Psychometric Evaluation of the Grit Psychological Resources Scale (GPRS)

Introduction

Grit is defined as the consistent pursuit of long-term goals, despite setbacks, adversity and challenges.¹ Favourable outcomes are thought to be associated with one’s consistency of interest (passion) and perseverance of effort (perseverance) toward higher-order goals.² Grittier individuals are adept at realigning lower-order activities and plans to facilitate the achievement of desired outcomes, often many years later. People with high levels of grit are reported to experience less stress and burnout, greater job satisfaction and retention,^3,4 and improved capacity for positive leadership.⁵ As individuals, businesses, and the larger global community navigate the post-COVID-19 era, goal-oriented capabilities are increasingly important for recovery from psychosocial and economic impact.^6,7 Recent studies have raised questions regarding the predictive value of grit, especially concerning the passion facet.^8,9 These findings suggest that more work is needed to better understand the construct, its predictors and effects. In contrast, other studies support the construct,^10,11 indicating a need to consider the psychological resources of grit and influence on passion and perseverance.

Psychological resources include personal characteristics, such as attributes (self-determination and optimism) and skills (communication and problem solving), that are valued and conserved or expanded to help buffer against stress.¹² In addition to providing a protective role, psychological resources can enhance health and wellbeing.¹³ In the context of grit, psychological resources may enhance people’s ability to better cope with stressful situations and find multiple pathways to persevere with their passions, rather than giving up or changing directions as seen in less gritty people.¹⁴ People with more resources are also likely to be well received in the workplace, as businesses work to strengthen their resolve against future health and economic challenges.¹⁵ As such, research on identifying and measuring psychological resources that can contribute to higher levels of grit has significant practical implications. For example, resources may help people persevere at work and with career goals in industries that place complex demands on professionals, such as healthcare and nursing.¹⁶

Informed by Duckworth’s¹⁷ previous work on grit, which identified four psychological resources of grit, interest, purpose, practice and hope, the authors have proposed a model outlining sixteen attributes associated with these resources.¹⁸ Although there are many scales that measure the attributes of interest, in general these instruments assess the concepts broadly and do not specifically assess behaviour in the context of grit. Thus, a new scale called the Grit Psychological Resources Scale, or GPRS, was developed by the authors. The intent of the GPRS is to assess the extent to which an individual has accumulated psychological resources of grit.

Construction of the GPRS scale is detailed in a previous paper.¹⁹ Two extensive literature reviews on grit and related concepts provided the theoretical and conceptual framework. Namely, a concept analysis of grit was performed to identify attributes, antecedents and consequences of grit.²⁰ The results of this concept analysis informed the development of an integrated model for the psychological resources of grit.¹⁸ Definitions and behavioural descriptions for the model elements were then generated and used to create a pool of potential scale items. Insights were also elicited from existing scales on wording, instructions and response options. The Delphi technique was then used with a panel of 30 experts to evaluate 100 items for content and face validity. Over a series of three online surveys, the expert panel rated and agreed on 20 items for the final GPRS.¹⁹ The aim of this study was to evaluate the psychometric properties of the GPRS.

Method

A quantitative study was conducted with adults aged 18 years and over. Participants completed online surveys at two time points, three months apart, between December 2020 and May 2021. The data collected was used to perform psychometric testing of the GPRS, including internal consistency, test–retest reliability, regression analyses, convergent validity and confirmatory factor analysis. The study complies with the Declaration of Helsinki and was approved by the University of Tasmania Social Sciences Human Research Ethics Committee, reference number S0023721 (H-72893).

Participants

Participants were invited to take part in the study through the first author’s professional networks, social media platforms, online communities, emails and snowball sampling techniques (people sharing pages and posts). The study aimed to recruit the recommended sample size of 10 respondents for each scale item, which equated to at least 200 participants.²¹ Study invitations included an internet link that directed people to the first online survey stored in a secure web-based application (REDCap 12.4.8). All participants were required to be over 18 years old and received an information sheet about the study and the option to have any questions answered before providing informed consent to participate.

Data Collection

The surveys set up in REDCap required participants to answer each question before proceeding to the next question, thus reducing the chance of missing data. In the first survey, participants were asked to answer demographic questions for group-level analysis, the GPRS, and four additional scales selected to test convergent validity. Three months later, participants were asked to complete the GPRS to assess test–retest reliability.

