inherent trade-off in the order of the requests: requesting demographical data first
(as opposed to requesting context data first) increases demographics disclosure (β =
0.315) and decreases context data disclosure (β = −0.928), and vice versa. In other
words, asking a certain type of data first increases its disclosure, but decreases the
disclosure of the other type of data.
There are two possible explanations for this effect. One is that users become more
wary of privacy threats as the data collected about them accumulates. In the interview
study, 5 participants mentioned that at some point, they felt that the combination of
several items they disclosed caused additional privacy concerns. An alternative explanation
is that users get tired of answering so many disclosure requests. Support for
this comes from the fact that 9 interviewees mentioned that they had to answer too
many requests, and 6 participants noted that they would decline disclosure if it would
take too much effort to disclose the information.
There are reasons to believe that the former explanation holds more ground. If the
latter explanation were correct, the effect should be most pronounced for demographics
disclosure, because in the current system it takes more effort to disclose demographical
data than context data (since demographics disclosure requires the user to key in the
data, whereas context disclosure merely requires users to click a ‘yes’ or ‘no’ button).
In fact, though, the effect is stronger for context data disclosure than for demographics
disclosure (see Table II). This is in line with Acquisti et al. [2011] who also find that
the order effect is strongest for more sensitive data.