Disentangling workplace sexism in age generations: A multi-group analysis on the effects on job satisfaction and task performance

Morando, Martina Article Disentangling workplace sexism in age generations: A multi-group analysis on the effects on job satisfaction and task performance Administrative Sciences Provided in Cooperation with: MDPI – Multidisciplinary Digital Publishing Institute, Basel Suggested Citation: Morando, Martina (2023) : Disentangling workplace sexism in age generations: A multi-group analysis on the effects on job satisfaction and task performance, Administrative Sciences, ISSN 2076-3387, MDPI, Basel, Vol. 13, Iss. 9, pp. 1-15, https://doi.org/10.3390/admsci13090210 This Version is available at: https://hdl.handle.net/10419/320776 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Citation: Morando, Martina. 2023. Disentangling Workplace Sexism in Age Generations: A Multi-Group Analysis on the Effects on Job Satisfaction and Task Performance. Administrative Sciences 13: 210. https://doi.org/10.3390/ admsci13090210 Received: 11 August 2023 Revised: 11 September 2023 Accepted: 18 September 2023 Published: 21 September 2023 Copyright: © 2023 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). administrative sciences Article Disentangling Workplace Sexism in Age Generations: A Multi-Group Analysis on the Effects on Job Satisfaction and Task Performance Martina Morando Section Psychology, Department of Educational Sciences, University of Catania, 95124 Catania, Italy; [email protected] Abstract: The gender gap is a current topical issue. Sexist assumptions that manifest as gender stereotypes are partially responsible for these inequalities. The ambivalent sexism theory argues that hostile sexism refers to explicitly antagonistic sexist attitudes, while benevolent sexism refers to apparently positive but implicitly malevolent attitudes. There has been evidence reported that benevolent sexism is detrimental to women’s personal and professional well-being, implies lower levels of career aspiration and impacts task performance. This study is aimed at examining the impact that the experience of benevolent and hostile sexism could have on performance and job satisfaction. A total of 402 female workers were enrolled. The results showed that an experience with benevolent sexism significantly decreased the positive relationship between work engagement, psychological capital and organisational support and outcomes. Conversely, hostile sexism only reduces job satisfaction in its interaction with work engagement and organisational support. Moreover, through a multi-group analysis, possible differences across age were examined in the theorised model. Here, the younger generation seems to be more affected and experience more benevolent sexism than the older generation, which is seen both in individual moderators and in their interactions with predictors. This study is helpful for a deeper comprehension of contemporary sexism, offering also suggestions for equality policies’ design. Keywords: sexism; gender equality; social sustainability; satisfaction; performance 1. Introduction Measuring gender equality is a complex and difficult issue, reflecting its multidimensional nature. To assist in the analysis of gender equality in the various European states, the European Institute for Gender Equality developed a tool known as the Gender Equality Index. This tool measures progress in the current state of gender roles and focuses on specific areas in need of improvement and change, thus providing valuable suggestions and guidance for the policy design of more effective gender equality measures. The Gender Equality Index tracks 32 different indicators under six different macrodimensions: ‘Work’, ‘Money’, ‘Knowledge’, ‘Time’, ‘Power’ and ‘Health’. Looking at the results referring to the year 2022, gender inequality seems to persist despite the attempts and efforts of states and various institutions. Hence, gender equality still remains a goal to be pursued for most European countries. The overall score obtained by the European Union (EU), in fact, was 68.6 points out of 100. In addition, for the first time since this instrument was first measured and used, there was a decline in several areas due to the uneven impact of the COVID-19 pandemic. However, progress in some domains prevented an overall shift in gender equality. Trying to summarise the results from the Gender Equality Index, however, it seems that there is still much work left to be done to achieve full and satisfactory gender equality. According to this index, the most egalitarian countries are Sweden (83.9), Denmark (77.8) and the Netherlands (77.3), while the least egalitarian are Greece (53.4), Romania (53.7) and Hungary (54.2). Italy for example, the territory where this study took Adm. Sci. 2023,13, 210. https://doi.org/10.3390/admsci13090210 https://www.mdpi.com/journal/admsci