Measures

The questionnaire battery consisted of the GPRS and four other scales associated with the nomological network of the psychological resources of grit.

Grit Psychological Resources Scale (GPRS)

The GPRS consists of 20 positively worded items with four sub-scales: interest, purpose, practice, and hope. The following instructions for completion of the scale were provided:

For the statements below, indicate your level of agreement on the scale provided. When responding, consider your behaviour in general. Think about current behaviour – not how you think you should or want to be. If your behaviour is considerably different at work, select the response that best reflects you outside work. There are no right or wrong answers.

The response options were “1 = very strongly disagree”, “2 = strongly disagree”, “3 = disagree”, “4 = neutral”, “5 = agree”, “6 = strongly agree”, and “7 = very strongly agree”. Psychometric properties of the GPRS are reported in the present paper.

Passion Scale

The Passion Scale is a 17-item scale that assesses harmonious passion (HP) and obsessive passion (OP).²² There are six items for each subscale of passion and five items to identify the presence of passion. For example, “This activity is in harmony with the other activities in my life” (HP), “I have difficulties controlling my urge to do my activity” (OP), and “I spend a lot of time doing this activity” (presence of passion). Responses are made on a 7-point Likert scale ranging from 1 (do not agree at all) to 7 (very strongly agree). Vallerand et al's²² original paper reported acceptable levels of reliability in four studies, with Cronbach’s alpha values ranging from 0.71 to 0.84 and 0.85 to 0.92 for the HP and OP sub-scales, respectively. Good psychometric properties were also reported for this scale across different languages and in the context of various activities in data from 19 studies.²³

Multidimensional Persistence Scale (MPS)

The MPS is a 13-item scale that assesses three dimensions of persistence: persistence despite difficulty (PDD; eg, “I keep on going when the going gets tough”), persistence despite fear (PDF; eg, “I tend to face my fears”), and inappropriate persistence (IP; eg, “Sometimes I find myself continuing to do something, even when there is no point in carrying on”).²⁴ Responses are made on a 7-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree). The MPS demonstrated acceptable reliability, with Cronbach’s alpha values for the PDD, PDF and IP sub-scales of 0.85, 0.93, and 0.76, respectively.²⁴

Life Engagement Test (LET)

The LET is a 6-item scale that assesses purpose in life; eg, “To me, the things I do are all worthwhile”.²⁵ Responses are made on a 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). Psychometric evaluation found the LET to be stable across gender, age, ethnicity, and time. Cronbach’s alpha values were satisfactory, ranging from 0.72 to 0.87 across eight samples.²⁵

Hope Scale

The Hope Scale, labelled The Future Scale for administration purposes, is a 12-item scale that assesses hope in terms of a person’s sense of successful agency (eg, “I energetically pursue my goals”) and pathways (eg, “I can think of many ways to get out of a jam”) to achieve goals.²⁶ Responses are made on an 8-point Likert scale ranging from 1 (definitely false) to 8 (definitely true). Testing of The Hope Scale across six samples yielded acceptable results for reliability, with Cronbach’s alpha values ranging from 0.74 to 0.84 for total scores and 0.71 to 0.76 and 0.63 to 0.80 for the agency and pathway sub-scales, respectively.²⁶

Statistical Analyses

SPSS for Mac 28.0 was used to perform all statistical analyses, except for CFA, which was conducted using Stata/BE 17.0 for Mac. Before conducting analyses, data were screened for errors and the distribution of scores, including normality. Additionally, total scores were calculated, and the age and country demographic variables collapsed into groups. The psychometric properties of the GPRS were examined using several tests for reliability and validity. Cronbach’s alpha was used for internal consistency.²⁷ An alpha coefficient of 0.70 is considered acceptable; however, above 0.80 is preferred for the psychometric evaluation of scales.²⁸ Test–retest reliability was assessed using the Pearson’s r correlation, with higher values indicating higher reliability.²⁹ Regression tests assessed if demographic variables influenced GPRS score and survey completion.

