Testing Robustness of UTAUT Model: An Invariance Analysis

Introduction

A strong theoretical model allows comparison of phenomena observed under different conditions. For instance, researchers often wish to analyze a data set that includes two or more user populations, use occasions, or technologies in use. In order to compare phenomena accurately across different conditions, however, the model’s measures must be invariant across them (Deng et al., 2008; Lee et al., 2007; Steinmetz et al., 2009). Otherwise, it is difficult to say whether any differences observed exist because of the conditions being compared, or because of differences in the measures themselves (Lai & Li, 2005; McCoy et al., 2007). Invariance analysis is a statistical test that uses multi-group confirmatory factor analysis to assess the reliability and validity of measurement instruments across individuals, groups, and contexts (Billiet, 2002; Doll et al., 1998; Jöreskog & Söbom, 1993; Klenke, 1992).

Technology adoption has long been a central issue for Information Systems researchers (Davis, 1989, Lin & Bhattacherjee, 2008, Rogers, 1995, Venkatesh et al., 2003, Zhang et al., 2008). The Technology Acceptance Model (TAM), TAM2, extended TAM, Theory of Reasoned Action (TRA), Theory of Planned Behavior (TPB), and other theories have been used to explain users’ behavior with regard to technology adoption. Recently, the Unified Theory on Acceptance and Use of Technology (UTAUT) proposed by Venkatesh et al. (2003) integrates eight theories of technology adoption and provides a comprehensive view of the factors affecting users’ adoption behavior. The UTAUT model consisted of four main constructs – performance expectancy, effort expectancy, social influence, and facilitating conditions – and four moderating variables: gender, age, experience, and voluntariness of use. In spite of the field’s broad reliance on technology adoption models, their invariance has not been extensively tested. In fact, while the invariance of TAM measurements has been assessed in a few prior studies (Lai & Li, 2005), to the best of our knowledge, the invariance of the UTAUT model has not been rigorously tested. A few studies have tested the validity of UTAUT’s measurement instruments across some conditions (Li & Kishore, 2006; Oshlyansky et al., 2007), but a full-range invariance analysis has yet to be performed.

Invariance analysis of a model is important because it can confirm that the model’s measures are invariant across conditions – ensuring that the model can be applied under different conditions without concerns about instrument reliability or validity. This study aims to demonstrate how a full-scale invariance analysis can be used for validating constructs in research models by applying it to UTAUT. The main goals of the study are:

• 1.

  To test the reliability and validity of the measurement instruments of the UTAUT model;

• 2.

  To assess the measurement equivalence of UTAUT instruments across different countries, technologies, and genders;

• 3.

  To identify the constructs responsible for any non-invariance (difference) found.