Database Reference
In-Depth Information
Validity
Validity is probably the biggest concern people have—they're afraid that because a paper prototype
doesn't look realistic, the results from usability testing it aren't realistic either. I'm often asked whether
paper prototyping finds important problems. Although this is an area where more research would be
helpful, the evidence I've seen supports the generalization that paper prototypes will find the same
number and types of problems as a more high-tech method. (I discussed some exceptions in the
previous chapter .)
Research
Some people may be convinced of paper prototyping's legitimacy simply by virtue of the fact that there's
a topic on the subject. (I wish!) Other people may be swayed by the case studies and examples that I
have liberally sprinkled throughout this topic, especially because they comprise the experience of
people at many different companies. But the more scientifically minded among your colleagues may
hunger for something more substantial than anecdotal evidence. They'll identify themselves by two
words: "Prove it."
I'm not a scientist, and you probably aren't either. There is a vast body of HCI (human-computer
interaction) literature, some of which is relevant to paper prototyping. Following I provide information
from several published papers pertaining to paper prototyping, and there are more in the References
section. I'm not attempting to be exhaustive; I'm assuming that you are a practitioner who doesn't need
a lot of depth on the research, just the reassurance that there is some legitimate basis to paper
prototyping.
Although I'm not a scientist, I do respect science. It's quite difficult to design a solid experiment,
especially for something that's vague and hard to quantify, like "usability problems." I have just enough
knowledge of experimental design to know that I'm not qualified to analyze the strengths and
weaknesses of the following studies from a methodological standpoint. I can, however, attest that their
findings ring true from the perspective of someone who's done a lot of paper prototyping. Without
further ado, here are abstracts from some research studies that pertain to paper prototyping's validity.
Virzi, Sokolov, and Karis (1996)
Abstract: "In two experiments, each using a different product (either a CD-ROM based electronic
topic or an interactive voice response system), we compared the usability problems uncovered
using low- and high-fidelity prototypes. One group of subjects performed a series of tasks using a
paper-based low-fidelity prototype, while another performed the same tasks using either a high-
fidelity prototype or the actual product. In both experiments, substantially the same sets of
usability problems were found in the low- and high-fidelityconditions [emphasis added].
Moreover, there was a significant correlation between the proportion of subjects detecting
particular problems in the low- and high-fidelity groups. In other words, individual problems were
detected by a similar proportion of subjects in both the low- and high-fidelity conditions. We
conclude that the use of low-fidelity prototypes can be effective throughout the product
Search WWH ::




Custom Search