Content and face validity were established during scale construction. A theoretical model for the psychological resources of grit informed construction of the initial bank of items. Content validation was strengthened by a Delphi study whereby 30 experts rated items for inclusion in the GPRS according to the extent to which statements reflected the domains of interest; ie, psychological resources. The Delphi process also provided evidence of face validity. Readability was assessed using the Microsoft Word for Mac version 16.46. Flesch Reading Ease scores between 70 and 80 are acceptable for the adult general public and a Flesch-Kincaid Grade Level of 8 indicates the text can be understood by 13 to 14-year-olds.³⁰ The Pearson’s r correlation was used to assess the relationship between other scales that measure conceptually similar concepts, with higher correlations considered evidence for convergent validity.³¹

Confirmatory factor analysis (CFA) was also conducted to examine scale validity. For scale development, exploratory factor analysis (EFA) can be used to establish the latent structure before verifying the relationship with CFA. However, given that the factor structure for the GPRS was established in prior theoretical and empirical studies (conceptual model and Delphi study), only CFA was performed.³² The following fit indices and thresholds were used to evaluate models:

• Chi-square test (X²), non-significant p-values suggest good fit.³³
• Root mean square error of approximation (RMSEA), values ≤0.08 indicates reasonable fit³⁴ and 0.08–0.10 indicates mediocre fit.³⁵
• Akaike information criterion (AIC)³⁶ and Bayesian information criterion (BIC),³⁷ smaller values are considered better fit.
• Comparative fit index (CFI)³⁸ and Tucker-Lewis index (TLI),³⁹ values ≥0.90 considered acceptable fit.
• Standardized root mean squared residual (SRMR),⁴⁰ values close to 0.08 suggest good fit.
• Factor loadings of ≥0.32 can be interpreted, ≥0.45 fair, ≥0.55 good, ≥0.63 very good, ≥0.71 excellent.^41,42 A cut-off of 0.3 was used in this study to retain items.

Results

Sample Characteristics

A total of 277 participants completed the first survey, and of these, 173 completed the second survey three months later (retention rate = 62.5%). Table 1 shows the demographic characteristics of the participants completing the first survey. Over half of participants (61%) were aged between 40 and 59 years (mean age = 47.2, SD 11.7) and three-quarters (77%) were female. Participants who were employed (240) worked in various industries, with over one-third in education and training (35%), with others in professional, scientific and technical services (19%), and healthcare and social assistance (14%). Nearly two-thirds of participants held postgraduate qualifications (60%), followed by undergraduate (24%) and vocational (11%). Of those currently undertaking study (98), most were pursuing postgraduate qualifications (61%). Participants were located in 37 countries across seven geographic regions, with the highest proportion (61%) residing in Oceania. Table 2 shows the descriptive statistics of the scales used in this study.

Table 1 Participant Demographics

Table 2 Descriptive Statistics of the Study Scales (N = 277)

Reliability

Table 3 presents the means and standard deviations for each item of the GPRS. The Cronbach’s alpha was 0.906 (0.909 for survey 2), suggesting very good internal consistency.²⁸ Of the 277 participants who completed the first survey, 173 (66%) also completed a second survey three months later. The Pearson’s r correlation coefficient between Time 1 and Time 2 was 0.75, p = < 0.001, providing evidence of test–retest reliability.²⁹

Table 3 GPRS Items, Means and Standard Deviations

Table 4 shows results of linear regression analyses predicting GPRS score based on age, gender and qualification. Univariate linear regression showed that age did not significantly predict GPRS score on the first survey (Β = 0.004, p > 0.05). Likewise, there was no significant relationship with gender (Β = 0.052, p > 0.05) or qualification (Β = 0.100, p > 0.05). Multivariate linear regression yielded similar results to the univariate analyses, F (3, 273) = 1.22, p > 0.05.

Table 4 Univariate and Multivariate Linear Regression Analyses Predicting GPRS Score Based on Age, Gender and Qualification (N = 277)

Table 5 presents results of logistic regression analyses predicting likelihood of completing the second survey based on age, gender and qualification. Univariate logistic regression did not reveal any significant association with age (OR = 1.01, p > 0.05) or gender (OR = 0.78, p > 0.05) on completion of the second survey. The odds of completing the second survey for participants who had completed a Master’s degree or above was 1.57 (95% CI, 0.96 to 2.57) times that of those with a Bachelor’s degree or below, although the result was not significant. Multivariate logistic regression yielded similar results to univariate analyses, X² (3, N = 277) = 5.92, p > 0.05, including the predictive effect of qualification (OR = 1.59). Results from the regression analyses provide further support for scale reliability.