Adm. Sci. 2023,13, 210 2 of 15 place, ranks 14th in the EU with a score of 65, just 3.6 points lower than the EU. Since 2010, Italy’s overall index has increased by 11.7 points, recording one of the largest long-term improvements among the Member States. A major step backwards, however, occurred in the work domain, where a not particularly high score (63.2 points) dropped further. As a result, Italy consistently ranks in last place among all the EU Member States in the work domain, with high levels of gender inequality, especially in the sub-domain of labour participation. Italy also seems to rank low with the gender employment gap (around 18%) and, again, with the percentage of employed women (51.1%). Reflecting on the current situation and discussing these inequalities and disparities does not seem utopian; on the contrary, the global pandemic seems to have exacerbated the difficulties already in place. Indeed, the COVID-19 pandemic was not only a health emergency but also contributed significantly to the deterioration of the economic crisis and the labour market in general. The impact on a global scale that people experienced is actually much broader and more complex than one might think. The economic and labour crisis caused by COVID-19 has significantly increased unemployment worldwide and worsened already precarious and difficult employment situations. This crisis also seems to have had a greater impact on certain groups of working men and women, thus increasing the inequalities. Among them, people in less protected and lower-paid jobs were those at greatest risk and who suffered the worst consequences, mainly including groups such as young and older workers, female workers and migrant workers. The pandemic crisis hit an economy and a society that already had low female employment rates, poor work–family reconciliation policies, very strict gender stereotypes and a very skewed, albeit very slowly evolving, division of labour in the family according to gender. The pandemic crisis has both accentuated these aspects and made visible the fragile balance on which they were held. In Italy, a country with a strong conservative tradition and rigid gender roles, the situation is certainly no different and opens up new and important reflections on inequalities and disparities, which deserve attention and in-depth study. Although the social position of women in recent decades appears significantly improved in many areas including the workplace, it is still difficult to speak of true gender equality since gender gaps continue to exist. Indeed, in Europe and in Italy, the position of women compared to men has not changed in relation to employment rates, success and career choice/possibility or equal pay. An important indicator of the efforts towards achieving gender equality in the workplace was the presence of approximately equal numbers of women and men in leadership positions. However, the situation in organisational contexts is often not so equal. Gender differences actually persist and overwhelmingly influence workplace outcomes (Barreto et al. 2009). The reasons underlying these differences are not women’s own attributes or attitudes; on the contrary, women are motivated and engaged in work activities in the same way and at the same level as men (Ellemers et al. 2004). A possible explanation can be found in gender discrimination. At the organisational level, sexist attitudes and beliefs are probably responsible for the approval of policies that hinder (rather than facilitate) the participation of women in the workforce (Benokraitis and Feagin 1986;Maier 1999). Several studies further prove that the endorsement of sexist beliefs (from male and female sources) is associated with the biased treatment of individual women, either in the personnel selection process, in the evaluation of their performance or in promotion decisions and opportunities (Bartol 1999;Ellemers et al. 2004;Heilman 2001). Despite these considerations, the literature reports that society currently tends to believe that gender discrimination is no longer a current problem (Swim et al. 1995;Tougas et al. 1995). Effectively, sexism still exists but has changed its nature by becoming more implicit, subtle and difficult to recognise (Zehnter et al. 2021;Morando et al. 2023). In spite of this, although expressed differently, sexist ideologies continue to be experienced and perceived by women, who are often affected and experience the strongest consequences. The ambivalent sexism theory introduced by Glick and Fiske (1996,1999) is an example of a contemporary form of sexism and argues that hostile sexism refers to explicitly Adm. Sci. 2023,13, 210 3 of 15 antagonistic sexist attitudes, while benevolent sexism refers to apparently positive but implicitly malevolent attitudes (Acar and Sümer 2018). Furthermore, while hostile sexism is a more blatant form of sexism, benevolent