Using Nominal Data to Examine Information Integration
PIJMEP Volume 31, Number 3, ISSN 0211-2159
Functional measurement studies typically collect numerical data in order to study judgment. The new Nanova (Nominal analysis of "variance") method allows for expansion of the paradigm to include the study of actual or projected behavior. In everyday life, people carry out actions that can be described using verbal labels, which are nominal data. Nanova is similar to analysis of variance in that significance statements assess the effect of experimentally manipulated factors. The way the methodology can extract cognitive strategies from behavioral actions is illustrated by considering a hypothetical burglar who attends to two safety features of the target homes under consideration. A real illustration is also presented, in which respondents reported both fear and projected actions in response to scenarios describing terrorist attacks. The emotional responses, reported as numbers, were analyzed with analysis of variance. The projected actions, reported nominally, were analyzed with the NANOVA computer program (Weiss, 2009). Two factors embedded in the scenarios, government announcement and public reaction, yielded similar effects on both kinds of response. Neither main effect was significant, nor were the anticipated effects of the variables obtained. With both response modes, the factors interacted significantly. (Contains 4 figures and 9 tables.)
Weiss, D.J. (2010). Using Nominal Data to Examine Information Integration. Psicologica: International Journal of Methodology and Experimental Psychology, 31(3), 441-459.