Table 5 Univariate and Multivariate Logistic Regression Analyses Predicting Likelihood of Completing the Second Survey Based on Age, Gender and Qualification (N = 277)

Validity

Content validity and face validity were established during the development of the GPRS, with experts agreeing on the final scale items. The Flesch Reading Ease score was 55.8 and the Flesch-Kincaid Grade Level was 8.1, which indicates that adults in the general public should understand the scale.³⁰ There were positive moderate correlations (Pearson’s r) between the GPRS and the passion (r = 0.34), persistence (r = 0.49), LET (r = 0.51) and hope scales (r = 0.67). All correlations were statistically significant demonstrating convergent validity (see Table 6).

Table 6 Pearson’s r Correlations Between Scales That Measure Similar Constructs (N = 277)

Confirmatory Factor Analysis

Nine model variations that the data could meet were selected and tested, including first, second and third-order models as follows:

• Model 1 was a first-order model with five items loading onto each of four factors (interest, purpose, practice, and hope). The factors were allowed to correlate with each other.
• Model 2 was a first-order model with ten items loading onto each of two factors (passion and perseverance). The factors were allowed to correlate with each other.
• Model 3 was a first-order model with all the items loaded onto one factor (grit psychological resources).
• Model 4 was a second-order model with four first-order factors (interest, purpose, practice, and hope; five items loaded onto each factor) and one second-order factor (grit psychological resources). The first-order factors were allowed to correlate with each other.
• Model 5 was a second-order model with two first-order factors (passion and perseverance; 10 items loaded onto each factor) and one second-order factor (grit psychological resources). The first-order factors were allowed to correlate with each other.
• Model 6 was a third-order model with four first-order factors (interest, purpose, practice, and hope; five items loaded onto each factor), two second-order factors (passion and perseverance) and one third-order factor (grit psychological resources). The first-order factors were allowed to correlate with each other. The two second-order factors were also allowed to correlate with each other.
• Model 7 was a second-order model with four first-order factors (interest, purpose, practice, and hope; five items loaded onto each factor) and one second-order factor (grit psychological resources).
• Model 8 was a second-order model with two first-order factors (passion and perseverance; 10 items loaded onto each factor) and one second-order factor (grit psychological resources).
• Model 9 was a third-order model with four first-order factors (interest, purpose, practice, and hope; five items loaded onto each factor), two second-order factors (passion and perseverance) and one third-order factor (grit psychological resources).

Table 7 shows the fit indices for models 1, 2, 3 and 7. The other five models did not converge, indicating they did not fit the data. All the first-order models (1, 2 and 3) and one second-order model (7) showed acceptable SRMR values. The RMSEA values for the four models compared for fit fell within the average range. However, SRMR values arguably yield more accurate rejection rates.⁴³ The AIC and BIC values were similar, with the lowest values preferred (model 1 and 7). Likewise, the CFI and TFI values did not vary much between models and were below the suggested cut-off. Overall, when considering both the theoretical and conceptual background for this research, model 7 is the preferred model.

Table 7 Confirmatory Factor Analysis Fit Indices for the GPRS (N = 277)

Figure 1 shows the standardised factor loadings and error variances of model 7. The factor loadings for items were all ≥0.49, with only one exception (0.3), 15 (75%) were >0.55 and 10 (50%) were ≥0.63. Loadings onto respective factors ranged from 0.49 to 0.63 for interest, 0.30 to 0.85 for purpose, 0.64 to 0.68 for practice, and 0.54 to 0.63 for hope. The loadings between the first-order factors and second-order factor were high: 0.90 for interest, 0.80 for purpose, 0.97 for practice, and 0.91 for hope. These findings provide further psychometric support for the second-order model, which yields a total score with four sub-scales.

Figure 1 Confirmatory factor analysis: the preferred model (Model 7).

Scoring and Interpretation

The average grit psychological resources (GPR) score is calculated by summing all the statement responses and dividing the total by 20. Possible scores range from 1 to 7 and indicate the GPR level as follows:

• Score of less than 2 indicates very low GPR = level 1, undeveloped.
• Score of 2 to less than 3 indicates low GPR = level 2, foundational.
• Score of 3 to less than 4 indicates low-moderate GPR = level 3, intermediate.
• Score of 4 to less than 5 indicates moderate-high GPR = level 4, developed.
• Score of 5 to less than 6 indicates high GPR = level 5, well-developed.
• Score of 6 to 7 indicates very high GPR = level 6, advanced.

The score levels and descriptors may be used to provide feedback on current GPR status and articulate meaningful changes over time.

