Ultimately, the justifications lower users’ disclosure rates. The total effects (see
Figure 4) of the justifications on disclosure are all negative, which means that the
baseline system without justification actually results in the highest disclosure rates.
These effects are again roughly the same size as the effects of privacy, collection and
control concerns. The interview study reveals that users typically treat the justification
message as a warning sign: 11 participants mentioned a low percentage in a justification
message as a reason not to disclose (whereas only 5 participants mentioned
a high percentage as a reason to disclose). It seems that the justification provide more
inhibition than encouragement.
Elsewhere [Knijnenburg and Kobsa 2013], we show that these effects occur regardless
of the percentage in the justification message (except for the ‘number of others’
justification, but even for that message a high percentage merely reduces the negative
effect, and never actually increases disclosure). This is in line with the interview
study results. Participants generally had a ‘cut-off’ percentage below which they would
not disclose something. Moreover, 14 participants would at several occasions refuse to
disclose something they deemed too private despite a high percentage.
In Knijnenburg and Kobsa [2013] we demonstrate however that choosing the justi-
fication based on characteristics of the individual user may be a way to increase both
disclosure and satisfaction.