The Italian version of the Passion for Work Scale: First psychometric evaluations

To assess the factorial validity of the scale, the two steps of analyses of this study considered two different samples of workers. Both groups of participants are random samples.

The first sample used to perform the exploratory factor analysis is a heterogeneous sample and is composed by 101 participants: 54.5% female, average age 42.29 years ( SD = 12.86), 79.2% work full-time, average seniority 15 years ( SD = 11.70). Moreover, 76% are employees, 13% workers (working class), and 11% managers.

Two hundred and thirty-four participants belonging to a professional association of project managers compose the second sample, used to perform the confirmatory factor analysis. In particular, 77.3% are male, average age 44.82 years ( SD = 6.90), 90.5% work full-time, average seniority 11 years ( SD = 6.55). Moreover, 65.3% of them are supervisors, 20.5% employees, and 14.2% directors.

Participants completed a self-report questionnaire composed by the following scales:

Passion . As mentioned, to measure passion, the Passion Scale by Vallerand et al. (2003) and adapted to work activity was used. The scale was assessed by 14 items (7 items for HP and 7 items for OP) on a Likert scale ranging from 1 to 7, 1 means not agree at all and 7 very strongly agree (HP: M = 4.25, SD = 1.39; OP: M = 2.03, SD = 1.01).

Job satisfaction . Job satisfaction was assessed with the scale by Pejtersen, Kristensen, Borg and Bjoner (2010) with a 4-item Likert scale ranging from 1 to 5, 1 means strongly disagree and 5 strongly agree . The Cronbach's alpha for the present study is .82 ( M = 3.43, SD = 0.72).

Life satisfaction . Life satisfaction was assessed with the scale by Pavot and Diener (1993) with a 4-item Likert scale ranging from 1 to 7, 1 means strongly disagree and 7 strongly agree . The Cronbach's alpha for the present study is .91 ( M = 4.42, SD = 1.29).

Positive emotions at work . Positive emotions at work were assessed with the scale by Warr (1990) and operationalized as psychological well-being. They were measured with a 6-item Likert scale ranging from 1 to 6, 1 means never and 6 always . The Cronbach's alpha for the present study is .75 ( M = 3.70, SD = 1.01).

Negative emotions at work . Negative emotions at work were assessed with the scale by Warr (1990) and operationalized as psychological discomfort. They were measured with a 6-item Likert scale ranging from 1 to 6, 1 means never and 6 always . The Cronbach's alpha for the present study is .84 ( M = 2.82, SD = 0.96).

Flow at work . Flow at work were assessed with the scale by Bakker (2008), using the Italian adaptation by Zito, Bakker, Colombo, and Cortese (2015) , with a 13-item Likert scale ranging from 1 to 7, 1 means never and 6 always . The Cronbach's alpha for the present study is .90 ( M = 3.91, SD = 1.05).

Before the administration of the Italian version of the passion for work scale, the original version was translated from English to Italian, and back translated from Italian to English in order to control and verify items conformity. Moreover, the scale has been adapted from “passion for a general activity” to “passion for work”. This process was made with the supervision of an expert with English as mother tongue, with a satisfactory correspondence between items. Data were collected through an online questionnaire, placed on a platform implemented by researchers, but with different procedures: the first sample was randomly contacted, as the second sample received an email from their professional association of project manager with a link to the questionnaire, also published on the intranet of the association website. The board of directors of the association authorized the study and, since there was no medical treatment or other procedures that could cause psychological or social discomfort to participants, additional ethical approval was not d.

Before starting the questionnaire, participants had information about the voluntary nature of the participation in the study, the anonymity of their data, and instructions to complete the questionnaire.

To assess the factorial structure of the Italian version of the Passion for Work Scale, data analyses followed two main stages: the first concerned the heterogeneous sample ( N = 101) for the exploratory factor analysis with SPSS 22. The second stage involved the sample of professional workers ( N = 234) for the confirmatory factor analyses, conducted with Mplus 7. The model goodness of fit was tested considering the following indices: chi-square value (? 2 ), Comparative Fit Index (CFI, Bentler, 1990 ), Tucker-Lewis Index (TLI; Tucker & Lewis, 1973 ), Root Mean Square Error of Approximation (RMSEA; Steiger, 1990 ), and Standardized Root Mean Square Residual (SRMR; Jöreskog & Sörbom, 1993 ).

Moreover, in order to assess the validity of the scale and deepen the psychometric characteristics of the Italian version of Passion for Work scale, the correlations (Pearson's r ) between the two dimensions of the Italian version of Passion for Work Scale and other construct that literature highlighted to be correlated with the dualistic model of passion were examined (in both the samples used for the exploratory and the confirmatory factor analyses; N = 335). In fact, in the light of the characteristics of the type of passion and of how the work activity is internalized ( Marsh et al., 2013; Vallerand et al., 2003; Vallerand & Houlfort, 2003 ), it is expected that HP is more correlated with job and life satisfaction, flow at work experiences, and positive emotions at work, and less with negative emotions at work; on the contrary OP is expected to correlate with HP. Also, an analysis of variance (independent samples t-test ) was conducted in order to deepen the characteristics of the passion construct, by comparing passion scores for women and men.

