Final Report Summary - SATISFOLLOWERS (Antecedents and outcomes of followers' perceived satisfaction with their role)
Project objectives. The project’s major aim was to investigate the important and understudied topic of followership at work, considered as complementary to leadership and equally worth investigating. The study of followership focused in particular on a new concept, followers’ satisfaction with their role (SF). Its three more specific objectives were to analyze SF’s antecedents, correlates and consequences, and thus be able to describe the nomological network of SF. To achieve these goals, the project intended to analyze pairs of leaders and followers and their potential interdependence. The study was also longitudinal and originally focused on SMEs.
Activities carried out for the project. A number of different activities were carried out to accomplish SatisFollowers goals. Listed in chronological order as far as possible, these activities included: analyzing the literature on followership and on the other topics investigated with the questionnaires; developing questionnaires on the Qualtrics platform; developing a scale to measure SF and pre-testing it using an online sample; contacting SMEs and associations which work with them (e.g. BIS, FSB, NECC, etc.) to present the project; organizing a conference at the Durham University Business School (DUBS) in July 2014 to present the project to companies; attending a course at Michigan State University (Professor D. Kashy and Dr. R. Ackerman) to learn an analysis technique for testing Actor-Partner Interdependence Models (APIMs), which is specifically suited to analyze pairs and interdependence between pairs; analysing the online sample, and presenting the topic in a symposium at the EAWOP Congress in Oslo (with a presentation entitled: Talking aloud: voice behaviors as an antecedent of followers’ satisfaction with their role) where she tested a chain of mediations using Mplus; preparing individual reports for three companies who took part in the project including comparative data collected with the online sample; analysing the partial dyadic matrix to test APIM models and to consolidate skills on this technique, and analyses on a larger sample collected at the end of the project; organizing other outreach activities for the project.
Contacting companies was found to be the major issue of SatisFollowers. Efforts to find contacts continued for the entire duration of the project, and in fact the questionnaire administration has not yet been closed: several different strategies were used to involve companies (apart from those mentioned above, these strategies include: informing potential participants about the project using the Internet and social media, e.g. using the DUBS website, its Facebook page, and newsletter; contacting a number of DUBS alumni through LinkedIn and the Alumni office; asking colleagues at the DUBS for potential contacts; organizing a raffle for respondents; presenting the project to the DBA students).
Main results. After developing the new measure for SF – which we defined and operationalized as supervisees’ satisfaction with their role, rather than satisfaction with the relationship with their leader/supervisor or with their job in general – the first question the project tried to answer was: is it a good measure? We thus pre-tested our scale on a sample consisting of 317 UK employees who had a direct supervisor. A CFA was performed using Mplus on the 6 SF items, 4 items of satisfaction with supervisor (JS_S) and 4 items of satisfaction with job (JS_J; taken from Spector’s scale, 1985/1994). The CFA yielded good results: all 6 items showed very high loadings, the correlations between the subscales were not too high and the fit indexes were fairly good [Chi-square = 329.166 with 73 df; RMSEA .10; CFI .91; TLI .89; SRMR .06]. The AVE of SF was .70. A CFA on the second sample of paired employees was performed and again showed good results. Thus, we considered the new instrument satisfactory for further research. We started to explore the outcomes of SF using the first sample we collected. For instance, we ran three hierarchical multiple regressions using work engagement (WE) and emotional exhaustion (EE) as dependent variables, and emotional demands, cognitive demands, LMX (the relationship quality between leader and follower) and SF as independent variables. In the first step of the multiple regression, we used several control variables. We found a significant impact of SF on WE and EE in the expected direction and, interestingly, we found that the coefficient of LMX dropped when SF is added in the regression, and is no longer significant. SF could then be a significant mediator of the relationship between LMX and our two outcomes. We also investigated correlations between variables to shed light on the SF nomological network. The pattern of correlations was in line with expectations, except for a non-significant correlation between SF and the dummy question which investigates the acceptance of the follower role (i.e. the “claim” made by supervisees for themselves, De Rue & Ashford, 2010). Among the antecedents, SF shows positive correlations (which are listed in decreasing order) with LMX, willingness to be a follower, positive ILTs, positive affectivity, and positive IFTs. It shows a negative correlation with negative affectivity, negative IFTs and ILTs. Among correlates, there is a strong positive relationship with both satisfaction with job and satisfaction with supervisor, a positive but lower correlation with satisfaction with colleagues. Among potential consequences, SF shows a positive correlation with work engagement, voice behaviors and motivation to lead and a negative correlation with emotional exhaustion and turnover intentions. Lastly, we scrutinized antecedents and consequences of SF on the sample of dyads/pairs we collected. A number of APIMs were tested on 112 pairs of supervisors and supervisees (who were randomly chosen by the researchers) in a number of different organizations. The APIM models showed interesting results and some very insightful interdependences. The most interesting results will not be discussed here, as the researchers are currently preparing them for publication. A single example of the type of considerations that the analysis technique allows us to make will be given here. A significant actor effect and a significant partner effect were found for job satisfaction: followers’ job satisfaction is higher if the leaders’ satisfaction in their role is higher. In greater detail, testing for nonindependence, job satisfaction between leaders and followers was moderately correlated (r .21, p .05). The actor effect of SF/SL for leaders is stronger than the actor effect for followers. This result was quite expected since the former are thinking of their “role of power” while the latter of their “subordinate role”. In addition, a positive partner effect (β .26, p .01) shows that followers are more satisfied with their job if leaders are more satisfied with their leaders’ role. The effects of actor and partner satisfaction in their role explain approximately 39% of the variance in job satisfaction (for both leaders and followers; Pseudo R2 =.397 and .395 respectively).
