Measuring internet use: comparisons of different assessments and with internet addiction

Abstract

Background: The measurement of Internet use is an under-documented topic despite the worldwide growth of problematic Internet use and the recent debate in addiction research suggesting that heavy use over time should be suitable for population-based assessment. This study compared different measures of Internet use to identify the most relevant one, and tested whether Internet use was a good proxy of self-reported Internet addiction.

Methods: Data were collected in the [email protected] study among 3054 Swiss 8th grade adolescents. Different assessments of Internet use (frequency, number of hours spent on Internet, quantity-frequency measure, latent score) were compared using rank invariance across instruments. We also examined associations of Internet use with health outcomes (Internet addiction, wellbeing, somatic health problems, and quantity of sleeping). Additionally, associations of Internet addiction with health outcomes were investigated.

Results: Quantity-frequency captured 82% of the variability of average quantity and 64% of the variability of frequency, whereas the latent score captured respectively 94% and 52%. Regression models showed that frequency displayed the lowest associations with Internet addiction (b = 0.128) and wellbeing (b=−0.038). The associations of other Internet use measures were very similar. Associations of Internet addiction with wellbeing were higher than those of Internet use with wellbeing, whereas the other associations were similar.

Conclusions: The quantity-frequency appeared as the most reliable and straightforward measure of Internet use. However, Internet use seemed to be only one factor in Internet addiction. This study was a first step toward proposing a unique relevant tool of Internet use in epidemiological research.

Keywords:

• Adolescents
• behavioral addiction
• internet use
• population-based sample

Declaration of interest

The authors declare no conflict of interest.

Funding information

This work was funded by Swiss National Science Foundation (FNS 105319_140354).