Sampling techniques include both non-probability or judgmental sampling andprobability sampling (Saunders et al., 2009).Probability sampling means that each case in a group has equal and knownopportunities selected by a random method, while non-probability sampling is,selected from a specific sample in the group in a non-random manner (Saunders et al.,2009, ). Non-probability sampling is more commonly used in market surveys thanprobabilistic sampling because non-probability sampling requires less time and costthan probability sampling. Therefore, due to time and budget constraints, non-probabilistic sampling is suitable for this study. Questionnaires are sent to familiaruniversity teachers to fill in, and then they send to their colleagues. The manner inwhich this particular kind of research is drafted by associates is called snowballsampling or chain sampling.This method of recruiting research subjects throughacquaintances is called snowball sampling or chain sampling (Etikan and Bala, 2017).3.3.3 Procedures for Data CollectionConcerning procedures for data collection, Brace (2018) introduced three main types:telephone surveys, self-completed questionnaires and face-to-face interviews. In orderto achieve the research purpose, this study decided to choose the self-fillingquestionnaire method. Also, this study uses a closed questionnaire based on theInternet, inviting the respondents to complete the questionnaire by visiting thewebsite. Compared with other methods, web-based self-completed questionnaireshave the advantages of lower cost and free choice of time for respondents to completethe questionnaires, which is also more efficient than telephone or face-to-face28
interviews. However, a significant drawback is that the interviewer is unable to clarifythe problem or misunderstanding in the presence (Brace, 2018).This study sent 120 questionnaires to target groups through the Internet.Questionnaires are distributed on the WJX website to facilitate data collection. Mostof the respondents were very willing to participate in the study. I have sent out 250questionnaires,andonly 13 respondents refused to fill out the questionnaire.Finally,237 valid questionnaires were, analysed, and the feedback rate of the questionnaire isreasonable.3.3.4 Analysis of Questionnaire DataData collected through questionnaires can be, analysed by SPSS, a computer programfor statistical analysis (Brace, 2018). In SPSS,researchers input data and definevariables, then use appropriate methods for analysis. The bar chart and pie chart canbe used to explain the necessary information. The statistical methods, such ascorrelation coefficient and regression analysis, can be used to analyse the correlationbetween work characteristics, salary and welfare, training and developmentopportunities, organisational climate and employee engagement, and the correlationbetween employee engagement and employee performance.