psychometric properties of the scale were only investigated in a collectivist culture. Future studies should, therefore, seek to identify the cross-cultural differences in other collectivist and individualist societies. Third, attrition is inevitable in longitudinal research, and this study was not exempt. In the baseline model, 2,607 elementary students responded; however, this decreased to 2,327 in the third year (a dropout rate of approximately 10.7%). At baseline, 2,590 middle school students participated; however, this decreased to 2,325 in the third year (a dropout rate of approximately 10.2%). Whether the participants who remained differed significantly from those who dropped out is unknown. As missing cases can lead to problems in longitudinal studies, additional analyses were conducted to examine the impacts of the missing data, which were input and then used to create complete datasets using the expectation-maximization algorithm. The results from these newly created datasets were compared with those of the main study and were found to be similar, which implied that the missing data had not had any serious influence on the overall conclusions.

Supporting information

S1 Table. Demographic characteristics (elementary: N = 2,327; middle: N = 2,325).

https://doi.org/10.1371/journal.pone.0296795.s001

(DOCX)

S2 Table. Standardized factor loadings for two-factor model of the Grit-S in elementary school at baseline.

https://doi.org/10.1371/journal.pone.0296795.s002

(DOCX)

S3 Table. Standardized factor loadings for the two-factor model of the Grit-S in middle school at baseline.

https://doi.org/10.1371/journal.pone.0296795.s003

(DOCX)

S4 Table. Standardized factor loadings in the configural invariance model of the Grit-S.

https://doi.org/10.1371/journal.pone.0296795.s004

(DOCX)

S5 Table. Standardized factor loadings in the configural invariance model of the Grit-S.

https://doi.org/10.1371/journal.pone.0296795.s005

(DOCX)

