Culture, demography and individuals' technology acceptance behaviour : a PLS based structural evaluation of an extended model of technology acceptance in South-Asian country context

摘要

The models that predict the factors of individuals’ acceptance behaviour are predominantly based on a technology acceptance model (TAM) or the TAM’s conceptualisation. Although the TAM has a parsimonious structure and good explanatory power across the time, population and context, it is still criticised by a number of researchers. Categorically, it is criticised due to: inherent ‘cultural bias’ that limits its generalisability across cultures (national to organisational level); its underlying conceptualisation of predicting acceptance behaviour solely based on an ‘individual-based reactions’ that limits its applicability over the group’s effect (normative and social influence); and finally, due to its presupposition to examine the effect of ‘external variables’ through the only mediation effect of beliefs’ perceived ease of use (PEOU) and perceived usefulness (PU) that limits its ability to be extended beyond its boundaries by adding further factors directly or indirectly affecting intention behaviour (BI). To overcome mentioned limitations, an extended technology acceptance model to suit a developing country context is presented. The model attempted to delineate the direct relationship between behavioural beliefs, normative and control beliefs, management support beliefs, and task-specific beliefs towards acceptance intention and usage. In addition, the model examined the overlooked moderating impact of demographic and situational variables (age, gender, organisational type, academic position, educational level, experience usage and voluntariness) and cultural dimensions (masculinity-femininity, individualism-collectivism, power distance, and uncertainty avoidance) on indirect relations predicting acceptance behaviour. This study used a quantitative methodology to investigate the correlational paths. Using a cross-sectional survey method, data was collected from 504 academics working in 25 public and private higher educational institutions in Pakistan. Overall response rate was 53.9% (i.e. 504 out of 935). After data-screening, the final model was tested with 380 subjects. Hypothetical relationships were examined using structural equation modelling (SEM) based on the partial least squares (PLS) at the first stage, and with analysis of moment structures (AMOS) at the second stage. The indirect exploratory effect of the moderators was examined using multi-group analysis (MGA) method. The study findings indicate that the extended model achieved an acceptable fit with the data and most of the hypothetical paths were significant. Specifically, in the direct relationships, out of 20 paths representing 12 hypotheses, 11 were supported leaving 9 as unsupported. The highest variance explained by the independent variables towards dependent variables was quite similar in PU and BI (R2=26% in both using PLS; R2=34% in BI and 33% in PU using AMOS). The highest significant path was perception of usefulness, followed by academic tasks and resource facilitations towards intention; and perception of ease of use, subjective norms, and institute support towards perception of usefulness. The findings of moderating factors i.e., demographics revealed that subjects younger in age, female in gender and bachelor degree in education were influenced by the perception of ease of use, and normative beliefs; control beliefs were influenced by the organisation being private; management support at institute level was more influential in private organisations with mandatory settings; and the effect of perception of usefulness and normative beliefs was decreased with the increased experience. From the cultural perspective, the highly sensitive path was between normative beliefs and the perception of usefulness, so that the effect was stronger for subjects who were feminine in nature, collectivist in society, and high on power distance. Demographic factor academic-position and cultural dimension uncertainty avoidance did not produce any moderation effect. Finally, based on the findings, limitations and implications for theory and practices are devised.