Setting and participants
A nationwide cross-sectional multicentre study was conducted using a total of 329 public health students undertaking postgraduate education as part of their Master’s degree. The study was conducted from January 16 to February 28, 2018.
We invited 31 universities, both public and private, to participate in the study. In 2017, there were a total of 1362 students [20] in Master’s degree programs (Level 7 of the European Qualifications Framework). Eight universities decided to participate in the study. In total, 653 students undertaking in-service training in Master’s degree programs were qualified for the study, and complete data sets were obtained from 329 students (response rate 50.38%). With this sample size and the number of public health students in Master’s degree programs in Poland (N = 1362), the error margin was 3.17% (95% confidence level).
Educational context
In Poland, public health education is conducted in line with the Bologna Process. Higher education programs are divided into: Bachelor’s degree studies (first-cycle studies), Master’s degree studies (second-cycle studies), and doctoral studies (third-cycle studies). These programs are conducted independently of each other, and universities have the ability or organize these cycles [21].
During second-cycle study programs, students need to achieve specific learning outcomes in terms of knowledge, skills, and social competences. Having completed a second-cycle study program, the graduate attains the 7th level of the European Qualifications Framework and obtains the professional title of Master of Public Health [22].
During second-cycle studies, students can choose a major (specialization path) that is interesting to them. Due to the non-regulated nature of study programs at the faculty of public health, each university determines its own curriculum and offered specializations. In Poland, there is a wide range of specialization paths, e.g., health education and social marketing, healthcare analytics, clinical research and health technology assessment, or epidemiology with health promotion elements [23] as well as European public health and lifestyle medicine and management in healthcare [24].
Measures
Four research tools were used in the study
The Authentic Leadership Self-Assessment Questionnaire (ALSAQ) developed by Walumbwa et al. [8] and recommended by Northouse [25] was used to perform the self-assessment. The Polish version of the ALSAQ, validated by Panczyk et al. [19], includes 16 items and allows the global indicator of Authentic Leadership skills and its three components: moral processing, self-awareness, and relational transparency to be measured. The Polish version of the ALSAQ has a good internal consistency (Cronbach’s alpha 0.84) and a test−retest analysis confirmed the stability of the measurement for the subscales and particular items. In our study, we only analyzed the global level of authentic leadership skills, i.e., the sum of the points from the three subscales listed above. We chose not to analyze these subscales separately because they correlate with one another.
The Moral Foundations Questionnaire (MFQ) was developed by Graham et al. [26]. The MFQ measures five universal moral foundations of harm/care, fairness/reciprocity, ingroup/loyalty, authority/respect, and purity/sanctity. The codes provide the basis for the evaluation of one’s behaviour for mortality [26]. We used the Polish version of the MFQ-PL questionnaire, which has good validity and reliability [27]. In order to ensure the validity of the psychological test, the study participants provided answers to all of the questions on the MFQ-PL questionnaire. However, based on the literature review, we only took the results from two subscales, i.e., harm/care and fairness/reciprocity, in account. The harm/care (KODT) subscale refers to empathy and compassion and the principles of not hurting other people and helping those who are weaker and people in need [28]. This subscale was selected for the study because it reflects an empathic attitude, which is a key aspect of authentic leadership [7]. The fairness/reciprocity (KODS) subscale refers to the reciprocity of help given and helping others in contrast to taking advantage of other people and only feigned involvement in action [28]. This subscale was chosen for the study because it is consistent with the general concept of authentic leadership, which involves treating others equally, providing equal opportunities, and acting for the benefit of the group [7].
The General Self-Efficacy Scale (GSES) was originally developed by Matthias Jerusalem and Ralf Schwarzer in 1981 and was designed to assess optimistic self-beliefs and the ability to cope with a variety of difficult situations in life. It is a short 10-item psychometric scale [15]. The scale is one-dimensional and enables a global self-efficacy measurement. We used the Polish version of the GSES, which has good validity and reliability [29].