References

1. Duckworth AL, Peterson C, Matthews MD, Kelly DR. Grit: Perseverance and passion for long-term goals. Journal of Personality and Social Psychology. 2007;92(6):1087–1101. pmid:17547490
- View Article
- PubMed/NCBI
- Google Scholar
2. Datu JAD. Beyond passion and perseverance: Review and future research initiatives on the science of grit. Frontiers in Psychology. 2021;11:1–17. pmid:33584397
- View Article
- PubMed/NCBI
- Google Scholar
3. Datu JAD, Yuen M, Chen G. Grit and determination: A review of literature with implications for theory and research. Journal of Psychologists and Counsellors in Schools. 2017;27(2):168–76.
- View Article
- Google Scholar
4. Fernández Martín FD, Arco Tirado JL, Hervás Torres M. Grit as a predictor and outcome of educational, professional, and personal success: A systematic Review. 2020;26(2):166–73.
- View Article
- Google Scholar
5. Bowman NA, Hill PL, Denson N, Bronkema R. Keep on truckin’or stay the course? exploring grit dimensions as differential predictors of educational achievement, satisfaction, and intentions. Social Psychological and Personality Science. 2015;6(6):639–45.
- View Article
- Google Scholar
6. Eskreis-Winkler L, Shulman EP, Beal SA, Duckworth AL. The grit effect: Predicting retention in the military, the workplace, school and marriage. Frontiers in Psychology. 2014;5:1–12. pmid:24550863
- View Article
- PubMed/NCBI
- Google Scholar
7. Sheldon KM, Jose PE, Kashdan TB, Jarden A. Personality, effective goal-striving, and enhanced well-being: Comparing 10 candidate personality strengths. Personality and Social Psychology Bulletin. 2015;41(4):575–85. pmid:25713170
- View Article
- PubMed/NCBI
- Google Scholar
8. Jin B, Kim J. Grit, basic needs satisfaction, and subjective well-being. Journal of Individual Differences. 2017;38(1):29–35.
- View Article
- Google Scholar
9. Kleiman EM, Adams LM, Kashdan TB, Riskind JH. Gratitude and grit indirectly reduce risk of suicidal ideations by enhancing meaning in life: Evidence for a mediated moderation model. Journal of Research in Personality. 2013;47(5):539–46.
- View Article
- Google Scholar
10. Duckworth AL, Quinn PD. Development and validation of the short grit scale (GRIT–S). Journal of Personality Assessment. 2009;91(2):166–74. pmid:19205937
- View Article
- PubMed/NCBI
- Google Scholar
11. Kim HM, Hwang MH. Validation of the Korean grit scale for children. The Journal of Education. 2015;35(3):63–74.
- View Article
- Google Scholar
12. Nishikawa K, Okugami S, Amemiya T. Development of the Japanese Short Grit Scale (Grit-S). Japanese Journal of Personality. 2015;24(2),167–9.
- View Article
- Google Scholar
13. Haktanir A, Lenz AS, Can N, Watson JC. Development and evaluation of Turkish language versions of three positive psychology assessments. International Journal for the Advancement of Counselling. 2016;38:286–97.
- View Article
- Google Scholar
14. Sarıçam H, Çelik İ, Oğuz A. Kisa azim (sebat) ölçeğinin türkçeye uyarlanmasi: geçerlik ve güvenirlik çalişmasi. International Journal of Turkish Literature Culture Education. 2016;5(2):927–935.
- View Article
- Google Scholar
15. Schmidt FTC, Fleckenstein J, Retelsdorf J, Eskreis-Winkler L, Möller J. Measuring grit. European Journal of Psychological Assessment. 2017;35(3):436–447.
- View Article
- Google Scholar
16. Wyszyńska P, Ponikiewska K, Karaś D, Najderska M, Rogoza R. Psychometric properties of the polish version of the short grit scale. Polish Psychological Bulletin. 2017(2):229–36.
- View Article
- Google Scholar
17. Sulla F, Renati R, Bonfiglio S, Rollo D. Italian students and the Grit-S: A self-report questionnaire for measuring perseverance and passion for long-term goals. IEEE International Symposium on Medical Measurements and Applications. 2018:1–5. https://doi.org/10.1109/MeMeA.2018.8438668
18. Arco-Tirado J, Fernández-Martín F,D., Hoyle RH. Development and validation of a Spanish version of the grit-S scale. Frontiers in Psychology. 2018;9:96. pmid:29467705
- View Article
- PubMed/NCBI
- Google Scholar
19. Marentes-Castillo M, Zamarripa J, Castillo I. Validation of the grit scale and the treatment self-regulation questionnaire (TSRQ) in the Mexican context. Revista Latinoamericana de Psicología. 2019;51(1):9–18.
- View Article
- Google Scholar
20. Kropáčová S, Slezáčková A, Jarden A. Škála vytrvalosti: Analýza psychometrických vlastností české verze pro dospělé. E-psychologie. 2018;12(2):27–40.
- View Article
- Google Scholar
21. Schmidt FTC, Sudzina F, Botek M. Psychometric assessment of the short grit scale among czech young adults. Journal of Psychoeducational Assessment. 2021;39(4):508–13.
- View Article
- Google Scholar
22. Li J, Zhao Y, Kong F, Du S, Yang S, Wang S. Psychometric assessment of the short grit scale among Chinese adolescents. Journal of Psychoeducational Assessment. 2018;36(3):291–6.
- View Article
- Google Scholar
23. Luo J, Wang M, Ge Y, Chen W, Xu S. Longitudinal invariance analysis of the short grit scale in Chinese young adults. Frontiers in Psychology. 2020;11:1–9. pmid:32265788
- View Article
- PubMed/NCBI
- Google Scholar
24. Zhong C, Wang M, Shou Y, Ren F, Zhang X, Li M, et al. Assessing construct validity of the grit-S in chinese employees. PLoS ONE. 2018;13(12):1–14. pmid:30576360
- View Article
- PubMed/NCBI
- Google Scholar
25. Shaban NM. Validation of grit scale in the Arabian context for Egyptian players. Ovidius University Annals, Series Physical Education & Sport/Science, Movement & Health. 2020;20(2):153–7.
- View Article
- Google Scholar
26. Frontini R, Monteiro D, Rodrigues F, Matos R, Antunes R. Adapting the short grit scale with exploratory structural equation modeling for Portuguese college students. Perceptual and Motor Skills. 2022;129(5):1428–42. pmid:35658701
- View Article
- PubMed/NCBI
- Google Scholar
27. Disabato DJ, Goodman FR, Kashdan TB. Is grit relevant to well‐being and strengths? evidence across the globe for separating perseverance of effort and consistency of interests. Journal of Personality. 2019;87(2):194–211. pmid:29573415
- View Article
- PubMed/NCBI
- Google Scholar
28. Datu JAD, Valdez JPM, King RB. Perseverance counts but consistency does not! validating the short grit scale in a collectivist setting. Current Psychology. 2016;35:121–30.
- View Article
- Google Scholar
29. Abu Hasan HE, Munawar K, Abdul Khaiyom JH. Psychometric properties of developed and transadapted grit measures across cultures: A systematic review. Current Psychology. 2020:1–19.
- View Article
- Google Scholar
30. Muenks K, Wigfield A, Yang JS, O’Neal C, R. How true is grit? assessing its relations to high school and college students’ personality characteristics, self-regulation, engagement, and achievement. Journal of Educational Psychology. 2017;109(5):599–620.
- View Article
- Google Scholar
31. Schmidt FTC, Fleckenstein J, Retelsdorf J, Eskreis-Winkler L, Möller J. Measuring grit: A German validation and a domain-specific approach to grit. European Journal of Psychological Assessment. 2019;35(3):436–447.
- View Article
- Google Scholar
32. Akos P, Kretchmar J. Investigating grit at a non-cognitive predictor of college success. The Review of Higher Education. 2017;40(2):163–86.
- View Article
- Google Scholar
33. Oriol X, Miranda R, Oyanedel JC, Torres J. The role of self-control and grit in domains of school success in students of primary and secondary school. Frontiers in psychology. 2017;8:1716. pmid:29075211
- View Article
- PubMed/NCBI
- Google Scholar
34. Clark KN, Dorio NB, Eldridge MA, Malecki CK, Demaray MK. Adolescent academic achievement: A model of social support and grit. Psychology in the Schools. 2020;57(2):204–21.
- View Article
- Google Scholar
35. Weng M, Liao C, Kwok O, Wu J. What is the best model of grit among junior high students: Model selection, measurement invariance, and group difference. Social Development. 2023;32(4).1117–33.
- View Article
