Empirical research on mobile Internet usage: A meta-analysis of the literature

• Empirical literature on mobile Internet (MI) usage intensity level and its determinants is synthesized.
• 175 investigations published between 2001 and 2012 are reviewed.
• Correlations between customer characteristics, MI attribute perceptions/behavioral intentions and factual use conditions and MI usage are meta-analyzed.
• Moderator analysis suggests that the approach to measuring MI usage behaviors (subjective versus objective) and geographic origin of a sample (Asia versus Europe/USA) had significant impacts on relationships between a number of predictors and MI usage.

This paper reviews 175 scholarly empirical publications on mobile Internet (MI) usage intensity levels and potential determinants of respective usage behavior at the individual subscriber level. Based on an overview of MI measurement approaches and units evidence on the development of the average MI data volume generated per subscriber over the last years is summarized. Interindividual variance in MI usage was found to be very large. This raises the question which factors contribute to explaining MI usage differences. A qualitative review and a meta-analysis of correlations between 22 variables grouped into four categories (country, personal user characteristics, MI attribute perceptions/behavioral intentions and factual use conditions) on the one side and MI usage criteria on the other were carried out. Enjoyment, behavioral intention to use MI, educational level, subscription of a flat rate and ease of use turned out to be the five antecedents with the largest sample size- and measurement error-corrected average correlations r¯a with MI usage criteria. Moderator analysis indicated that the approach to measuring MI usage behaviors (subjective versus objective) and geographic origin of a sample (Asia versus Europe/USA) significantly altered relationships between a number of predictors and MI usage. Specifically, mono-method work which measures both explanatory factors and MI usage in one questionnaire was prone to overestimate relationships between independent study variables and self-reports of MI usage. The review derives open research issues in three territories, namely measurement of MI usage, choice of MI usage determinants as well as study sampling and design.