Skills acquired by the researcher thanks to the project’s activities. Summarizing, Dr. Gatti learnt (apart from specific knowledge about the topic): how to use Qualtrics to develop surveys and collect samples of participants on that platform; she learnt new methods such as SEMs on Mplus and APIMs on SPSS; different strategies for building contacts with companies and organizations in the UK; how to present a project in the UK, preparing the material and the presentation for such meetings; to administer questionnaires face-to-face in a UK company; networking strategies in a UK university (e.g. to develop a marketing strategy to communicate the project). Beside these new skills, the researcher improved all her skills in doing research.
Other activities carried out to improve the researcher’s career prospects. In order to further the fellow’s career and to cope with the problem of finding contacts, several other activities were planned and carried out during this time: the researcher was recommended by the scientist in charge as a reviewer for the EAWOP Journal and she wrote her first review for an international journal; she was introduced to a number of colleagues and started to work on several publications, e.g. with Professor Schyns and Professor Marc van Veldhoven, and with the former and Professor Rosalie Hall; she wrote and submitted an article on the topic of leadership with Professor Robert Lord – her formal mentor at the Business School – as the first author; the researcher was invited to give three presentations to talk about the process of writing a Marie Curie proposal; she participated in several meetings of the Leverhulme network on Implicit Leadership Theories (ILTs) and Implicit Followership Theories (IFTs); she participated in several lectures organized at the DUBS with outstanding researchers in the leadership field (e.g. Professor Mary Uhl-Bien, Professor David Day, etc.); she organized two meetings of the reading group at the DUBS on inconsistent leadership and on followership and SF as a valuable organizational outcome.
Expected final results and their potential impact and use. A first final result, which is in many respects consistent with the project’s topic, is that the experience has helped “this dyad” in establishing an enduring work relationship. Professor Schyns and Dr. Gatti, are in fact still working together collecting data to increase the sample size for longitudinal analyses. Being able to test some of the analyses performed on the first sample from a longitudinal perspective would be an excellent achievement. In any case, the APIMs on the first sample can provide information about the supervisor-supervisee relationship from a particular perspective which will offer the researchers several opportunities for publication.
Activities carried out for the project. A number of different activities were carried out to accomplish SatisFollowers goals. Listed in chronological order as far as possible, these activities included: analyzing the literature on followership and on the other topics investigated with the questionnaires; developing questionnaires on the Qualtrics platform; developing a scale to measure SF and pre-testing it using an online sample; contacting SMEs and associations which work with them (e.g. BIS, FSB, NECC, etc.) to present the project; organizing a conference at the Durham University Business School (DUBS) in July 2014 to present the project to companies; attending a course at Michigan State University (Professor D. Kashy and Dr. R. Ackerman) to learn an analysis technique for testing Actor-Partner Interdependence Models (APIMs), which is specifically suited to analyze pairs and interdependence between pairs; analysing the online sample, and presenting the topic in a symposium at the EAWOP Congress in Oslo (with a presentation entitled: Talking aloud: voice behaviors as an antecedent of followers’ satisfaction with their role) where she tested a chain of mediations using Mplus; preparing individual reports for three companies who took part in the project including comparative data collected with the online sample; analysing the partial dyadic matrix to test APIM models and to consolidate skills on this technique, and analyses on a larger sample collected at the end of the project; organizing other outreach activities for the project.
Contacting companies was found to be the major issue of SatisFollowers. Efforts to find contacts continued for the entire duration of the project, and in fact the questionnaire administration has not yet been closed: several different strategies were used to involve companies (apart from those mentioned above, these strategies include: informing potential participants about the project using the Internet and social media, e.g. using the DUBS website, its Facebook page, and newsletter; contacting a number of DUBS alumni through LinkedIn and the Alumni office; asking colleagues at the DUBS for potential contacts; organizing a raffle for respondents; presenting the project to the DBA students).