- Google Scholar
36. Alan S, Boneva T, Ertac S. Ever failed, try again, succeed better: Results from a randomized educational intervention on grit. The Quarterly Journal of Economics. 2019;134(3):1121–62.
- View Article
- Google Scholar
37. Cosgrove JM, Brown KP, Beddoes ZE, Bartholomew JB, Castelli DM. Physical education and student grit. Research Quarterly for Exercise and Sport. 2016;87(S2):A64.
- View Article
- Google Scholar
38. Peña P A., Duckworth AL. The effects of relative and absolute age in the measurement of grit from 9th to 12th grade. Economics of Education Review. 2018;66:183–90.
- View Article
- Google Scholar
39. West MR, Kraft MA, Finn AS, Martin RE, Duckworth AL, Gabrieli CFO, et al. Promise and paradox: Measuring students’ non-cognitive skills and the impact of schooling. Educational Evaluation and Policy Analysis. 2016;38(1):148–70.
- View Article
- Google Scholar
40. Lim HJ, Lee S. How grit-S items work in measuring growth: Focusing on item discrimination and sensitivity to change. Journal of Psychoeducational Assessment. 2023;41(2):175–93.
- View Article
- Google Scholar
41. Wu H, Estabrook R. Identification of confirmatory factor analysis models of different levels of invariance for ordered categorical outcomes. Psychometrika. 2016;81(4):1014–45. pmid:27402166
- View Article
- PubMed/NCBI
- Google Scholar
42. Ha HS, Hwang JK, Kim SE, Lee YH. Korean children and youth panel survey: Annual report of year 2020(20-R12). National Youth Policy Institute. 2020;1–260. https://lib.nypi.re.kr/search/detail/CATTOT000000040179
- View Article
- Google Scholar
43. Garrido LE, Abad FJ, Ponsoda V. A new look at horn’s parallel analysis with ordinal variables. Psychological Methods. 2013;18(4):454. pmid:23046000
- View Article
- PubMed/NCBI
- Google Scholar
44. Golino HF, Epskamp S. Exploratory graph analysis: A new approach for estimating the number of dimensions in psychological research. PLoS ONE. 2017;12(6):e0174035. pmid:28594839
- View Article
- PubMed/NCBI
- Google Scholar
45. Epskamp S, Fried EI. A tutorial on regularized partial correlation networks. Psychological Methods. 2018;23(4):617–34. pmid:29595293
- View Article
- PubMed/NCBI
- Google Scholar
46. Golino H, Shi D, Christensen AP, Garrido LE, Nieto MD, Sadana R, et al. Investigating the performance of exploratory graph analysis and traditional techniques to identify the number of latent factors: A simulation and tutorial. Psychological Methods. 2020;25(3):292–320. pmid:32191105
- View Article
- PubMed/NCBI
- Google Scholar
47. Epskamp S, Cramer AOJ, Waldorp LJ, Schmittmann VD, Borsboom D. Qgraph: Network visualizations of relationships in psychometric data. Journal of Statistical Software. 2012;48:1–18.
- View Article
- Google Scholar
48. Hu L, Bentler PM. Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: a Multidisciplinary Journal. 1999;6(1):1–55.
- View Article
- Google Scholar
49. Greiff S, Heene M. Why psychological assessment needs to start worrying about model fit. Hogrefe Publishing. 2017;33(5);313–7.
- View Article
- Google Scholar
50. Shi D, Lee T, Maydeu-Olivares A. Understanding the model size effect on SEM fit indices. Educational and Psychological Measurement. 2019;79(2):310–34. pmid:30911195
- View Article
- PubMed/NCBI
- Google Scholar
51. Monroe S, Cai L. Evaluating structural equation models for categorical outcomes: A new test statistic and a practical challenge of interpretation. Multivariate Behavioral Research. 2015;50(6):569–83. pmid:26717119
- View Article
- PubMed/NCBI
- Google Scholar
52. Xia Y, Yang Y. The influence of number of categories and threshold values on fit indices in structural equation modeling with ordered categorical data. Multivariate Behavioral Research. 2018;53(5):731–55. pmid:30477318
- View Article
- PubMed/NCBI
- Google Scholar
53. Xia Y, Yang Y. RMSEA, CFI, and TLI in structural equation modeling with ordered categorical data: The story they tell depends on the estimation methods. Behavior Research Methods. 2019;51:409–28. pmid:29869222
- View Article
- PubMed/NCBI
- Google Scholar
54. McNeish D. Dynamic fit index cutoffs for factor analysis with likert, ordinal, or binary responses. PsyArXiv Preprints [Preprint]. 2023.: https://osf.io/preprints/psyarxiv/tp35s/
55. Wolf MG, McNeish D. Dynamic Model Fit. Version 1.1.0 [Software]. 2020. www.dynamicft.app
56. Millsap RE, Yun-Tein J. Assessing factorial invariance in ordered-categorical measures. Multivariate Behavioral Research. 2004;39(3):479–515. https://doi.org/10.1207/S15327906MBR3903_4
- View Article
- Google Scholar
57. Liu Y, Millsap RE, West SG, Tein J, Tanaka R, Grimm KJ. Testing measurement invariance in longitudinal data with ordered-categorical measures. Psychological Methods. 2017;22(3):486. pmid:27213981
- View Article
- PubMed/NCBI
- Google Scholar
58. Savalei V, Brace JC, Fouladi RT. We need to change how we compute RMSEA for nested model comparisons in structural equation modeling. Psychological Methods. 2023. pmid:36622720
- View Article
- PubMed/NCBI
- Google Scholar
59. Rosseel Y. Lavaan: An R package for structural equation modeling. Journal of Statistical Software. 2012;48:1–36.
- View Article
- Google Scholar
60. Jorgensen TD, Pornprasertmanit S, Schoemann AM, Rosseel Y. semTools: useful tools for structural equation modeling. R package. Version 0.5–1 [software]. 2018. https://CRAN.R-project.org/package=semTools
61. Muthén LK, Muthén BO. Mplus User’s Guide. 8th ed. Los Angeles; 1998–2017.
62. R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing. 2021.: https://www.R-project.org/
63. Cheung GW, Rensvold RB. Cross-cultural comparisons using non-invariant measurement items. Applied Behavioral Science Review. 1998;6(1):93–110.
- View Article
- Google Scholar
64. Cheung GW, Rensvold RB. Testing factorial invariance across groups: A reconceptualization and proposed new method. Journal of Management. 1999;25(1):1–27.
- View Article
- Google Scholar
65. Richardson GB, Smith R, Lowe L, Acquavita SP. Structure and longitudinal invariance of the short alcohol and alcohol problems perception questionnaire. Journal of Substance Abuse Treatment. 2020;115:108041. pmid:32600628
- View Article
- PubMed/NCBI
- Google Scholar
66. Seddig D, Leitgöb H. Approximate measurement invariance and longitudinal confirmatory factor analysis: Concept and application with panel data. Survey Research Methods. 2018;12(1):29–41.
- View Article
- Google Scholar
67. Svetina D, Rutkowski L, Rutkowski D. Multiple-group invariance with categorical outcomes using updated guidelines: An illustration using M plus and the lavaan/semtools packages. Structural Equation Modeling: A Multidisciplinary Journal. 2020;27(1):111–30.
- View Article
- Google Scholar
68. Usher EL, Li CR, Butz AR, Rojas JP. Perseverant grit and self-efficacy: Are both essential for children’s academic success? Journal of Educational Psychology. 2019;111(5):877.
- View Article
- Google Scholar
69. Kwan VSY, Bond MH, Singelis TM. Pancultural explanations for life satisfaction: Adding relationship harmony to self-esteem. Journal of Personality and Social Psychology. 1997;73(5):1038. pmid:9364759
- View Article
- PubMed/NCBI
- Google Scholar
70. Markus HR, Kitayama S. Culture and the self: Implications for cognition, emotion, and motivation. Psychological Review. 1991;98(2):224.
- View Article
- Google Scholar
71. Bowen NK, Masa RD. Conducting measurement invariance tests with ordinal data: A guide for social work researchers. Journal of the Society for Social Work and Research. 2015;6(2):229–49.
- View Article
- Google Scholar
72. Marsh HW, Guo J, Parker PD, Nagengast B, Asparouhov T, Muthén B, et al. What to do when scalar invariance fails: The extended alignment method for multi-group factor analysis comparison of latent means across many groups. Psychological Methods. 2018;23(3):524. pmid:28080078
- View Article
- PubMed/NCBI
- Google Scholar