The Youth Leadership Life Skills Development (YLLSD) Scale was developed by Seevers et al. based on Miller’s concept of leadership life skills development [13]. The YLLSD Scale contains 30 items from seven domains (communication skills, decision-making skills, skills for getting along with others, learning skills, management skills, skills for understanding yourself, and skills for working with groups) that together form a complete picture of leadership skills. The final summated scale of 30 indicators had a Cronbach’s alpha reliability coefficient of .98. In our study, we chose to only analyze decision-making skills, as these skills appear to be crucial for public health specialists [1].
Additionally, the research tool was supplemented with a question concerning declared participation in social skills training. For this purpose, the following yes/no question was asked: Have you participated in training/workshops on non-technical skills (e.g., leadership, communication, social competences, etc.)? Based on the answers to this question, the study group was divided into two subgroups. The first group comprised students who declared that they had participated in at least one life skills training session (extra LST group), while the second group comprised students who declared that they had not participated in such training (no extra LST group).
Model assumptions
Based on the formulated hypotheses (H1a–H1c), a theoretical model was developed (Fig. 1) by assuming the impact of self-efficacy (GSES) and fairness in group cooperation (F) on decision-making skills (DMS). Moreover, DMS influence the development of authentic leadership skills (ALS). Empathic attitudes (C/E) are an important addition to this model, as they directly affect ALS and interact with F.
Data collection
The data were collected by means of a paper questionnaire distributed among a group of students at the end of regular university classes. Trained interviewers limited themselves to stating the aim of the study and informing the participants how to fill in the questionnaire. They were also responsible for collecting the completed questionnaires and securing them prior to sending them to the central unit coordinating the study. ABBYY® FlexiCapture version 9.0 software was used to digitize the paper questionnaire data. Questionnaires with missing data were rejected and were not included in the analysis.
Data analysis
In order to analyze the variables collected in the study, we used descriptive statistics (mean, standard deviation) and structure coefficients (numbers and frequency). The chi-square independence test and Student’s t test were used to compare the two subgroups (extra life skills training group vs. no extra life skills training group) in terms of the examined characteristics depending on the type of variable (categorical or continuous variables, respectively). The calculations were performed with the use of the STATISTICA package, version 13.3 (Tibco Software Inc., Palo Alto, CA, United States). A 5% level of significance was set.
All analyses were carried out using the structural equation modelling software program Mplus version 7.0 [30]. We used two-group structural equation modelling: extra LST vs. no extra LST groups. The aim of this analysis was to determine whether the relationships between the variables that were assumed theoretically would be confirmed by the collected empirical data. For this purpose, the model parameters (path coefficients, variance, and covariance) were estimated and used to build the theoretical variance-covariance matrix of the variables used in the model (Fig. 2). We verified whether the calculated model parameters differed in the extra LST vs. no extra LST groups. Maximum likelihood estimation with robust standard errors was used to calculate the parameters of the structural model.
The fit of the model was assessed via the following statistics and indices: the chi-square test of model fit (CMIN), normal chi-square (CMIN/DF), the Comparative Fit Index (CFI), the Tucker–Lewis Index (TLI), the Root Mean Square Error of Approximation (RMSEA), and the Standardized Root Mean Square Residual (SRMR). For the evaluation of the model, the chi-square statistics were expected to be nonsignificant. Both the CFI and TLI evaluate the fit of a user-specified solution in relation to a more restricted nested baseline model, in which the covariance among all of the input indicators is fixed to zero or has no relationship among the variables that are posited; in other words, the number of dependent variables is equal to the number of factors. The TLI imposes an additional correction for over-parameterization [31]. The expected values of recommended indices were as follows: χ2 divided by the degrees of freedom (CMIN/DF) ≤ 3.00; RMSEA < 0.080 and SRMR < 0.050; and CFI and TLI > 0.95 [32].