The effect of warning placement was investigated using a task-analytic approach to generate alternative warning locations. Within the context of a realistic task, 60 subjects used a consumer product that displayed a warning in one of four conditions. In two of the conditions the warnings were positioned so as to temporarily interfere with task performance. Dependent measures included attention to and behavioral compliance with the warning. The effect of warning placement was significant; the percentage of subjects who noticed the warning ranged from 0% to 93%, and the percentage who complied ranged from 0% to 53% across the four warning conditions. It was also found that warnings that temporarily interfered with task performance were significantly more effective than those that did not. The results are supportive of a task-analytic approach to warning design that pays particular attention to the cognitive aspects of task performance. The results also illustrate that, although alternative warning locations for a given product may satisfy commonly accepted placement criteria, they can differ dramatically in their effectiveness.

Warning labels can be formatted in a variety of ways. ANSI Z535.4 and ISO 3864-2 are two voluntary standards that provide guidance on label format. Two studies examined people's perception of (Study 1) and behavior in response to (Study 2) warning labels formatted in the following ways: (1) Non-standard format, (2) ANSI-format, (3) ISO-format, and (4) a combination of the ANSI- and ISO-format. The results of Study 1 show that lay people think that warning format will significantly influence people's behavior in terms of noticing and complying with warnings. However, the results of Study 2 showed that people's actual behavioral compliance is not significantly influenced by the standard warning formats compared to the Non-standard condition. Subjects did report reading more of the Non-standard label than the other labels. The results of these studies are consistent with previous findings in the warning literature. Implications of these results are discussed.

No abstract

Within the context of a realistic task (setting up an office space), 84 subjects used a two-drawer file cabinet that displayed a warning label with different numbers of statements—one, three or five. In each of these labels, a target statement, indicating that the bottom drawer was to be filled before the top one to avoid tipping, appeared first in the list. One additional warning label condition had three statements with the target statement appearing last. The warning label was affixed to the file cabinet in such a way that subjects had to interact with and remove it in order to fill the file drawer(s). Results demonstrated that almost all of the subjects noticed the label, but very few read or complied with the target warning statement regardless of its location within the list (first or last) or the number of total items on the label (one, three or five).