sexism is a more subtle form of it. Both forms generally have concrete implications and significantly affect members of minority groups, although benevolent/subtle sexism has a differently stronger impact. Indeed, the literature on the topic suggests that more subtle forms of prejudgment influence individuals in the way they feel and act, even leading them to unconsciously behave in ways that help justify or perpetuate their own inferiority (Steele and Aronson 1995;Ellemers and Barreto 2008). Perceiving oneself as a victim of sexism has a strong and complex impact on an individual’s well-being and affects one’s existence in a holistic way. Perceiving oneself as a victim of discrimination implies devaluing the social group to which one belongs (a fundamental aspect of social identity) and represents a devaluation of a part of oneself, also communicating social exclusion to others (Branscombe et al. 1999;Schmitt et al. 2002). A woman who feels she is a target of gender discrimination, for example, not only has to face the negative outcome this entails but also has to realise that something fundamental about her (her gender) is devalued by others. Thus, while blatant sexism leads to attention and emotional and negative reactions to the source of the mistreatment, more subtle sexism leads to self-directed negative emotions, such as anxiety and self-directed anger (Barreto and Ellemers 2005). These emotions also influence attitudes and behaviour and, consistent with the concept of stereotype threat, cause the targets of discrimination to have poorer performance precisely because they believe themselves to be the targets of a negative stereotype, in a kind of self-fulfilling prophecy (Steele 1997). Moreover, as benevolent sexism implicitly suggests women’s lack of competencewhile providing ambivalent messages and praise, it contributes to women doubting their ability to perform well and experiencing a decrease in self-esteem and self-confidence, as well as feeling anxious during tasks. This definitely impairs their performance (Dardenne et al. 2007). Different studies revealed the impact of benevolent sexism on women’s personal and professional well-being (Rubin et al. 2019;Dardenne et al. 2007;Moya et al. 2007), as better illustrated in the following Section 4. Starting from these premises, the aim of the study is to examine the impact that the experience of benevolent and hostile sexism could have on performance and job satisfaction in a sample of female workers. To do so, common predictors in the literature that significantly predict the outcomes were selected, introducing experience with benevolent and hostile sexism as moderators and investigating their role in modifying the predictive power of the predictors. Specifically, work engagement and psychological capital were selected as personal variables and organisational support as a variable measuring the perception of the organisation. Work engagement, for example, was selected because it is one of the most important predictors with significant impacts on the level of job performance (Grobelna 2019) and for increasing job satisfaction (Zalewska 2020;Hakanen and Schaufeli 2012;Corbeanu and Iliescu 2023). Additionally, PsyCap was selected (in its individual components as well as a higher-order construct) due to it being commonly associated with job satisfaction (Görgens-Ekermans and Steyn 2016;Platania and Paolillo 2022). PsyCap has been found to predict task performance, whether it was self-rated, supervisor-rated or objectively measured (Avey et al. 2011). It has also been linked to contextual performance, viz., organisational citizenship behaviours and other forms of desirable performance criteria, including creative performance, problem-solving and innovation (Newman et al. 2014). The literature maintains s that individuals’ POS helps boost their obligations toward an organisation in order to reciprocate favourably. There has been evidence reported that individuals’ POS enhances both in-role performance, such as goal attainment, and extra-role performance, such as helping and supportive behaviour toward coworkers (Eisenberger et al. 2001). A meta-analysis conducted by Rhoades and Eisenberger (2002) revealed that favourable treatments, such as rewards from the organisation, beneficial working conditions and fairness, received by employees are directly linked to POS. Moreover, POS promotes Adm. Sci. 2023,13, 210 4 of 15 auspicious outcomes such as high job satisfaction, lower turnover, enhanced dedication, positive emotions and better performance (Yu and Frenkel 2013;Koopmans et al. 2015). Additionally, to provide a broader examination and understanding, the study aimed at examining, through a multi-group analysis, the possible difference across age in the theorised model. Indeed, the literature provides us with some insight into how sexist attitudes might differ across the lifespan, but few studies have been carried out in relation to the possible different experiences with the two forms of sexism (hostile and benevolent). Surveys and polls seem to report that younger women who experience more sexism are more likely to experience negative effects on their health, well-being and work performance (Hammond et al. 2018;Buscemi et al. 2016). 