[ref1] 1. Duckworth AL, Peterson C, Matthews MD, Kelly DR. Grit: Perseverance and passion for long-term goals. Journal of Personality and Social Psychology. 2007;92(6):1087–1101. pmid:17547490
View Article
PubMed/NCBI
Google Scholar

[2] View Article

[3] PubMed/NCBI

[4] Google Scholar

[ref2] 2. Datu JAD. Beyond passion and perseverance: Review and future research initiatives on the science of grit. Frontiers in Psychology. 2021;11:1–17. pmid:33584397
View Article
PubMed/NCBI
Google Scholar

[6] View Article

[7] PubMed/NCBI

[8] Google Scholar

[ref3] 3. Datu JAD, Yuen M, Chen G. Grit and determination: A review of literature with implications for theory and research. Journal of Psychologists and Counsellors in Schools. 2017;27(2):168–76.
View Article
Google Scholar

[10] View Article

[11] Google Scholar

[ref4] 4. Fernández Martín FD, Arco Tirado JL, Hervás Torres M. Grit as a predictor and outcome of educational, professional, and personal success: A systematic Review. 2020;26(2):166–73.
View Article
Google Scholar

[13] View Article

[14] Google Scholar

[ref5] 5. Bowman NA, Hill PL, Denson N, Bronkema R. Keep on truckin’or stay the course? exploring grit dimensions as differential predictors of educational achievement, satisfaction, and intentions. Social Psychological and Personality Science. 2015;6(6):639–45.
View Article
Google Scholar

