Page 187 -

156 Chapter 5 Understanding how interfaces affect users

a new question to it. Students enjoyed the experience and were more willing to con-
tinue working with the computer than were other students who were not praised by
the computer for doing the same things. In another study, Walker et al. (1994) com-
pared people's responses to a talking-face display and an equivalent text-only one
and found that people spent more time with the talking-face display than the text-
only one. When given a questionnaire to fill in, the face-display group made fewer
mistakes and wrote down more comments. In a follow-up study, Sproull et al.
(1996) again found that users reacted quite differently to the two interfaces, with
users presenting themselves in a more positive light to the talking-face display and
generally interacting with it more.

Evidence against the motion
Sproull et al.'s studies also revealed, however, that the talking-face display made
some users feel somewhat disconcerted and displeased. The choice of a stern talk-
ing face may have been a large contributing factor. Perhaps a different kind of re-
sponse would have been elicited if a friendlier smiling face had been used.
Nevertheless, a number of other studies have shown that increasing the "human-
ness" of an interface is counterproductive. People can be misled into believing that
a computer is like a human, with human levels of intelligence. For example, one
study investigating user's responses to interacting with agents at the interface rep-
resented as human guides found that the users expected the agents to be more hu-
manlike than they actually were. In particular, they expected the agents to have
personality, emotion, and motivation-even though the guides were portrayed on
the screen as simple black and white static icons (see Figure 5.8). Furthermore, the
users became disappointed when they discovered the agents did not have any of
these characteristics (Oren et al., 1990). In another study comparing an anthropo-
morphic interface that spoke in the first person and was highly personable (HI
THERE, JOHN! IT'S NICE TO MEET YOU, I SEE YOU ARE READY NOW)
with a mechanistic one that spoke in third person (PRESS THE ENTER KEY TO

Figure 5.8 Guides of histori-
cal characters.