Journal of Occupational

Introduction
Job satisfaction is a major psychosocial behavioural concept that has attracted a great deal of attention in organisational research. Many definitions have been given for job satisfaction, with most of which having taken into account similar considerations, such as the employee weighting of 'input' versus 'output/reward' or 'pain' versus 'pleasure' [1, 2]. Accordingly, an individual is expected to exhibit positive or negative emotional reactions to working conditions in the work environment depending on how favourable these conditions are [3]. Employee expectations and personality traits could play a key role in determining how to produce such reactions [4]. For instance, females are generally more satisfied than males with working in a given work environment, which implies females’ expectations from their job could be less than those of their male counterparts on average [5].
There is a global decrease in the job satisfaction level among doctors, even in the western countries. For instance, doctors in Germany have been reported to be seeking to work abroad despite available internal vacancies [6]. The worst conditions exist in the developing countries with a decrease in the number of healthcare workforce [7]. The job satisfaction level has far-reaching consequences for employees' emotional wellbeing and organisational commitment. Dissatisfied healthcare providers could adversely affect productivity and the quality of care. In addition, low levels of satisfaction among patients with healthcare services, an increase in the level of morbidity, and high hospital mortality rates have been shown to be associated with low job satisfaction levels among healthcare providers [7, 8].
Job satisfaction is a complex construct measured by two general approaches. The first approach adopts single-item global measures known to be more biased because of being mostly correlated with factors intrinsic to the job; however, these measures have psychometric and practical advantages [9, 10]. The second approach adopts multi-dimensional measures that provide a more comprehensive overview of work environment aspects. Work environment facets could be differentially weighted by employees. However, they will not contribute to determining employees' response if measured by a single-item global measure. Several multidimensional job satisfaction instruments have been developed and validated to measure satisfaction with different facets of the work environment.
The Minnesota Satisfaction Questionnaire (MSQ) is a multidimensional job satisfaction instrument useful in estimating global satisfaction among employees [11]. This questionnaire is based on the theory of work adjustment which assumes work adjustment outcomes, i.e. satisfaction in this case, could be explained by how well an individual’s abilities match the work requirement and how well their needs match the reinforcement in the work environment [11]. Thus, facets in the work environment could be grouped into intrinsic and extrinsic ones. An intrinsic facet includes domains related to the job itself, while an extrinsic facet is made up of domains usually within the control of the employers, such as salary, management or team relationships [10, 12, 13]. Job satisfaction measurement across work facets is complimentary to the overall global measure of job satisfaction [14-16]. However, the mere use of single-item global measures could be misleading because they are considered biased [9].
It is suggested that the single-item global measure of job satisfaction is more correlated to the intrinsic facets; however, the facets of multidimensional measures which are assessed by the single-item global measure for predicting overall job satisfaction have not yet been well researched. Using the Spector’s job satisfaction survey, no facet of the job satisfaction measure predicted the overall single-item job satisfaction in the employees of a professional bank in Calabar, Nigeria [17]. However, using the same instrument, five facet satisfaction measures independently predicted the overall single-item job satisfaction measure among resident doctors in Calabar, including payment, contingent rewards, the operating procedure, communication, and autonomy [18]. To our knowledge, the MSQ has not yet been used in assessing job satisfaction among doctors in Nigeria, and the facet satisfaction measures that predict the overall single-item measure have not been studied as yet. Against this background, we aim to assess domain-specific job satisfaction among resident doctors in selected teaching hospitals in Southwestern Nigeria and to determine the facet satisfaction measures independently associated with the overall single-item job satisfaction measure.
 
Materials and Methods
This cross-sectional web-based survey was conducted among resident doctors practicing in four selected teaching hospitals in Southwestern Nigeria;  namely:  Ladoke Akintola University of Technology (LAUTECH) teaching hospital Ogbomosho,  Lagos  State University Teaching  Hospital  (LASUTH),  Lagos University Teaching  Hospital  (LUTH),  and University College Hospital (UCH) Ibadan. The teaching hospitals were selected based on the convenience of access to the resident doctors' electronic mails with the association executives' permission and consent, and also with the assurance of absolute confidentiality and anonymity. All available 643 resident doctors were contacted via an email that directed them to a survey link; thus, it was a full population survey requiring no sampling as all of the participants were expected to respond.
The survey tool contained some background information about the respondents, a single-item overall job satisfaction measure, and the short-form version of the Minnesota Satisfaction Questionnaire (Weiss et al., 1967). The short-form version consisted of 20 items describing satisfaction with the following facets of (1) ability utilisation, (2) achievement, (3) activity, (4) advancement, (5) authority, (6) company policies and practices, (7) compensation, (8) co-workers, (9) creativity, (10) independence, (11) moral values, (12) recognition, (13) responsibility, (14) security, (15) social service, (16) social status, (17) supervision-human relations, (18) supervision-technical (supervisor competence), (19) variety, and (20) working conditions. The short-form version of the MSQ consisted of the three scales of intrinsic satisfaction, extrinsic satisfaction, and general satisfaction. The responses to these items were weighed and scored accordingly within the range of 1-5, with 1 = very dissatisfied, 2 = dissatisfied, 3 = neither satisfied nor dissatisfied, 4 = satisfied, and 5 = very satisfied. Thus, higher scores indicated greater satisfaction. The reliability of the MSQ has been documented in an African setting. The Cronbach’s alpha for the intrinsic and extrinsic subscales were 0.79 and 0.82, respectively, while it was 0.86 for the general MSQ scale, which demonstrated reliable internal consistency [19]. The single-item overall job satisfaction measure was similarly scored as well.
The general satisfaction scale assigned scores to all 20 mentioned MSQ items, with the scores ranging from 20 to 100 [11]. The intrinsic satisfaction scale consisted of only 12 items of 1, 2, 3, 4, 7, 8, 9, 10, 11, 15, 16, and 20, with the sum score range of 12–60 [11]. In addition, the extrinsic satisfaction scale consisted of only six items of 5, 6, 12, 13, 14 and 19, with the sum score range of 6-30 [11]. Items 17 and 18 are only contained in the general satisfaction scale like other MSQ items, but not included in the intrinsic and extrinsic satisfaction scales. The scores of the MSQ items could be ranked to indicate areas of higher or lesser satisfaction rates. Past research has categorised the general satisfaction scale into moderate (score 60 and lower), moderate to not fully satisfied (the score range of 61-79), and satisfied (score 80 and higher) [20]. In addition, the percentiles for the three scales could be reported to provide norm group data. In addition, there is no information about any previously documented norm group of doctors anywhere in the world. Thus, this study used score 80 and higher for a composite estimate of global job satisfaction as it was adopted by several other studies [20-22].
Statistical analysis was conducted using SPSS Statistics V.25.0 (IBM Corp. NY, United States). Percentages were reported for the distribution of the respondents according to sociodemographic characteristics as well as for the level of satisfaction with each facet. The mean ± SD were reported for the scores of each facet and for the three scales of general, intrinsic, and extrinsic satisfaction. The Pearson's correlation coefficient was reported for the correlation between the scores of each facet and the scores of the single-item overall satisfaction measure. Besides, facets with statistically significant correlations (p ≤ 0.05) were entered into a linear regression model of the overall single-item job satisfaction measure. The effect of the facets on overall job satisfaction were controlled for age and work experience. The control of age and work experience is justified on the basis of the fact that an increase in age has been consistently associated with higher job satisfaction levels in clinical medicine [23, 24]. In addition, work experience was known to have influenced the job satisfaction level [25]. P-value ≤ 0.05 was also reported as a statistically significant predictor of overall job satisfaction.
 
