{PSOCIAL}

Purpose holds an important place in the research on wellness, including the areas of prevention/health promotion, resilience, and psychological wellbeing. On one hand, purpose in life is an important predictor of mental health and psychological well-being (Bronk, 2014). In Carol Ryff’s foundational studies of wellbeing (1989) and her current work (2018), purpose is a component of psychological wellbeing. Likewise, purpose and meaning are central to theoretical models and empirical studies of psychological well-being (Li et al., 2021) and resilience (for reviews, see Bonanno, 2021; Grych et al., 2015). Similarly, purpose is consistently positively correlated with life satisfaction (Heng et al., 2020; Leria-Dulčić & Salgado-Roa, 2019; Schulenberg & Melton, 2010).

Furthermore, an absence of purpose in life has been linked to hopelessness and existential frustration or emptiness (Huamaní & Ccori, 2016; Schimmoeller & Rothhaar, 2021). There is a general consensus that a lack of purpose is associated with mental health problems including depression and suicidality (Heisel & Flett, 2004; Laird et al., 2019; Nkyi & Ninnoni, 2020; Straus et al., 2019; Sutin et al., 2018), and even a risk factor for mortality in older adults (Boyle et al., 2009; Hill & Turiano, 2014; Krause, 2009; Pinquart, 2002). In sum, across numerous studies, purpose in life is a protective factor for physical and psychological wellness, and a lack of purpose is a risk factor for worse mental health.

Various self-report scales have been developed to measure purpose, and one of the most widely used is the Purpose in Life Test (PIL), developed by Crumbaugh and Maholick (1964). Different versions and adaptations of this instrument have been developed and used in a variety of contexts including Latin America (Hayashi & Tieko Esmerelles, 2018; Martínez Ochoa et al., 2018; Martínez Ortiz et al., 2012; Nascimento & Dias, 2019; Simkin et al., 2018), North America (Bonebright et al., 2000; Durant et al., 1995; Jackson & Coursey, 1988; Schulenberg & Melton, 2010), Europe (Brunelli et al., 2012; García-Alandete et al., 2013; Haugan & Moksnes, 2013; Jonsén et al., 2010; Konkolÿ & Martos, 2006), Asia (Chang & Dodder, 1983; Kim et al., 2001; Law, 2012; Okado, 1998; Shek, 1993), Africa (Stones & Philbrick, 1980), and Oceania (Dyck, 1987; Marsh et al., 2003).

Various researchers have proposed modifications to the number of items in the instrument. Originally, Marsh et al. (2003) suggested the elimination of three of the scale’s twenty original items, on the basis that these items did not seem to contribute significantly to the scale. Law (2012) proposed retaining only seven items to measure the existential aspect of purpose, specifically. In 2010, Schulenberg and Melton concluded that many of the items on the original 20-item PIL did not measure purpose, but rather other constructs such as depression and boredom. Rigorous factor-analytic methodology identified three items that loaded extremely well together, and the fourth item was included on the basis of face validity and improved psychometrics (Schulenberg & Melton, 2010; Schulenberg et al., 2011). On this basis, the PIL-SF was developed, retaining items 3, 4, 8, and 20 from the original scale (Appendix; Schulenberg & Melton, 2010; Schulenberg et al., 2011).

The PIL-SF (Schulenberg et al., 2011) has consistently, inversely predicted forms of psychopathology, such as depression, substance use, and posttraumatic stress. Likewise, the PIL-SF positively predicts resilience and life satisfaction. The PIL-SF differentiates the specific aspect of the construct that is progress towards meaningful goals, whereas other English and Spanish instruments often lack this specificity. For example, the PIL-SF measures a different construct than the Meaning in Life Questionnaire (MLQ; Steger et al., 2006). In both the five-item English version (Steger et al., 2006) and a three-item Spanish version that was developed in Chile (Steger & Samman, 2012), the MLQ – Presence of Meaning subscale assesses current meaning and purpose in life. The distinction is that the PIL-SF specifically assesses the goal-directed aspect of purpose, whereas the MLQ – Presence subscale does not. Another measure of meaning and purpose that has been translated to Spanish is the Schedule for Meaning in Life Evaluation (SMiLE; Monforte-Royo et al., 2011). This questionnaire identifies a person’s primary sources of meaning, rather than the extent to which they perceive they currently have a purpose in life. Finally, the 20-item PIL has already been shown to be distinct from the Spanish language Seeking of Noetic Goals test (SONG), in that the PIL measures current presence of meaning and purpose while the SONG measures search for meaning (García-Alandete et al., 2018). Therefore, the PIL-SF is unique among Spanish language scales: it specifically measures goal-oriented purpose in life.

Even though various studies have supported the factorial structure of the Purpose in Life Test – Short Form proposed by Schulenberg et al. (2011) in such diverse contexts as the United States (Schulenberg et al., 2011) and China (Law, 2012; Wang et al., 2016), the psychometric properties of the Spanish-language version have not yet been evaluated. For this reason, the present study aimed to evaluate the reliability, factor structure, convergent validity, and discriminant validity of the Spanish PIL-SF. We hypothesized the following. (1) The PIL-SF would have adequate internal consistency. (2) The PIL-SF would have a one-factor structure. (3) There would be small to modest significant correlations between the PIL-SF and measures of wellbeing and spirituality. (4) There would be significant inverse correlations between the PIL-SF and measures of negative affect including depression and anxiety.