2. Materials and Methods 2.1. Participants and Procedure A total of 402 Italian female workers from Italian companies were enrolled for the study. The age of participants ranged between 18 and 64 (M = 35.31; S.D. = 4.91). In total, 47.4% of them work in a STEM workplace (science, technology, engineering and math) and 56.4% in a non-STEM workplace. Out of the 402 participants, 61.7% were employed in the private sector and 38.3% in the public sector. Of the companies involved, 27% belonged to large enterprises, 32% to medium-sized enterprises, 22% to small enterprises and 19% to micro enterprises. Furthermore, the majority of the sample reported having a full-time (78%) and open-ended (57%) contract. As for geographical distribution, the sample came from Northern Italy (19%), Central Italy (27%) and Southern Italy (54%). The study was carried out in accordance with the Declaration of Helsinki, and the protocol was authorised by the Internal Ethics Committee of the Department of Education Sciences (Psychology Section) of the University of Catania (Ierb-Edunict-2020/2); the relevant research procedures followed all the guidelines of the AIP (Italian Psychology Association) and its Ethics Council. Participants were recruited through convenience sampling, using an online survey. A survey link was posted on social media platforms, such as Facebook, Twitter and Instagram, and the survey link was sent to acquaintances via email and to the HR manager of some Italian companies in the period between October 2022 and May 2023. By clicking on the link, participants received an information sheet and an informed consent form, which, once accepted, led to the survey with instructions on how to complete it. Participation was voluntary. 2.2. Measures 2.2.1. Utrecht Work Engagement Scale The Work engagement construct was measured using the 9-item Utrecht Work Engagement Scale. Schaufeli et al.’s (2003) scale, the Italian version of which is by Balducci et al. (2010), explores the experience of aspects of the construct through three item scales: Vigour (VI), Dedication (DE) and Absorption (AB). Item responses were given on a frequency scale, using a Likert scale ranging from 0 (never) to 6 (always). 2.2.2. Compound PsyCap Scale (CPC-12) The Compound Psychological Capital Scale (CPC-12), conceptualised by Lorenz et al. (2016) and translated and validated in the Italian context by Platania and Paolillo (2022), was used to assess the psychological capital construct. The psychological capital construct is inspired by positive psychology and, more specifically, positive organisational behaviour. It is defined as a set of different positive individual resources that enable and influence the capital to achieve desired goals. The 12-item scale actually evaluates four different factors (optimism, resilience, hope and self-efficacy), grouped into one general factor known as psychological capital. Each item can be answered using a 7-point Likert agreement scale. Adm. Sci. 2023,13, 210 5 of 15 2.2.3. Perceived Organisational Support Perceived organisational support was measured using the eight-item scale developed by Eisenberger et al. (1997), Italian adaptation by Battistelli and Mariani (2011). Employees were asked to rate the degree of perceived organisational support using a seven-point Likert scale ranging from 1 = strongly disagree to 7 = strongly agree. 2.2.4. Experience with Ambivalent Sexism Inventory For the measurement of experience with ambivalent sexism and hostile sexism, the Experience with Ambivalent Sexism Inventory (EASI) by Salomon et al. (2020) was chosen. Participants rated the frequency with which they encountered each of 28 hostile and benevolent sexism behaviours. All items were rated on 7-point scales from 1 = this has never happened to me to 7 = this has happened very frequently (more than twice per week). This inventory consists of four different factors: hostile sexism, heterosexual intimacy, benevolent sexism and heterosexual hostility. Specific to this study, items measuring hostile sexism (eight items) and benevolent sexism (eight items) were selected. 2.2.5. Generic Job Satisfaction Scale For the measurement of the job satisfaction construct, the Generic Job Satisfaction Scale, by Macdonald and MacIntyre (1997), was selected. The choice of this scale depended on its advantages: it is brief, easily administered in workplaces and reliable for a wide variety of occupations. The scale consists of 10 items and has a 5-point Likert agreement scale as a response option. 