*3.6. Social Comparison*

As explained in Section 2.6, the position of users in the ranking is based on their indicated preference. If people prefer upward comparison, they are shown users who perform better, and vice versa. This implies that users are always at the top or bottom position in the overview, irrespective of their performance. Although a study has shown that it matters to take the preferred direction of social comparison into account [27], the specific implementation might still allow room for improvement. A negative consequence of the current design decision is that users may become demotivated if they do not see any acknowledgment of their efforts. Therefore, a less strict selection of other users to show in the ranking might work better to motivate users through social comparison.

In addition, social comparison is more effective if you know the people you are comparing with. If users only have a few Facebook connections that are also using the system, it is likely that they mostly see anonymized other users that they do not know. Therefore, it might be better to allow adding connections via the system directly, or to invite friends to start using the system as well. Another

option could be to select similar users (in terms of occupational status, home town, gender, age, etc.) to show in the ranking, and to show and emphasize these similarities in the design, in order to strengthen the perceived closeness to the other users.

Related to this is the issue that social comparison might not be equally beneficial for all types of people. For example, it is expected that patients and individuals managing chronic conditions are not so much interested in social comparison, but could benefit from social support. Although our system targets healthy individuals, in general it is important to take such personal characteristics into account when reasoning about the specific behavior change techniques that are applied by the system to the users.

A final consideration is of ethical nature. In the current implementation, it is easily possible that user A is shown the data of user B, but not vice versa. This means that individual reciprocity of information sharing is not ensured, which could cause objections from potential users. In that case, a more sophisticated selection mechanism should be developed, in which such reciprocity is maintained.