Results
A total of 195 doctors responded to the survey, out of which eight questionnaires were unusable and did not contribute data to the survey, with the response rate having been 30.3%. The respondents' mean age was 36.49 ± 5.414, among whom 60% aged 30-39, 114 (60.9%) individuals were female, 109 (58.1%) individuals were male, and 114 (61.0%) individuals worked at UCH. About three-quarters (143; 76.4%) were married, more than two-thirds of the respondents were senior registrars (131; 69.8%), and 155 (83.2%) of them had at least six years of work experience since graduation from the medical school. The mean year of work experience since graduation was 9.81 ± 4.2. More than 155 (80%) individuals had worked for more than five years. The more commonly represented specialties included surgery (37; 20.0%), internal medicine (34; 18.2%), community medicine (29; 15.3%), and family medicine (25; 14.7%) (Table 1).

The mean score of the single-item overall job satisfaction was 2.72 ± 1.056. Only 50 (26.7%) of the respondents expressed they were satisfied (45; 24.1%) or very satisfied (5; 2.7%), in general, with their jobs. The other categories of the responses included 'neither satisfied nor dissatisfied' (55; 29.4%), 'dissatisfied' (57; 30.5%), and 'very dissatisfied' (25; 13.4%). On the general satisfaction scale with score 100, the 25th, 50th, and 75th percentile scores were 51, 58, and 65, respectively. In addition, only 1.9% of the respondents were fully satisfied based on the cut-off of ≥ 80, which in fact included the fully satisfied respondents. When the undecided group was added to the fully satisfied respondents, a cut-off of about 47.5% was created for the moderate to not fully satisfied and fully satisfied respondents.
Concerning the facet-specific satisfaction scores, the mean of the general satisfaction scale was 57.68 ± 11.32, the mean of the intrinsic factors was 37.40 ± 7.70, and that of extrinsic factors was 15.48 ± 3.856. Table 2 shows the distribution of the responses for each MSQ domain as well as the mean scores for each work facet. The respondents expressed an above average level of satisfaction in only two domains, including compensation (56.7%) and creativity (52.5 %) (Table 2). In contrast, the level of satisfaction was very low for the facets of recognition (4.8%), supervision-human relations (5.3%), and responsibility (7.5%).

The single-item overall job satisfaction score had a significant positive correlation with scores for each domain of the MSQ instrument (Table 3). Similarly, it had a significant positive correlation with the general sum score (r = 0.625, p < 0.001), the sum score for intrinsic factors (r = 0.596, p < 0.001), and the sum score for extrinsic factors (r = 0.476, p < 0.001). However, the correlation between the single-item overall satisfaction score and the general domain sum score was much stronger than the correlation between the single-item overall satisfaction score and the individual scores of the MSQ domains.

The model of the single-item measure was a dependent variable, and all domains were independent variables. With the age and years of work experience kept constant, it was shown that only ability utilisation (p = 0.005), advancement (p < 0.001), supervision-human relations (p = 0.035), variety (p = 0.012), and working conditions (p = 0.001) significantly predicted overall job satisfaction in the assessment made by the single-item measure (Table 4). For a unit increase in the satisfaction score for ability utilisation, there was a unit increase of 0.208 in the overall job satisfaction score as assessed by the single-item measure, with all other variables kept constant. Similarly, for a unit increase in the satisfaction score for advancement, there was a corresponding unit increase of 0.314 in the overall job satisfaction score. In addition, for a unit increase in the satisfaction score for supervision-human relations, there was a unit increase of 0.222 in the overall job satisfaction score, with all other variables kept constant. Finally, for a unit increase in the satisfaction scores for variety and working conditions, there were the unit increases of 0.235 and 0.299, respectively, in the overall job satisfaction score, with other variables kept constant.