All the physicians and nurses in this hospital were invited to participate in this study. The valid number of participants was 376, comprising 42 physicians and 334 nurses. Data was collected internally through the 2014 SAQ-C from JCT (Chi et al., 2017). The questionnaire had 46 items examining their attitudes towards aspects such as job satisfaction, teamwork climate, stress recognition, emotional exhaustion, work-life balance, safety climate, and perception of management. For all the aspects except work-life balance, a 5-point Likert scale ranging from 1 = strongly disagree to 5 = strongly agree was used (Chi et al., 2017). Since the work-life balance used a 4-point Likert to measure the frequency per week, it was excluded from this study. This study used linear regression with forward selection to analyze the data. The linear regression started with an empty set and continually added one attribute at a time. Only the attribute that gave the highest performance was added to the selection at each step. All statistical analyses were carried out using SPSS software version 18.
In this analysis, the predictor variable was, in this case, supervisors/managers. The α = 0.05, and the adjusted R-square values range between 0.048 and 0.138(Chi et al., 2017). According to the regression analysis, teamwork climate had a negative correlation with the supervisor/manager and the years of experience in the position. This means that nurses and physicians who are not in charge are less satisfied with the teamwork climate, job satisfaction, perceptions of management, stress recognition, safety climate, and working conditions (Chi et al., 2017). Similarly, physicians and nurses with many years of experience in their positions were also less satisfied. The regression analysis found that the supervisor/ manager and experience in the position negatively impacted the safety climate. This means that the healthcare workers who are not in leadership positions and those with many years of experience in their positions are less satisfied. Furthermore, the analysis found that job satisfaction is affected by variables such as experience in position, age, and supervisor/manager. More so, nurses tend to have less satisfaction in stress recognition.
Perceptions of management are negatively impacted by supervisor/manager and experience in the position. This means that the healthcare staff that do not hold supervisor/manager positions and have many years of experience in their positions are less satisfied with their perception of management. The dimension of working conditions was impacted by supervisor/ manager, years of experience, and age. Elderly employees were more satisfied with their working conditions, and those with more experience were less satisfied. Lastly, it was not possible to draw a linear regression between emotional exhaustion and the ten demographic variables. As per the regression analysis, nurses and physicians with supervisory/managerial positions and much experience had the greatest effect on the patient safety culture. Therefore, it is safe to conclude that the job position had some bearing on patient safety culture.
Strengths and weaknesses of linear regression
One of the strengths of linear regression, as shown in the article, is simplicity. This analysis is straightforward and provides a direct relationship between dependent and independent variables (Schober & Vetter, 2021). Also, linear regression is easy to interpret as it provides the direction and strength between variables. Moreover, linear regression makes it possible for researchers to analyze large sets of data using fewer resources compared to other analysis methods (Schober & Vetter, 2021). Besides, linear regression tests variables using tools such as homoscedasticity, linearity, and independence of errors.
However, linear regression presents several weaknesses. For instance, linear regression is increasingly sensitive to outliers to the extent that a few data points can affect the model parameters inappropriately (Schober & Vetter, 2021). Also, there are instances when linear regression assumes linearity between variables. This means that the analysis may provide inaccurate results if the true relationship is non-linear. Moreover, linear regression is considered not the most effective method for analyzing complex and non-linear relationships between variables (Schober & Vetter, 2021). More so, multicollinearity issues may occur when linear regression is used to analyze highly correlated variables. This may make it difficult to understand the real effect of each dependent variable.
Remedies to address these weaknesses
The issue of outliers can be remedied by using graphical methods to identify outliers, using other techniques such as Huber regression, and replacing extreme values with less extreme ones (Sun e
Order this paper