App stores allow globally distributed users to submit user feedback, in the form of user reviews, about the apps they download. Previous research has found that many of these reviews contain valuable information for software evolution, such as bug reports or feature requests, and has designed approaches for automatically extracting this information. However, the diversity of the feedback submitted by users from diverse cultural backgrounds and the consequences this diversity might imply have not been studied so far. In this paper, we report on a cross-cultural study where we investigated cultural differences in app store reviews and identified correlations to cultural dimensions taken from a well-established cultural model. We analyzed 2,560 app reviews written by users from eight countries with diverse national culture. We contribute evidence about the influence of cultural factors on characteristics of app reviews. Our results also help developers of automated feedback analysis tools to avoid cultural bias when choosing their algorithms and the data for training and validating them.
ABSTRACTApp stores allow globally distributed users to submit user feedback, in the form of user reviews, about the apps they download. Previous research has found that many of these reviews contain valuable information for software evolution, such as bug reports or feature requests, and has designed approaches for automatically extracting this information. However, the diversity of the feedback submitted by users from diverse cultural backgrounds and the consequences this diversity might imply have not been studied so far.In this paper, we report on a cross-cultural study where we investigated cultural differences in app store reviews and identified correlations to cultural dimensions taken from a well-established cultural model. We analyzed 2,560 app reviews written by users from eight countries with diverse national culture. We contribute evidence about the influence of cultural factors on characteristics of app reviews. Our results also help developers of automated feedback analysis tools to avoid cultural bias when choosing their algorithms and the data for training and validating them.