[16] View Article

[17] Google Scholar

[ref6] 6. Eskreis-Winkler L, Shulman EP, Beal SA, Duckworth AL. The grit effect: Predicting retention in the military, the workplace, school and marriage. Frontiers in Psychology. 2014;5:1–12. pmid:24550863
View Article
PubMed/NCBI
Google Scholar

[19] View Article

[20] PubMed/NCBI

[21] Google Scholar

[ref7] 7. Sheldon KM, Jose PE, Kashdan TB, Jarden A. Personality, effective goal-striving, and enhanced well-being: Comparing 10 candidate personality strengths. Personality and Social Psychology Bulletin. 2015;41(4):575–85. pmid:25713170
View Article
PubMed/NCBI
Google Scholar

[23] View Article

[24] PubMed/NCBI

[25] Google Scholar

[ref8] 8. Jin B, Kim J. Grit, basic needs satisfaction, and subjective well-being. Journal of Individual Differences. 2017;38(1):29–35.
View Article
Google Scholar

[27] View Article

[28] Google Scholar

[ref9] 9. Kleiman EM, Adams LM, Kashdan TB, Riskind JH. Gratitude and grit indirectly reduce risk of suicidal ideations by enhancing meaning in life: Evidence for a mediated moderation model. Journal of Research in Personality. 2013;47(5):539–46.
View Article
Google Scholar

[30] View Article

[31] Google Scholar

[ref10] 10. Duckworth AL, Quinn PD. Development and validation of the short grit scale (GRIT–S). Journal of Personality Assessment. 2009;91(2):166–74. pmid:19205937
View Article
PubMed/NCBI
Google Scholar

[33] View Article

[34] PubMed/NCBI

[35] Google Scholar

[ref11] 11. Kim HM, Hwang MH. Validation of the Korean grit scale for children. The Journal of Education. 2015;35(3):63–74.
View Article
Google Scholar

[37] View Article

[38] Google Scholar

[ref12] 12. Nishikawa K, Okugami S, Amemiya T. Development of the Japanese Short Grit Scale (Grit-S). Japanese Journal of Personality. 2015;24(2),167–9.
View Article
Google Scholar

[40] View Article

[41] Google Scholar

[ref13] 13. Haktanir A, Lenz AS, Can N, Watson JC. Development and evaluation of Turkish language versions of three positive psychology assessments. International Journal for the Advancement of Counselling. 2016;38:286–97.
View Article
Google Scholar

[43] View Article

[44] Google Scholar

[ref14] 14. Sarıçam H, Çelik İ, Oğuz A. Kisa azim (sebat) ölçeğinin türkçeye uyarlanmasi: geçerlik ve güvenirlik çalişmasi. International Journal of Turkish Literature Culture Education. 2016;5(2):927–935.
View Article
Google Scholar

[46] View Article

[47] Google Scholar

[ref15] 15. Schmidt FTC, Fleckenstein J, Retelsdorf J, Eskreis-Winkler L, Möller J. Measuring grit. European Journal of Psychological Assessment. 2017;35(3):436–447.
View Article
Google Scholar

[49] View Article

[50] Google Scholar

[ref16] 16. Wyszyńska P, Ponikiewska K, Karaś D, Najderska M, Rogoza R. Psychometric properties of the polish version of the short grit scale. Polish Psychological Bulletin. 2017(2):229–36.
View Article
Google Scholar

[52] View Article

[53] Google Scholar

[ref17] 17. Sulla F, Renati R, Bonfiglio S, Rollo D. Italian students and the Grit-S: A self-report questionnaire for measuring perseverance and passion for long-term goals. IEEE International Symposium on Medical Measurements and Applications. 2018:1–5. https://doi.org/10.1109/MeMeA.2018.8438668

[ref18] 18. Arco-Tirado J, Fernández-Martín F,D., Hoyle RH. Development and validation of a Spanish version of the grit-S scale. Frontiers in Psychology. 2018;9:96. pmid:29467705
View Article
PubMed/NCBI
Google Scholar

[56] View Article

[57] PubMed/NCBI

[58] Google Scholar

[ref19] 19. Marentes-Castillo M, Zamarripa J, Castillo I. Validation of the grit scale and the treatment self-regulation questionnaire (TSRQ) in the Mexican context. Revista Latinoamericana de Psicología. 2019;51(1):9–18.
View Article
Google Scholar

[60] View Article

[61] Google Scholar

[ref20] 20. Kropáčová S, Slezáčková A, Jarden A. Škála vytrvalosti: Analýza psychometrických vlastností české verze pro dospělé. E-psychologie. 2018;12(2):27–40.
View Article
Google Scholar

[63] View Article

[64] Google Scholar

[ref21] 21. Schmidt FTC, Sudzina F, Botek M. Psychometric assessment of the short grit scale among czech young adults. Journal of Psychoeducational Assessment. 2021;39(4):508–13.
View Article
Google Scholar

[66] View Article

[67] Google Scholar

[ref22] 22. Li J, Zhao Y, Kong F, Du S, Yang S, Wang S. Psychometric assessment of the short grit scale among Chinese adolescents. Journal of Psychoeducational Assessment. 2018;36(3):291–6.
View Article
Google Scholar