2.2.6. Individual Work Performance Questionnaire (IWPQ) The Individual Work Performance Questionnaire by Koopmans et al. (2015) was used to measure the concept of performance. Specifically for this study, only the items measuring task performance were selected. All participants responded to the five items using a 7-point Likert scale. Task performance can be defined as “the competence with which individuals perform substantive or technical tasks that are fundamental to their work” (Campbell 1990). The behaviours used to describe performance often include quantity and quality of work, professional skills and knowledge (Campbell 1990;Rotundo and Sackett 2002). On the task performance scale, work planning and organisation, result-oriented work, prioritisation and efficient work are evaluated. 2.3. Data Analysis Data were analysed using SPSS 27.0, and AMOS 27.0. SPSS (version 27.0 for Windows; IBM Corp., Armonk, NY, USA) was used for the descriptive analysis and bivariate correlations of the variables in this study. AMOS (version 27.0) and the structural equation Model (SEM) were used to test the fit of the measurement model, the interaction effects and the multigroup analysis. To evaluate the models’ goodness of fit, we used the comparative fit index (CFI), the root mean square error of approximation (RMSEA) and the standardized root mean square residual (SRMR). χ2 values and ∆χ2 values were also used and were presented between the competing models, and the index ∆ CFI was used with values not exceeding 0.01, indicating the equivalence of models in terms of fit (Meade et al. 2008). The structural equation model (SEM) was also used to test moderation. The ‘interaction term’ method was used, in which a product term of the independent variable and the moderator was formed. To evaluate invariance tests and the multigroup analysis, two group were extracted from the original sample: ‘younger workers’ (177, 44.03%) and ‘older workers’ (225, 55.97%). To perform this analysis and examine the differences, the sample was divided according to the commonly understood indications and subdivisions on generations; hence, it was considered the youngest workers to be those belonging to Generation Y (Millennials, born Adm. Sci. 2023,13, 210 6 of 15 between 1983 and 1994) and Generation Z (born between 1995 and 2010), while the oldest were those belonging to the Boomers—those born between 1946 and 1964—and Generation X—those born between 1965 and 1982. Afterward, the maximum likelihood (ML) estimation method was used to adjust the model individually to each group, eliminating the items that did not contribute to the adjustment quality, and then the model estimation was tested across groups (Hayes and Scharkow 2013;Byrne 2010). The data analysis involved three different steps: testing measurement model invariance; testing structural model invariance; and testing structural-path-coefficient differences. The invariance of the measurement model across the two groups was tested by comparing the unconstrained model (i.e., with all parameters free) to the model with measurement weights constrained (i.e., the measurement model per se). In the second step of this data analysis, the invariance of the structural model was tested across the two groups, considering the model with measurement weights constrained as being correct and comparing the model with the unconstrained structural coefficients to the constrained model (i.e., with structural weights constrained). Because of the results of the multigroup invariance, the differences in the structural path coefficients were investigated. 3. Results 3.1. Descriptive Statistic, Correlation and Reliability The findings in Table 1show the total mean scores for each variable of the sample. Furthermore, to estimate the internal consistency of the model, the Cronbach’s alpha coefficient was calculated for each of the variables examined. The construct validity of the model was examined in terms convergent validity and discriminant validity, which indicate the internal structure of the respective domains. Convergent validity was confirmed by the average variance extracted (AVE) and composite reliability (CR) values. The reliability analysis confirmed that the final model had a good level of internal consistency, as the Cronbach’s alpha value was above 0.8. The composite reliability values of the constructs ranged from 0.69 to 0.89, demonstrating convergent validity. Table 1also reports the results of the intercorrelations; in our study, task performance had positive correlations with work engagement (r = 0.56), psychological capital (r = 0.39) and organisational support (r = 0.26). Similarly, job satisfaction was positively correlated with the same variables (WE (r = 0.38); PsyCap (r = 0.23); POS (r = 0.19)). Quite the reverse, the variables measuring experience with benevolent sexism and hostile sexism showed negative correlations with the other constructs, except