Main results. After developing the new measure for SF – which we defined and operationalized as supervisees’ satisfaction with their role, rather than satisfaction with the relationship with their leader/supervisor or with their job in general – the first question the project tried to answer was: is it a good measure? We thus pre-tested our scale on a sample consisting of 317 UK employees who had a direct supervisor. A CFA was performed using Mplus on the 6 SF items, 4 items of satisfaction with supervisor (JS_S) and 4 items of satisfaction with job (JS_J; taken from Spector’s scale, 1985/1994). The CFA yielded good results: all 6 items showed very high loadings, the correlations between the subscales were not too high and the fit indexes were fairly good [Chi-square = 329.166 with 73 df; RMSEA .10; CFI .91; TLI .89; SRMR .06]. The AVE of SF was .70. A CFA on the second sample of paired employees was performed and again showed good results. Thus, we considered the new instrument satisfactory for further research. We started to explore the outcomes of SF using the first sample we collected. For instance, we ran three hierarchical multiple regressions using work engagement (WE) and emotional exhaustion (EE) as dependent variables, and emotional demands, cognitive demands, LMX (the relationship quality between leader and follower) and SF as independent variables. In the first step of the multiple regression, we used several control variables. We found a significant impact of SF on WE and EE in the expected direction and, interestingly, we found that the coefficient of LMX dropped when SF is added in the regression, and is no longer significant. SF could then be a significant mediator of the relationship between LMX and our two outcomes. We also investigated correlations between variables to shed light on the SF nomological network. The pattern of correlations was in line with expectations, except for a non-significant correlation between SF and the dummy question which investigates the acceptance of the follower role (i.e. the “claim” made by supervisees for themselves, De Rue & Ashford, 2010). Among the antecedents, SF shows positive correlations (which are listed in decreasing order) with LMX, willingness to be a follower, positive ILTs, positive affectivity, and positive IFTs. It shows a negative correlation with negative affectivity, negative IFTs and ILTs. Among correlates, there is a strong positive relationship with both satisfaction with job and satisfaction with supervisor, a positive but lower correlation with satisfaction with colleagues. Among potential consequences, SF shows a positive correlation with work engagement, voice behaviors and motivation to lead and a negative correlation with emotional exhaustion and turnover intentions. Lastly, we scrutinized antecedents and consequences of SF on the sample of dyads/pairs we collected. A number of APIMs were tested on 112 pairs of supervisors and supervisees (who were randomly chosen by the researchers) in a number of different organizations. The APIM models showed interesting results and some very insightful interdependences. The most interesting results will not be discussed here, as the researchers are currently preparing them for publication. A single example of the type of considerations that the analysis technique allows us to make will be given here. A significant actor effect and a significant partner effect were found for job satisfaction: followers’ job satisfaction is higher if the leaders’ satisfaction in their role is higher. In greater detail, testing for nonindependence, job satisfaction between leaders and followers was moderately correlated (r .21, p .05). The actor effect of SF/SL for leaders is stronger than the actor effect for followers. This result was quite expected since the former are thinking of their “role of power” while the latter of their “subordinate role”. In addition, a positive partner effect (β .26, p .01) shows that followers are more satisfied with their job if leaders are more satisfied with their leaders’ role. The effects of actor and partner satisfaction in their role explain approximately 39% of the variance in job satisfaction (for both leaders and followers; Pseudo R2 =.397 and .395 respectively).
Skills acquired by the researcher thanks to the project’s activities. Summarizing, Dr. Gatti learnt (apart from specific knowledge about the topic): how to use Qualtrics to develop surveys and collect samples of participants on that platform; she learnt new methods such as SEMs on Mplus and APIMs on SPSS; different strategies for building contacts with companies and organizations in the UK; how to present a project in the UK, preparing the material and the presentation for such meetings; to administer questionnaires face-to-face in a UK company; networking strategies in a UK university (e.g. to develop a marketing strategy to communicate the project). Beside these new skills, the researcher improved all her skills in doing research.
Other activities carried out to improve the researcher’s career prospects. In order to further the fellow’s career and to cope with the problem of finding contacts, several other activities were planned and carried out during this time: the researcher was recommended by the scientist in charge as a reviewer for the EAWOP Journal and she wrote her first review for an international journal; she was introduced to a number of colleagues and started to work on several publications, e.g. with Professor Schyns and Professor Marc van Veldhoven, and with the former and Professor Rosalie Hall; she wrote and submitted an article on the topic of leadership with Professor Robert Lord – her formal mentor at the Business School – as the first author; the researcher was invited to give three presentations to talk about the process of writing a Marie Curie proposal; she participated in several meetings of the Leverhulme network on Implicit Leadership Theories (ILTs) and Implicit Followership Theories (IFTs); she participated in several lectures organized at the DUBS with outstanding researchers in the leadership field (e.g. Professor Mary Uhl-Bien, Professor David Day, etc.); she organized two meetings of the reading group at the DUBS on inconsistent leadership and on followership and SF as a valuable organizational outcome.
Expected final results and their potential impact and use. A first final result, which is in many respects consistent with the project’s topic, is that the experience has helped “this dyad” in establishing an enduring work relationship. Professor Schyns and Dr. Gatti, are in fact still working together collecting data to increase the sample size for longitudinal analyses. Being able to test some of the analyses performed on the first sample from a longitudinal perspective would be an excellent achievement. In any case, the APIMs on the first sample can provide information about the supervisor-supervisee relationship from a particular perspective which will offer the researchers several opportunities for publication.