Sample A was comprised of 368 university students from Buenos Aires. Participants were adults ages 19 to 35 (M = 23.88; SD = 3.57), both men (23.3%) and women (77.7%). Sample B was comprised also by university students (N=336). Participants in Sample B were men (25.6%), women (74.1%), and other/non-binary (0.3%), and ranged in age from 19 to 55 years (M =25.09, SD = 5.97).

All data were collected through anonymous self-report survey questionnaires. For Samples A and B, participants were recruited via a link to the online survey that was distributed on social networks. Participation was voluntary, without compensation, and informed consent was obtained digitally before the surveys began. In the instructions, they were informed that their survey responses were anonymous and that the data obtained from the study would be used exclusively for scientific ends under Argentine National Law 25,326 protecting participants’ personal information and respecting their anonymity. No measure translation was conducted for this study, as all measures had previously been translated. The questionnaire for Sample A was composed of the Purpose in Life Test – Short Form (PIL-SF) and a sociodemographic questionnaire. Sample B was part of a larger study containing several questionnaires including the full form of the Purpose in Life Test (PIL), the ASPIRES short form, the NEO PI-R, the EBA, a sociodemographic questionnaire, and other questionnaires.

First, the internal consistency of the PIL-SF was analyzed utilizing the omega coefficient (interpretation based on Dunn et al., 2014) and Cronbach’s alpha coefficient (interpretation based on DeVellis, 2003). Second, confirmatory factor analysis (CFA) with maximum likelihood (ML) estimation was conducted to assess construct validity, in accord with recommendations from Jackson et al. (2009). The ML method was used for pairwise deletion of missing values. The model was later evaluated with the goodness-of-fit indices χ., the Comparative Fit Index (CFI), the Tucker-Lewis Index/Non-Normed Fit Index (TLI/NNFI), the Root Mean Square Error of Approximation (RMSEA) and the Standardized Root Mean Square Residual (SRMR), following recommendations from Hooper et al. (2008). The CFA was conducted with the lavaan package in R 3.3.4/RStudio, and all other statistical analyses of the data were performed using SPSS 27.

As shown in Table 1, McDonald’s omega for the PIL-SF was adequate for Samples A and B (based on cutoffs from Dunn et al., 2014; Green & Yang, 2015). Internal consistency calculated with Cronbach’s alpha was also adequate for the PIL-SF with both samples (based on cutoffs from DeVellis, 2003). Across administrations for Samples A and B, the PIL-SF means, standard deviations, and alpha coefficients were similar.

Descriptive Statistics for Versions of the Purpose In Life Test and its Short Form

As shown in Table 2, the model demonstrated excellent fit for a 1-factor model by the indices calculated. When administered both independently (Sample A) and embedded in the long form (Sample B), the PIL-SF fit was above the cutoff for the CFI and the Tucker-Lewis Index TLI, within recommended range for the RMSEA and RMSEA confidence interval (both were close to zero), and below the cutoff for SRMR (cutoffs from Hooper et al., 2008). For the long form of the PIL, fit statistics were inadequate, or outside of conventional cutoffs, for three of the indices that have conventional cutoffs (χ2, CFI, TLI). The RMSEA and SRMR had acceptable fit for the long form, but poorer fit than the PIL-SF did on these indices. As shown in the structural model for Sample A (Figure 1) and Sample B (Figure 2), all four PIL-SF items loaded well onto the overall factor.

Confirmatory Factor Analysis of Purpose in Life Test and its Short Form

Structural model of the PIL-SF using Sample A data. Path coefficients are standardized. Standard errors are shown in parentheses. N = 368.

Structural model of the PIL-SF using Sample B data. Path coefficients are standardized. Standard errors are shown in parentheses. N = 336.

For Sample A, Pearson correlations of each item with the other items in the independently administered PIL-SF were acceptable, 0.38 < r < 0.62, all p’s < .001 (Kline, 2005). For Sample B, Pearson correlations of each item with the other items in the PIL-SF as embedded in the long form were also acceptable, 0.42 < r < 0.51, all p’s < .001 (Kline, 2005).

As shown in Table 3, with Sample B, bivariate Pearson correlations were computed between the PIL, PIL-SF, and the measures of wellbeing (Satisfaction with Life Scale (SWLS), EBA Positive), negative emotions (EBA Negative, NEO PI-R Anxiety, NEO PI-R Depression), and spirituality (ASPIRES Connectedness and Universality subscales). The PIL-SF and PIL were strongly correlated with each other. Both the PIL and PIL-SF had similar correlation coefficients indicating moderate, significant, positive correlations with satisfaction with life and positive emotionality. Both the PIL and PIL-SF had similar correlation coefficients indicating significant correlations with negative emotionality (weak) and depression (moderate). The PIL was significantly correlated with anxiety but the PIL-SF was not. Both the PIL and PIL-SF were weakly, significantly, positively correlated with spiritual connectedness. Only the PIL-SF was significantly and positively, albeit weakly, correlated with universality.