for their intercorrelations (r = 0.66). This evidence was consistent with the current literature on the topic and the insights drawn for the elaboration of the following moderation and multi-group model. Table 1. Mean scores, reliability, composite reliability, average variance extracted and intercorrelations (N = 402). αCR AVE Mean (SD) 1 2 3 4 5 6 1Work Engagement (WE) 0.91 0.89 0.75 4.5 (1.1) 1 2Psychological Capital (PsyCap) 0.90 0.84 0.81 4.5 (0.9) 0.35 ** 1 3Organisational Support (POS) 0.92 0.85 0.81 3.4 (0.6) 0.44 ** 0.44 * 1 4Experience with Benevolent Sexism 0.89 0.87 0.91 3.1 (1.42) −0.19 ** −0.15 ** −0.14 ** 1 5Hostile Sexism (ASI) 0.92 0.69 0.79 2.7 (1.10) −0.14 ** −0.09 ** −0.04 0.66 ** 1 6Task Performance 0.87 0.76 0.81 4.3 (1.3) 0.56 ** 0.39 ** 0.26 ** −0.13 ** −0.14 ** 1 7Job Satisfaction 0.89 0.79 0.84 4.34 (1.7) 0.38 ** 0.23 ** 0.19 ** −0.12 ** −0.11 ** 0.35 ** Note: ** correlations are significant at the p< 0.001 level; * correlations are significant at the p< 0.05 level. Adm. Sci. 2023,13, 210 7 of 15 3.2. Confirmatory Factor Analysis to Test the Model In order to confirm the goodness of the model hypothesised and presented in Figure 1, a comparison of two different models was performed. A series of confirmatory factor analyses (CFAs), according to Harman’s one-factor test, were run in order to diagnose the extent to which the variance of the common method could be a problem. The CFAs were performed using robust maximum likelihood estimation to examine the structure of the constructs. Two different models were tested and compared. Adm. Sci. 2023, 13, x FOR PEER REVIEW 7 of 16 4 Experience with Benevolent Sexism 0.89 0.87 0.91 3.1 (1.42) −0.19 ** −0.15 ** −0.14 ** 1 5 Hostile Sexism (ASI) 0.92 0.69 0.79 2.7 (1.10) −0.14 ** −0.09 ** −0.04 0.66 ** 1 6 Task Performance 0.87 0.76 0.81 4.3 (1.3) 0.56 ** 0.39 ** 0.26 ** −0.13 ** −0.14 ** 1 7 Job Satisfaction 0.89 0.79 0.84 4.34 (1.7) 0.38 ** 0.23 ** 0.19 ** −0.12 ** −0.11 ** 0.35 ** Note: ** correlations are significant at the p < 0.001 level; * correlations are significant at the p < 0.05 level. 3.2. Confirmatory Factor Analysis to Test the Model In order to confirm the goodness of the model hypothesised and presented in Figure 1, a comparison of two different models was performed. A series of confirmatory factor analyses (CFAs), according to Harman’s one-factor test, were run in order to diagnose the extent to which the variance of the common method could be a problem. The CFAs were performed using robust maximum likelihood estimation to examine the structure of the constructs. Two different models were tested and compared. The first model (1) included three second-order factors (work engagement, PsyCap, job satisfaction) and three first-order factors (experience with benevolent sexism, hostile sexism, task performance). The results showed that the first model (1) provided a good fit to the data: χ2 [132] = 565.896, p < 0.001, χ2/df = 4.28, RMSEA = 0.06, CFI = 0.91, TLI = 0.90, and SRMR = 0.04. In the second model (2), the CFA included the same factors, except that it considered all scales with a single-factor structure, in which all indicators were loaded on a single factor. The results of this model provided a worse fit to the data (χ2 [164] = 659.847, p < 0.001, χ2/df = 4.02, RMSEA = 0.12, CFI = 0.77, TLI = 0.69, and SRMR = 0.07). The differences were significant when comparing the chi-square values and degrees of freedom of both models (Δχ2 (32) = 93.951, p < 0.0). Model 1 showed the best fit to the data. Figure 1. Theoretical model. 3.3. Main Effects and Moderating Effects As suggested by Cohen et al. (2013), the moderation model was proposed and declined in three stages. In the first phase, the effect of the independent variables was tested, i.e., work engagement, psychological capital and organisational support on the dependent variables (job satisfaction and task performance). In the second phase, the effect of each of the moderator variables (experience with benevolent sexism and hostile sexism) on the dependent variables was tested. Finally, in the third stage, the interaction terms (each independent variable × each of the moderators) and their effect on the dependent variables were introduced. Before calculating the interaction terms, the predictor and moderator variables were centred to minimise multicollinearity between the interactions and their individual components (Aiken et al. 1991). To identify the form of moderation, when significant, the regression model was plotted at two values of the moderator variable, i.e., Figure 1. Theoretical model. The first model (1) included three second-order factors (work engagement, PsyCap, job satisfaction) and three first-order factors (experience with benevolent sexism, hostile sexism, task performance). The results showed that the first model (1) provided a good fit to the data: χ2 [132] = 565.896, p< 0.001, χ2 /df = 4.28, RMSEA = 0.06, CFI = 0.91, TLI = 0.90, and SRMR = 0.04. In the second model (2), the CFA included the same factors, except that it considered all scales with a single-factor structure, in which all indicators were loaded on a single factor. The results of this model provided a worse fit to the data ( χ2 [164] = 659.847, p< 0.001, χ2 /df = 4.02, RMSEA = 0.12, CFI = 0.77, TLI = 0.69, and SRMR = 0.07). The differences were significant when comparing the chi-square values and degrees of freedom of both models (∆χ2(32) = 93.951, p< 0.0). Model 1 showed the best fit to the data. 3.3. Main Effects and Moderating Effects As suggested by Cohen et al. (2013), the moderation model was proposed and declined in three stages. In the first phase, the effect of the independent variables was tested, i.e., work engagement, psychological capital and organisational support on the dependent variables (job satisfaction and task performance). In the second phase, the effect of each of the moderator variables (experience with benevolent sexism and hostile sexism) on the dependent variables was tested. Finally, in the third stage, the interaction terms (each independent variable × each of the moderators) and their effect on the dependent variables were introduced. Before calculating the interaction terms, the predictor and moderator variables were centred to minimise multicollinearity between the interactions and their individual components (Aiken et al. 1991). To identify the form of moderation, when significant, the regression model was plotted at two values of the moderator variable, i.e., one standard deviation above the mean and one standard deviation below the mean. The results of the moderation analyses are presented in Table 2. As displayed in Table 2, in Step 1, the perceived levels of organisational support, work engagement and psychological capital had a significant and positive impact on task performance and on job satisfaction. Conversely, in Step 2, experience with benevolent sexism and hostile sexism revealed a significant and negative impact on the same outcomes, task performance and job satisfaction, showing that it is harmful to women’s performance and wellbeing at work (Morando et al. 2023;Dardenne et al. 2007;Dumont et al. 2010;Jones Adm. Sci. 2023,13, 210 8 of 15 et al. 2014;Vescio et al. 2005). Consistent with the literature, a different reaction and effect of women to hostile and benevolent sexism were found. Indeed, hostile sexism is reported to have a significantly less strong effect than the experience with benevolent sexism because it differently impacts women’s cognitions, emotions and behaviour. Whereas experiences with hostile sexism tend to arouse anger and arousal that fade away relatively quickly, interactions with benevolent sexism tend to arouse self-doubt, rumination and low-level arousal that simmers for a relatively long time. Hence, this has a different effect on both performance and overall satisfaction related to performance and the work environment (Pacilli et al. 2019;Barreto et al. 2010;Dardenne et al. 2007;Dumont et al. 2010). Table 2. Results of moderation analysis. Task Performance Job Satisfaction Step 1 β β Work Engagement 0.31 * 0.22 * Psychological Capital 0.27 * 0.33 * Organisational Support 0.40 * 0.34 * Step 2 Experience with Benevolent Sexism −0.18 * −0.13 * Hostile Sexism −0.10 * −0.09 * Step 3 Work Engagement ×Experience with Benevolent Sexism −0.11 * −0.08 * Psychological Capital ×Experience with Benevolent Sexism −0.11 * −0.12 * Organisational Support × Experience with Benevolent Sexism −0.13 * −0.09 * Work Engagement ×Hostile Sexism n.s. −0.07 * Psychological Capital ×Hostile Sexism n.s. n.s. Organisational Support ×Hostile Sexism −0.08 * −0.11 * Note * p< 0.01, n.s. = not significant. Regarding the moderating effects of the relationship between the workplace predictors and the experience with benevolent sexism and hostile sexism, the results showed several significant and negative interaction effects on the perceived job outcomes. Indeed, experience with benevolent sexism significantly decreased the positive relationship between each of the predictors (WE, PsyCap and POS) and task performance and job satisfaction. This implies that direct experience with benevolent sexism has a harmful influence on both the individual contribution that each worker provides in his or her work and the perceived support and involvement from the company. Less significant results are found in the interaction of hostile sexism with workplace predictors. Actually, hostile sexism reduces job satisfaction in its interaction with work engagement and organisational support. 