Measuring Mobile Phone Use: Self-Report Versus Log Data

In this paper we examine how well two types of self-report measures adequately operationalize frequency of mobile phone use by comparing them to server log data. Our self-report measures of voice and SMS text messaging activity are drawn from a nationally representative survey of adults living in Norway. Our results show that self-report measures only correlate moderately to the server log data. This result poses serious epistemological implications for the many influential studies that use similar self-report measures. We further conduct an exploratory multivariate analysis to examine the extent to which demographic traits are associated with under- and over-reporting. Gender, age, household size, marital status, and employment status were all found to be significant predictors of under- or over-reporting.