Improving consent in large scale mobile HCI through personalised representations of data

In using 'app store'-style software repositories to distribute research applications, substantial ethical challenge exists in gaining informed consent from potential participants. Standard 'terms and conditions' pages are commonly used, but we find they fail to communicate relevant information to users. We suggest interrupting use of an application with a visual representation of collected data, rather than merely providing a description at first launch. Data collected, but not uploaded, before this can be used to create personalised examples of what will be shared. We experiment with different ways of presenting this information and allowing opt-out mechanisms, finding that users are more concerned when presented with a visual, personalised representation, and consequently stop using the application sooner. We observe a particular difference in non-English speakers, suggesting that our proposed approach might be especially appropriate for global trials, where not all users will be able to understand researchers' disclosures of data logging intent.