a. Sampling
Data for this study was collected from full-time employed frontline nurses aged 21 years and above, including both registered and assistant nurses of a 795-bed general and acute care hospital in Singapore – Khoo Teck Puat Hospital. To attain the objective of this study, stratified sampling method was adopted in which reported number of nurses in the hospital (n=1364) was stratified into four strata to provide a precise and representative sample based on the projected proportion of Baby Boomers, generation X, Generation Y/Millennials, and Generation Z nurses. A power analysis was performed which determined a minimum sample requirement of n=769 nurses with parameters set to achieve statistical significance level of 0.05 and power level of 0.80. To reduce sample bias considering that studies have revealed significant impact of job designation on work values and attitudes [16] , this study aims to obtain a representative sample reflecting the proportion of job designations (5% Nursing supervisors, 79% Registered nurses, and 16% Assistant nurses) as well as the proportion of generation cohorts (4% Baby boomers, 7.2% Generation X, 68.4% Generation Y, and 24% Generation Z),
b. Data collection
Data collection was performed using a two-section survey. The first section collected un-identifiable demographic characteristics of the participants including age group (generation cohort), gender, ethnicity, nationality, religion, marital status, children, designation, educational qualification, and years of working experience. The second section was a cross-sectional, self-administered survey adapted from the study by Gursoy, Chi, and Karadag [17]. The Work Value and Attitude Scale comprises 25 items reflecting seven constructs of work attitude, values and engagement - Work Centrality (five items), Non-compliance (four items), Technology Challenge (three items), Work-life Balance (five items), Leadership (two items), Power (four items), and Recognition (two items). Each of the 25 subscales within the instrument uses a 5-point Likert Scale, ranging from 1 to 5 , with score 1 representing strongly disagree, score 2 disagree, score 3 neural, score 4 agree, and score 5 strongly disagree.
Considering the restrictions to physical contact due to the COVID-19 pandemic and varying shifts of frontline nurses, the survey was transcribed and administered through an online platform - form.gov.sg. In accordance with procedures approved by ethics review board, an electronic mail was sent to all nursing supervisors within the institution to disseminate the study information to frontline nurses in all inpatient wards, accompanied by a participant information sheet and study recruitment poster which embedded a QR code for participants to assess the online questionnaire at convenience.
c. Data analysis
Quantitative data were analysed through SPSS, using descriptive statistics to analyse the demographics of the sample and multiple tests to obtain inferential statistical data. A Cronbach Alpha’s reliability test was conducted to measure reliability of each of the seven constructs in the Work Value and Attitude Scale. A Cronbach Alpha value of .60 to .70 indicates an acceptable level of reliability and .80 or greater an excellent level [18]. The Shpiro-Wilks test for normality was performed to assert symmetry of the sample means. As the result of the normality assumption test also indicated that the data are not normally distributed with p-value of <0.001 for all dimensions, non-parametric statistical tests were employed for further analysis of the data collected. Kruskal-Wallis statistical tests were then conducted to establish if there were any significant differences among different generation cohorts in each construct of the work value and attitude scale, and further elucidate if other demographical factors significantly impact nurses’ work values and attitudes. For binary variables such as gender and children, the Mann Whitney U test was employed to identify differences between groups. To identify significant differences, the Dunn’s post hoc test was performed for categorial variables that were significant which tested all possible 2-way comparisons of the medians between subgroups. Although significant differences between gender and children groups were noted in several constructs, no post hoc tests are available for binary variables.
d. Ethical considerations
This study has been approved by the National Healthcare Group Domain Specific Review Board (DSRB) and permission to conduct the study was obtained from the chief nurse of the hospital. The participant information sheet described the purpose of the study, and provided contact information of the principal investigator should respondents have any additional questions. The administered online survey was anonymous as data collected involved no data attributable to personnel identification. As explicitly indicated in the participant information sheet, completion of the online survey indicated the participants’ implied consent to participate in the study. All data collected through the online platform was encrypted end-to-end and stored in encrypted format, ensuring that only researchers can access and view responses with a private digital key. Upon completion of data collection, all data were kept in a hospital password-protected computer.