[69] View Article

[70] Google Scholar

[ref23] 23. Luo J, Wang M, Ge Y, Chen W, Xu S. Longitudinal invariance analysis of the short grit scale in Chinese young adults. Frontiers in Psychology. 2020;11:1–9. pmid:32265788
View Article
PubMed/NCBI
Google Scholar

[72] View Article

[73] PubMed/NCBI

[74] Google Scholar

[ref24] 24. Zhong C, Wang M, Shou Y, Ren F, Zhang X, Li M, et al. Assessing construct validity of the grit-S in chinese employees. PLoS ONE. 2018;13(12):1–14. pmid:30576360
View Article
PubMed/NCBI
Google Scholar

[76] View Article

[77] PubMed/NCBI

[78] Google Scholar

[ref25] 25. Shaban NM. Validation of grit scale in the Arabian context for Egyptian players. Ovidius University Annals, Series Physical Education & Sport/Science, Movement & Health. 2020;20(2):153–7.
View Article
Google Scholar

[80] View Article

[81] Google Scholar

[ref26] 26. Frontini R, Monteiro D, Rodrigues F, Matos R, Antunes R. Adapting the short grit scale with exploratory structural equation modeling for Portuguese college students. Perceptual and Motor Skills. 2022;129(5):1428–42. pmid:35658701
View Article
PubMed/NCBI
Google Scholar

[83] View Article

[84] PubMed/NCBI

[85] Google Scholar

[ref27] 27. Disabato DJ, Goodman FR, Kashdan TB. Is grit relevant to well‐being and strengths? evidence across the globe for separating perseverance of effort and consistency of interests. Journal of Personality. 2019;87(2):194–211. pmid:29573415
View Article
PubMed/NCBI
Google Scholar

[87] View Article

[88] PubMed/NCBI

[89] Google Scholar

[ref28] 28. Datu JAD, Valdez JPM, King RB. Perseverance counts but consistency does not! validating the short grit scale in a collectivist setting. Current Psychology. 2016;35:121–30.
View Article
Google Scholar

[91] View Article

[92] Google Scholar

[ref29] 29. Abu Hasan HE, Munawar K, Abdul Khaiyom JH. Psychometric properties of developed and transadapted grit measures across cultures: A systematic review. Current Psychology. 2020:1–19.
View Article
Google Scholar

[94] View Article

[95] Google Scholar

[ref30] 30. Muenks K, Wigfield A, Yang JS, O’Neal C, R. How true is grit? assessing its relations to high school and college students’ personality characteristics, self-regulation, engagement, and achievement. Journal of Educational Psychology. 2017;109(5):599–620.
View Article
Google Scholar

[97] View Article

[98] Google Scholar

[ref31] 31. Schmidt FTC, Fleckenstein J, Retelsdorf J, Eskreis-Winkler L, Möller J. Measuring grit: A German validation and a domain-specific approach to grit. European Journal of Psychological Assessment. 2019;35(3):436–447.
View Article
Google Scholar

[100] View Article

[101] Google Scholar

[ref32] 32. Akos P, Kretchmar J. Investigating grit at a non-cognitive predictor of college success. The Review of Higher Education. 2017;40(2):163–86.
View Article
Google Scholar

[103] View Article

[104] Google Scholar

[ref33] 33. Oriol X, Miranda R, Oyanedel JC, Torres J. The role of self-control and grit in domains of school success in students of primary and secondary school. Frontiers in psychology. 2017;8:1716. pmid:29075211
View Article
PubMed/NCBI
Google Scholar

[106] View Article

[107] PubMed/NCBI

[108] Google Scholar

[ref34] 34. Clark KN, Dorio NB, Eldridge MA, Malecki CK, Demaray MK. Adolescent academic achievement: A model of social support and grit. Psychology in the Schools. 2020;57(2):204–21.
View Article
Google Scholar

[110] View Article

[111] Google Scholar

[ref35] 35. Weng M, Liao C, Kwok O, Wu J. What is the best model of grit among junior high students: Model selection, measurement invariance, and group difference. Social Development. 2023;32(4).1117–33.
View Article
Google Scholar

[113] View Article

[114] Google Scholar

[ref36] 36. Alan S, Boneva T, Ertac S. Ever failed, try again, succeed better: Results from a randomized educational intervention on grit. The Quarterly Journal of Economics. 2019;134(3):1121–62.
View Article
Google Scholar

[116] View Article

[117] Google Scholar

[ref37] 37. Cosgrove JM, Brown KP, Beddoes ZE, Bartholomew JB, Castelli DM. Physical education and student grit. Research Quarterly for Exercise and Sport. 2016;87(S2):A64.
View Article
Google Scholar

[119] View Article

[120] Google Scholar

[ref38] 38. Peña P A., Duckworth AL. The effects of relative and absolute age in the measurement of grit from 9th to 12th grade. Economics of Education Review. 2018;66:183–90.
View Article
Google Scholar

[122] View Article

[123] Google Scholar

[ref39] 39. West MR, Kraft MA, Finn AS, Martin RE, Duckworth AL, Gabrieli CFO, et al. Promise and paradox: Measuring students’ non-cognitive skills and the impact of schooling. Educational Evaluation and Policy Analysis. 2016;38(1):148–70.
View Article
Google Scholar

