A Comparative Usability Evaluation of Traditional Password Managers - Information Security and Cryptology-ICISC 2010 - page 248

Ta b l e 5 . Scores for each PM with respect to PC1 and PC2

PC1 PC2
USB 0.612 -0.075
Phone -0.013 -0.244
Online -0.599 0.319

A similar and independent analysis for Students and Non-Students indicate the fol-
lowing (results shown in Figures 6(b) and 6(c), resp.). For students, USB is better com-
pared to both Phone and Online, which form a cluster together, whereby Online is better
than Phone. On the other hand, for Non-Students, USB and Phone form a cluster with
each other (Phone is better than USB) which compares favorably with Online.

4 3

3.5 3.5

2.5

3 3

2

2.5

2.5

2 1.5

2

1.5

1.5

1

1

1

0.5

0.5

0.5

0

0

0

(a) All Users (b) Students (c) Non-Students

Fig. 6. Post Test Cluster Analysis Based On Principal Components (Dissimilarity vs. PM)

In summary, our Post-Test analysis shows that, for all of our users, Online was sur-
prisingly the last choice (despite its better usability as indicated via our During Test
analysis in Section 5.1). Users either preferred USB or Phone. This can be credited to
the fact that users felt that managing their passwords locally on their own devices gives
them a sense of control and authority, as shown by their ratings for Giving Control.

5.3 Final User Preferences

After having performed the usability experiments with the PMs, we also polled for
users' final preference based on their experience. We posed the subjects three questions
and asked them to rank the PMs based of their order of preference:
1. Which password manager do you prefer the most?
2. Which password manager according to you offers better security and protection of
your passwords?
3. Which password manager according to you is most convenient to use?
The responses we received are depicted in Table 6. While a large fraction of Non-
Students shows a strong liking for Phone, most Students' preference was either USB or
Online (although most of them selected Phone to be most secure). In short, our overall
analysis, presented in Section 5.2, conforms well with the final preferences provided by
our users (i.e, All Users, Students and Non-Students).