Finally, in order to assess the internal consistencies of measures, Cronbach's alphas were calculated both for the two factors of the Italian version of the Passion for Work scale, and for each measure considered in the study.

Correlations, analyses of variance and Cronbach's alphas were performed with SPSS 22.

The exploratory factor analysis has been conducted with the maximum likelihood (ML) ion from the 14 items, through different solutions in order to explore data: eigenvalues > 1 with two factors ion, and then the ment of one, two, and three factors with no rotation, with Varimax, with Oblimin, and with Promax rotation for each number of factors ion. In line with theoretical bases and previous empirical findings, the best result was the two-factor solution with Oblimin rotation (Kaiser's normalization) and, as expected and in line with the original scale by Vallerand et al. (2003) , factors can be named Harmonious Passion and Obsessive Passion. Within this solution, eliminating item 7 from the dimension of HP was chosen (original item “I am completely taken with this activity”, Italian translation “sono completamente preso da questo lavoro”), since it overlapped with the two factors. Actually, in the Italian language, this item could suggest also the sense to be very occupied in the activity, in line with the concept of obsessive thought. The exploratory factor analysis with 13 items was hence conducted to verify the new factor solution, with good results. The chosen solution resulted therefore composed by 13 items (see Table 1): HP (six items, ? = .92) and OP (seven items, ? = .87). The factor loading for the 13-item solution ranges between |.68| and |.95| for HP, and between |.56| and |.79| for OP. Moreover, the Bartlett's Test of Sphericity is significant ( p < .000), thus making the factor analysis possible, and the KMO is very satisfactory (.87).

The factor solution absorbs 61.5% of the total variance (HP explains 44.1% of the variance and OP explains 17.4%) and this is in line with the original 14-item scale by Vallerand et al. (2003) that absorbed 54.7% of the total variance.

The correlation between the two factors HP and OP is r = .36.

The confirmatory factor analysis was performed on the second sample ( N = 234). As shown in Table 2 , different models were tested with ML method: with one factor, with two factors with the original composition of factors, and with 13 items without item 7, on the basis of the exploratory factor analysis. Model 1 and 2 gave unsatisfactory fit indices and Model 3 showed fit indices not completely acceptable. Evaluating the modifications indices calculated by Mplus (default option Modification M.I. – 3.840), the fit improved by adding two correlations in the model estimation (Model 4). More specifically, correlations between items 5 and 6 and between items 8 and 9 were d. At a semantic level, the suggested correlations made sense: items 5 (“This activity is in harmony with the other activities in my life”) and 6 (“For me it is a passion, that I still manage to control”) in fact they seem to be particularly linked to the aspects of the harmonious balance of the working activity with the individual's life and of the non-pervasiveness of work in the lifetime. Moreover, at a semantic level, items 8 (“I cannot live without it”) and 9 (“The urge is so strong, I can’t help doing this activity”) seem to be linked to the exclusivity of the work and its pervasiveness, as if the work were related to the identity of the person characterizing his/her life. Therefore, Model 4 was calculated obtaining good fit indices. Moreover, this inclusion did not modify other estimated parameters, and the dualistic model of passion emerged from the analysis. It has to be noted that the RMSEA value can be improved, but other modifications were not added in the estimation of the model, in order to have a cleaner model. These are preliminary analyses, and according to MacCallum, Browne and Sugawara (1996) , RMSEA values ranging between .08 and .10 are still acceptable.

Deepening Model 4, all items load only on the intended factors and factors loading range between |.34| and |.93| for HP, and between |.46| and |.85| for OP (see Figure 1).

Figure 1

Results of the Confirmatory Factor Analysis ( N = 234).

The correlation between factors is positive, in line with the original scale of passion by Vallerand et al. (2003).

Results of correlations between the two dimensions of the Italian version of the Passion for Work Scale (that the literature indicates as correlated with the two elements of passion), job satisfaction, life satisfaction, positive and negative emotions at work, and flow at work are shown in Table 3 . HP shows high and positive correlations with flow at work ( r = .71), job satisfaction (r = 54), positive emotions at work ( r = .53), life satisfaction ( r = .48), and OP (r = .30), as it shows a negative correlation with negative emotions at work ( r = -.27). As for OP, it shows positive correlations with flow at work ( r = .38), negative emotions at work ( r = .20), job satisfaction (r = .16), and – though weak – with positive emotions at work ( r = .13).

The results of the t -test for independent samples was not significant: in this sample, there are not differences in the perception of passion, both HP, t(330) = -.901, p = .37, and OP, t(330) = .009, p = .99, for women (Mean HP = 24.71, SD = 8.21; Mean OP = 14.24, SD = 7.90) and men (Mean HP = 25.59, SD = 8.38; Mean OP = 14.24, SD = 6.72).