The average subscale scores are calculated by summing the statement responses for each psychological resource and dividing the total by 5: statements 1 to 5 for interest, statements 6 to 10 for purpose, statements 11 to 15 for practice, and statements 16 to 20 for hope. Subscale scores can be interpreted similarly to the GPR score and used to provide more detailed information on developmental opportunities. In addition, given that one’s perceived GPR may be influenced by the surrounding environment, separate instructions have been developed for administering the scale in the workplace. At a workplace level, respondents can answer from an individual perspective (ie, behaviour when doing work that contributes to individual goals) or team perspective (ie, behaviour when doing work that contributes to team goals).

Discussion

This study aimed to evaluate the psychometric properties of the GPRS. Evidence of reliability was supported by two primary types of evidence: internal consistency and test–retest reliability.⁴⁴ The GPRS showed high internal consistency.²⁸ Evidence of test–retest reliability was also established, indicating that scores are stable across time.²⁹ Further evidence of scale reliability was sought by considering the potential influence of demographic variables via standard multiple regression and logistic regression. Age, gender and qualification did not significantly predict scores or influence the completion of the second survey. However, a more heterogeneous sample may have yielded different results and should be explored in future studies.⁴⁵ Taken together, the tests of reliability indicate that the GPRS is stable and consistent; ie, measuring the variables as intended.⁴⁴

Evidence of validity was supported by several types of evidence including scale content (face and content validity), item response processes (reading ease and understandability), association with other instruments (convergent validity), internal structure (CFA), and interpretation of scale scores.⁴⁴ Three previous studies established the theoretical and conceptual basis for creation of the GPRS, and face and content validity. The Flesch reading ease and Flesch-Kincaid grade level indicated that the scale would be suitable for general population adults.³⁰ Additionally, analysis revealed significant moderate correlations between the GPRS with other similar scales, indicating the scale accounts for new and unique variation beyond what is already established in the current literature.⁴⁶

Confirmatory factor analysis provided support for four models, with all items loading well onto their respective factors and correlations between factors suggesting a strong association. Considering the results of fit indices, together with the theoretical and conceptual background for this study, the second-order model (model 7) was superior to the other three first-order models. The preferred model has four first-order factors (interest, purpose, practice, and hope) that load onto a second-order factor (grit psychological resources). Overall, the results from the psychometric evaluation demonstrate that the GPRS is a reliable and valid instrument with which to assess the psychological resources of grit that an individual or team has accumulated.

Limitations and Future Research

The scale has only been tested on adults and the psychometric properties on children are unknown. A gold standard does not exist for measuring the psychological resources of grit, so a complete assessment of validity was not possible. The sample consisted largely of females aged between 40 and 60 with university qualifications. Thus, scores may vary with different populations. The study used self-report questionnaires to obtain information and respondents may not have answered truthfully or accurately.⁴⁷ Future studies should include more heterogeneous samples, including equal representation of gender and age and more diverse educational backgrounds. The scale’s predictive, concurrent and discriminant validity could be investigated further. Finally, the scale could be translated into other languages to enable psychometric evaluation in non-English speaking countries.

Practical Implications

The GPRS provides an instrument that leaders, coaches, educators, and practitioners can easily administer to assess the level of psychological resources accumulated by an individual or team. This information can inform where to focus development efforts and, in turn, help to maximise the use of available resources for building capabilities. Furthermore, we expect this scale to be useful in understanding how the psychological resources of grit contribute to enabling more passion and perseverance for long-term goals.¹ It can also be used to understand the role of psychological resources in behavioural change.⁴⁸ For example, in the workplace, can they predict improved performance, and how can they influence other important metrics, including retention, engagement and culture? The scale could also inform which positive psychology interventions, such as those for character strengths⁴⁹ and other development approaches like coaching,⁵⁰ are suitable In addition, the scale provides some flexibility for use in teams, where team participants may complement each other’s strengths and weaknesses among the subscale scores. Finally, it can be used to gather information on the effect development initiatives are having at an individual, team and organisational level. The GPRS is publicly available for use, provided the authors and original source are correctly attributed (see Supplementary File).

Conclusion

The results of this study provide evidence for the structural stability of the 20-item GPRS. Scale development is an ongoing process and further study is required to evaluate additional types of validity. Additional testing should also involve more diverse populations. Using the GPRS with individuals or teams to assess the level of psychological resources accumulated will enable more targeted development of capabilities for goal achievement. Moreover, insights from GPRS scores can assist with evaluating personal and professional development.