Global Journal of Management and Business Research, E: Marketing, Volume 23 Issue 1

IV. R esearch M ethodology a) Measurement of Constructs To test the proposed model, a survey was conducted introducing all the constructs’ measurements statements of the conceptual model. The measurements items for all the constructs were adapted from the previously validated scale for ensuring the content validity. Considering the context of the study, some items were modified to fit the context. All the measurement items and their sources are listed in the appendix B. b) Questionnaire Design, Sample Size and Data Collection A well-structured questionnaire was developed in two parts. First part of the questionnaire contains the demographic information of the respondents. At the very first of the questionnaire it was mentioned that the respondents are free to answer and they can withdraw their responses anytime from online submission. Further, at the last part of the first part of the questionnaire it was also mentioned that the respondents with no experience of using WFT to not to attempt to answer second part since the study only considered experienced users of WFT. Experience users are critical for truly depicting the picture of understanding influential factors as well as the differential effect of in deciding to adopt and use WFT. However, the questionnaire was developed on Google Docs and disseminated using email as well as social media platform such as Messenger, Whats App, IMO, etc. This method of data collection is best suitable during the pandemic as well as ease of use, and time- saving. However, sample size was determined following the recommendation of MacCallum, Widaman, Zhang, and Hong (1999) who recommended that respondents to construct ratio is 20: 1. Our proposed model contains 10 constructs. Accordingly, a minimum sample size of 200 is suitable for this study. However, following the convenient sampling technique, the data were collected from 300 respondents for better explanatory precision of the model. Some responses were found problematic because of non-response bias, outliers problems etc. Finally, 288 responses were retained to test the relationship between variables and a set of hypotheses. c) Data Analysis Technique This study is an attempt to explore the key influential factors determining the use of WFT as well as to investigate the differential effects of age on the influential factors and dependent constructs. This study used Partial Least Squares-Structural Equation Modeling (PLS-SEM) to investigate the measurement and parameters of the constructs and the relationships between exogenous and endogenous variables. PLS- SEM is widely used in information system research (Marcoulides and Saunders, 2006). An important advantage of PLS-SEM is that the small sample size could be operated in PLS-SEM. PLS-SEM is more appropriate for the analysis like determination of influential factors (Hair et al., 2016). As a PLS-SEM technique, this study employed SmartPLS 3.0 to analysis the data. Further, this study also employed SPSS 23 for determining the reliability and validity of the data. V. R esults a) Demographic Properties of the Respondents Out of 288 respondents, 54.51% were males and 45.49% were females. 51.40% respondents were less than the age of 40 and rest of them were 41 to 55 years. The majority of the respondents were service holders (41.66%). 30.23% respondents were engaged in business and 28.11% were students. The majority of the respondents completed their Masters (52.08%), where as 30.21% respondents completed their Honors. Few of the respondents completed Higher-Secondary School Certificate (13.55%) and PhD (4.16%). 40.97% have been using WFT for less than a year, whereas 34.02% respondents have 1 to 2 years of using WFT. 24.99% respondents have 2 years of experience using WFT. b) Common Method Bias (CMB) Since the data is self-reported, CMB was inspected to circumvent the future risk to the outcome of the analysis. For examining CMB, primarily, this study checked the Harman’s single factor test and found that 27.54% variance is explained by a single factor from the overall variance where Podsak off & Organ (1986) suggested that less than 50% indicates no presence of CMB problem. Correlation matrix was also checked and the correlation matrix indicates that there was no correlation greater than 0.90, a sign of absence of CMB. In addition, the CMB issues was also investigated based on Variance Inflation Factors (VIF) values (Table 1). All the VIF values range was 1.371 to 2.617 which are less than 3.3 (Kock, 2015). The statistical evidence indicates that CMB is not an issue for this study. c) Measurement Model Validation PLS algorithm was carried out in the SmartPLS3.0. One-tailed test with 0.05 significance level was employed. Based on the result, we checked the internal reliability, convergent validity, and discriminant validity for assessing the measurement model fit and validation (Hair et al., 2013). Internal reliability was ensured by assessing Cronbach's Alpha and Dijkstra- Henseler's rho ( ρ A). Cronbach's Alpha values for all constructs ranged from 0.752 to 0.894 (Table 1) and Henseler's rho values ranged from 0.786 to 0.901, indicates the reliability criteria successfully surpassed (Henseler et al., 2009). Understanding the Age Differences in Adopting WFTs: An Extension of the UTAUT2 Model 6 Global Journal of Management and Business Research Volume XXIII Issue I Version I Year 2023 ( )E © 2023 Global Journals