3.4. Multigroup Analysis: A Comparison of the Different Effects between Younger and Older Generation In order to examine the possible existence of differences in the structural model for two different age groups (‘younger’ and ‘older’), we performed a multi-group analysis. To investigate these group differences, ascertainment is required that differences exist between these groups and that those differences derive from structural differences in the path coefficients between the groups. In particular, prior to assessing any evidence of structural pathway equality (structural invariance test), an important check needs to be carried out as to whether the measurement parameters operate in the same way for both groups (measurement invariance test) (Cheung and Rensvold 2002;De Pasquale et al. 2022; Platania et al. 2022a,2022b;Morando and Platania 2022). As the first step of this analysis, the invariance of the measurement model was assessed across the two groups, comparing the unconstrained model (i.e., with all parameters free) with the model with measurement weights constrained to be equal across the group (Tan and Pektas 2020). Adm. Sci. 2023,13, 210 15 of 15 Podsakoff, Philip M., Scott B. MacKenzie, Jeong-Yeon Lee, and Nathan P. Podsakoff. 2003. Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology 88: 879. [CrossRef] [PubMed] Rhoades, Linda, and Robert Eisenberger. 2002. Perceived organizational support: A review of the literature. Journal of Applied Psychology 87: 698–714. [CrossRef] Rotundo, Maria, and Paul R. Sackett. 2002. The relative importance of task, citizenship, and counterproductive performance to global ratings of job performance: A policy-capturing approach. Journal of Applied Psychology 87: 66–80. [CrossRef] Rubin, Mark, Stefania Paolini, Emina Subaši´c, and Anna Giacomini. 2019. A confirmatory study of the relations between workplace sexism, sense of belonging, mental health, and job satisfaction among women in male-dominated industries. Journal of Applied Social Psychology 49: 267–82. [CrossRef] Salomon, Kristen, Jennifer K. Bosson, Mona El-Hout, Elizabeth Kiebel, Sophie L. Kuchynka, and Samantha L. Shepard. 2020. The experiences with ambivalent sexism inventory (EASI). Basic and Applied Social Psychology 42: 235–53. [CrossRef] Schaufeli, Wilmar B., Arnold B. Bakker, and Marisa Salanova. 2003. Utrecht work engagement scale-9. Educational and Psychological Measurement. [CrossRef] Schmitt, Michael T., Nyla R. Branscombe, Diane Kobrynowicz, and Susan Owen. 2002. Perceiving discrimination against one’s gender group has different implications for well-being in women and men. Personality and Social Psychology Bulletin 28: 197–210. [CrossRef] Steele, Claude M. 1997. A threat in the air: How stereotypes shape intellectual identity and performance. American Psychologist 52: 613–29. [CrossRef] Steele, Claude M., and Joshua Aronson. 1995. Stereotype threat and the intellectual test performance of African Americans. Journal of Personality and Social Psychology 69: 797–811. [CrossRef] Stevenson, Olivia, and Ashley Batts Allen. 2017. Women’s empowerment: Finding strength in self-compassion. Women & Health 57: 295–310. Swim, Janet K., Kathryn J. Aikin, Wayne S. Hall, and Barbara A. Hunter. 1995. Sexism and racism: Old-fashioned and modern prejudices. Journal of Personality and Social Psychology 68: 199–214. [CrossRef] Szymanski, Dawn M., Arpana Gupta, Erika R. Carr, and Destin Stewart. 2009. Internalized misogyny as a moderator of the link between sexist events and women’s psychological distress. Sex Roles 61: 101–9. [CrossRef] Tan, Seref, and Sami Pektas. 2020. Examining the Invariance of a Measurement Model by Using the Covariance Structure Approach. International Journal of Contemporary Educational Research 7: 27–39. [CrossRef] Tougas, Francine, Rupert Brown, Ann M. Beaton, and Stéphane Joly. 1995. Neosexism: Plus ça change, plus c’est pareil. Personality and Social Psychology Bulletin 21: 842–49. [CrossRef] Vescio, Theresa K., Sarah J. Gervais, Mark Snyder, and Ann Hoover. 2005. Power and the creation of patronizing environments: The stereotype-based behaviors of the powerful and their effects on female performance in masculine domains. Journal of Personality and Social Psychology 88: 658–72. [CrossRef] [PubMed] Yamamoto, Takehiro, and Ken-ichi Ohbuchi. 2011. Negative effects of the exposure to benevolent sexism on women’s task performance. Tohoku Psychologica Folia 70: 36–43. Yu, Chongxin, and Stephen J. Frenkel. 2013. Explaining task performance and creativity from perceived organizational support theory: Which mechanisms are more important? Journal of Organizational Behavior 34: 1165–81. [CrossRef] Zalewska, Anna M. 2020. Work-Family and Family-Work Conflicts and Health: The Protective Role of Work Engagement and Job-Related Subjective Well-Being. Medycyna Pracy 33–47. [CrossRef] Zehnter, Miriam K., Francesca Manzi, Patrick E. Shrout, and Madeline E. Heilman. 2021. Belief in sexism shift: Defining a new form of contemporary sexism and introducing the belief in sexism shift scale (BSS scale). PLoS ONE 16: e0248374. [CrossRef] Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.