[125] View Article

[126] Google Scholar

[ref40] 40. Lim HJ, Lee S. How grit-S items work in measuring growth: Focusing on item discrimination and sensitivity to change. Journal of Psychoeducational Assessment. 2023;41(2):175–93.
View Article
Google Scholar

[128] View Article

[129] Google Scholar

[ref41] 41. Wu H, Estabrook R. Identification of confirmatory factor analysis models of different levels of invariance for ordered categorical outcomes. Psychometrika. 2016;81(4):1014–45. pmid:27402166
View Article
PubMed/NCBI
Google Scholar

[131] View Article

[132] PubMed/NCBI

[133] Google Scholar

[ref42] 42. Ha HS, Hwang JK, Kim SE, Lee YH. Korean children and youth panel survey: Annual report of year 2020(20-R12). National Youth Policy Institute. 2020;1–260. https://lib.nypi.re.kr/search/detail/CATTOT000000040179
View Article
Google Scholar

[135] View Article

[136] Google Scholar

[ref43] 43. Garrido LE, Abad FJ, Ponsoda V. A new look at horn’s parallel analysis with ordinal variables. Psychological Methods. 2013;18(4):454. pmid:23046000
View Article
PubMed/NCBI
Google Scholar

[138] View Article

[139] PubMed/NCBI

[140] Google Scholar

[ref44] 44. Golino HF, Epskamp S. Exploratory graph analysis: A new approach for estimating the number of dimensions in psychological research. PLoS ONE. 2017;12(6):e0174035. pmid:28594839
View Article
PubMed/NCBI
Google Scholar

[142] View Article

[143] PubMed/NCBI

[144] Google Scholar

[ref45] 45. Epskamp S, Fried EI. A tutorial on regularized partial correlation networks. Psychological Methods. 2018;23(4):617–34. pmid:29595293
View Article
PubMed/NCBI
Google Scholar

[146] View Article

[147] PubMed/NCBI

[148] Google Scholar

[ref46] 46. Golino H, Shi D, Christensen AP, Garrido LE, Nieto MD, Sadana R, et al. Investigating the performance of exploratory graph analysis and traditional techniques to identify the number of latent factors: A simulation and tutorial. Psychological Methods. 2020;25(3):292–320. pmid:32191105
View Article
PubMed/NCBI
Google Scholar

[150] View Article

[151] PubMed/NCBI

[152] Google Scholar

[ref47] 47. Epskamp S, Cramer AOJ, Waldorp LJ, Schmittmann VD, Borsboom D. Qgraph: Network visualizations of relationships in psychometric data. Journal of Statistical Software. 2012;48:1–18.
View Article
Google Scholar

[154] View Article

[155] Google Scholar

[ref48] 48. Hu L, Bentler PM. Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: a Multidisciplinary Journal. 1999;6(1):1–55.
View Article
Google Scholar

[157] View Article

[158] Google Scholar

[ref49] 49. Greiff S, Heene M. Why psychological assessment needs to start worrying about model fit. Hogrefe Publishing. 2017;33(5);313–7.
View Article
Google Scholar

[160] View Article

[161] Google Scholar

[ref50] 50. Shi D, Lee T, Maydeu-Olivares A. Understanding the model size effect on SEM fit indices. Educational and Psychological Measurement. 2019;79(2):310–34. pmid:30911195
View Article
PubMed/NCBI
Google Scholar

[163] View Article

[164] PubMed/NCBI

[165] Google Scholar

[ref51] 51. Monroe S, Cai L. Evaluating structural equation models for categorical outcomes: A new test statistic and a practical challenge of interpretation. Multivariate Behavioral Research. 2015;50(6):569–83. pmid:26717119
View Article
PubMed/NCBI
Google Scholar

[167] View Article

[168] PubMed/NCBI

[169] Google Scholar

[ref52] 52. Xia Y, Yang Y. The influence of number of categories and threshold values on fit indices in structural equation modeling with ordered categorical data. Multivariate Behavioral Research. 2018;53(5):731–55. pmid:30477318
View Article
PubMed/NCBI
Google Scholar

[171] View Article

[172] PubMed/NCBI

[173] Google Scholar

[ref53] 53. Xia Y, Yang Y. RMSEA, CFI, and TLI in structural equation modeling with ordered categorical data: The story they tell depends on the estimation methods. Behavior Research Methods. 2019;51:409–28. pmid:29869222
View Article
PubMed/NCBI
Google Scholar

[175] View Article

[176] PubMed/NCBI

[177] Google Scholar

[ref54] 54. McNeish D. Dynamic fit index cutoffs for factor analysis with likert, ordinal, or binary responses. PsyArXiv Preprints [Preprint]. 2023.: https://osf.io/preprints/psyarxiv/tp35s/

[ref55] 55. Wolf MG, McNeish D. Dynamic Model Fit. Version 1.1.0 [Software]. 2020. www.dynamicft.app

[ref56] 56. Millsap RE, Yun-Tein J. Assessing factorial invariance in ordered-categorical measures. Multivariate Behavioral Research. 2004;39(3):479–515. https://doi.org/10.1207/S15327906MBR3903_4
View Article
Google Scholar