Bivariate Pearson correlations of the Purpose in Life Test Short Form with other scales

The present study found good internal consistency for both the short and long forms, as calculated by both coefficients alpha and omega. Both coefficients were included due to alpha’s enduring popularity, despite criticism in the statistical literature (McNeish, 2018), and because omega is not yet common enough to be useful for comparing to other, similar scales. The reliability of this Spanish version of the PIL-SF is similar to reliability statistics reported by other studies of the PIL-SF in other countries (Law, 2012; Schulenberg et al., 2011).

In the present study, the CFA for the PIL-SF supported a unidimensional model. This corroborates findings from the original English language version of the PIL-SF, which also has a one-factor structure (Schulenberg et al., 2011). The original developmental work on the English-language version of the PIL-SF (Schulenberg & Melton, 2010; Schulenberg et al., 2011) concluded that, because many of the original PIL items do not load onto the overall Purpose factor, the PIL-SF was a more parsimonious, precise way to measure purpose in life than the 20-item scale. Our findings of superior fit indices (CFI, TLI) for the PIL-SF compared to the PIL corroborate the English PIL/PIL-SF studies; the PIL-SF does indeed appear to be a more parsimonious measure of purpose in life.

Convergent validity was demonstrated for the PIL-SF through its positive associations with scales of wellbeing, spirituality, and the PIL long form. Positive and moderate correlations were found between purpose and both satisfaction with life and positive affect. Purpose is often considered a component of psychological wellbeing, alongside life satisfaction and positive affect (Ryff, 1989, 2014; Seligman, 2011). Therefore, these correlations support the construct validity of the PIL-SF as a measure of purpose in life. The PIL and PIL-SF were both associated with spiritual connectedness and the PIL-SF was associated with spiritual universality. These elements of spirituality have been closely linked to meaning-making in the theoretical literature (Park et al., 2017). Although we hypothesized a larger correlation between the PIL-SF and the spirituality scales, the weak correlations suggest that the PIL-SF is a sufficiently precise measure of purpose and does not muddle this construct with spirituality. The PIL and PIL-SF are strongly positively correlated. This is unsurprising for a scale nested within another scale, and would be insufficient on its own to establish the convergent validity of the PIL-SF. However, as an addition to the aforementioned correlations, this corroborates the convergent validity of the PIL-SF as a measure of purpose in life.

In addition to convergent validity with purpose and wellbeing measures, discriminant validity was shown through inverse correlations between the PIL-SF and measures of negative affect. The PIL-SF was moderately, inversely, and significantly correlated with depression; weakly, inversely, and significantly correlated with negative emotionality; and the correlation with anxiety was not significant. Depression, more so than anxiety, has been used to establish construct validity in psychometric work on the English form of the PIL and PIL-SF, in part because hopelessness and despair are more theoretically opposite of purpose than other negative emotions (Frankl, 1956/2006). Therefore, while we hypothesized larger and significant correlations for the PIL-SF and anxiety or general negative affect, our findings nevertheless support the construct validity of the Spanish PIL-SF as a measure of purpose in life.

Limitations include that both samples were predominantly female and younger adults. These convenience samples are not representative of the general Argentine population. Further research should be conducted to validate the PIL-SF in other countries with other Spanish-speaking populations around the world. While further research is essential, this initial validation of the Spanish-language PIL-SF for Argentine use is an important step toward developing a psychometrically sound version with international applicability. Researchers working with Spanish speakers of other nationalities or cultures, or with Argentinians from other demographic groups, such as youth, should examine the validity of the Spanish PIL-SF with their population of interest.

Because the short form demonstrates stronger model fit comparable to the long form, the PIL-SF shows promise as a brief self-reported purpose in life with Spanish-speaking populations. We acknowledge that both the Spanish-speaking literature and broader global literature remains murky with regard to differentiating, classifying, and ordering meaning and purpose. While researchers have yet to reach a consensus as to whether purpose is a component of meaning (Martela & Steger, 2016; Park & Blake, 2020; Park et al., 2017) or whether meaning is an aspect of purpose (e.g., Bronk, 2012; Bronk et al., 2009; Damon et al., 2003; Schaefer et al., 2013), there is consensus in defining purpose as goal-directedness towards meaningful living. A measure of purpose, specifically, is necessary for such future research. Therefore, the Spanish PIL-SF has potential utility in empirical exploration of the differences, similarities, and hierarchy of meaning and purpose. Furthermore, at least three papers in the literature have asserted that some of the PIL items excluded from the PIL-SF seem to measure constructs other than purpose (Law, 2012; Marsh et al., 2003; Schulenberg & Melton, 2010). Therefore, especially when the intent is to measure purpose as a variable distinct from related constructs like meaning or spirituality, researchers should consider the short form to be the most precise option. We recommend that researchers utilize the PIL-SF when assessing the extent to which individuals currently perceive themselves as having purpose in life, or goals and aims towards living meaningfully.