[181] View Article

[182] Google Scholar

[ref57] 57. Liu Y, Millsap RE, West SG, Tein J, Tanaka R, Grimm KJ. Testing measurement invariance in longitudinal data with ordered-categorical measures. Psychological Methods. 2017;22(3):486. pmid:27213981
View Article
PubMed/NCBI
Google Scholar

[184] View Article

[185] PubMed/NCBI

[186] Google Scholar

[ref58] 58. Savalei V, Brace JC, Fouladi RT. We need to change how we compute RMSEA for nested model comparisons in structural equation modeling. Psychological Methods. 2023. pmid:36622720
View Article
PubMed/NCBI
Google Scholar

[188] View Article

[189] PubMed/NCBI

[190] Google Scholar

[ref59] 59. Rosseel Y. Lavaan: An R package for structural equation modeling. Journal of Statistical Software. 2012;48:1–36.
View Article
Google Scholar

[192] View Article

[193] Google Scholar

[ref60] 60. Jorgensen TD, Pornprasertmanit S, Schoemann AM, Rosseel Y. semTools: useful tools for structural equation modeling. R package. Version 0.5–1 [software]. 2018. https://CRAN.R-project.org/package=semTools

[ref61] 61. Muthén LK, Muthén BO. Mplus User’s Guide. 8th ed. Los Angeles; 1998–2017.

[ref62] 62. R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing. 2021.: https://www.R-project.org/

[ref63] 63. Cheung GW, Rensvold RB. Cross-cultural comparisons using non-invariant measurement items. Applied Behavioral Science Review. 1998;6(1):93–110.
View Article
Google Scholar

[198] View Article

[199] Google Scholar

[ref64] 64. Cheung GW, Rensvold RB. Testing factorial invariance across groups: A reconceptualization and proposed new method. Journal of Management. 1999;25(1):1–27.
View Article
Google Scholar

[201] View Article

[202] Google Scholar

[ref65] 65. Richardson GB, Smith R, Lowe L, Acquavita SP. Structure and longitudinal invariance of the short alcohol and alcohol problems perception questionnaire. Journal of Substance Abuse Treatment. 2020;115:108041. pmid:32600628
View Article
PubMed/NCBI
Google Scholar

[204] View Article

[205] PubMed/NCBI

[206] Google Scholar

[ref66] 66. Seddig D, Leitgöb H. Approximate measurement invariance and longitudinal confirmatory factor analysis: Concept and application with panel data. Survey Research Methods. 2018;12(1):29–41.
View Article
Google Scholar

[208] View Article

[209] Google Scholar

[ref67] 67. Svetina D, Rutkowski L, Rutkowski D. Multiple-group invariance with categorical outcomes using updated guidelines: An illustration using M plus and the lavaan/semtools packages. Structural Equation Modeling: A Multidisciplinary Journal. 2020;27(1):111–30.
View Article
Google Scholar

[211] View Article

[212] Google Scholar

[ref68] 68. Usher EL, Li CR, Butz AR, Rojas JP. Perseverant grit and self-efficacy: Are both essential for children’s academic success? Journal of Educational Psychology. 2019;111(5):877.
View Article
Google Scholar

[214] View Article

[215] Google Scholar

[ref69] 69. Kwan VSY, Bond MH, Singelis TM. Pancultural explanations for life satisfaction: Adding relationship harmony to self-esteem. Journal of Personality and Social Psychology. 1997;73(5):1038. pmid:9364759
View Article
PubMed/NCBI
Google Scholar

[217] View Article

[218] PubMed/NCBI

[219] Google Scholar

[ref70] 70. Markus HR, Kitayama S. Culture and the self: Implications for cognition, emotion, and motivation. Psychological Review. 1991;98(2):224.
View Article
Google Scholar

[221] View Article

[222] Google Scholar

[ref71] 71. Bowen NK, Masa RD. Conducting measurement invariance tests with ordinal data: A guide for social work researchers. Journal of the Society for Social Work and Research. 2015;6(2):229–49.
View Article
Google Scholar

[224] View Article

[225] Google Scholar

[ref72] 72. Marsh HW, Guo J, Parker PD, Nagengast B, Asparouhov T, Muthén B, et al. What to do when scalar invariance fails: The extended alignment method for multi-group factor analysis comparison of latent means across many groups. Psychological Methods. 2018;23(3):524. pmid:28080078
View Article
PubMed/NCBI
Google Scholar

[227] View Article

[228] PubMed/NCBI

[229] Google Scholar

Figures

Abstract

Introduction

Materials and methods

Participants

Measures

Data analysis

Results

Factor analysis and network estimation

Cross-Sectional measurement invariance

Longitudinal measurement invariance

Reliability analyses

Discussion

Supporting information

S1 Table. Demographic characteristics (elementary: N = 2,327; middle: N = 2,325).

S2 Table. Standardized factor loadings for two-factor model of the Grit-S in elementary school at baseline.

S3 Table. Standardized factor loadings for the two-factor model of the Grit-S in middle school at baseline.

S4 Table. Standardized factor loadings in the configural invariance model of the Grit-S.

S5 Table. Standardized factor loadings in the configural invariance model of the Grit-S.

References