Published online by Cambridge University Press: 01 February 2018
COGNITIVE MAPS AND AGGREGATE SIMILARITY MATRICES
The following maps and tables show the results of the Triad Test (Q20) aggregated across respondents from each university. In the triad test, for each respondent, each pair of countries (items) is given a score of 1 for each triad in which the pair are not circled (i.e. the third country is judged most different; thus, within that triad, the pair not selected are judged most similar). The maximum score from each respondent is 4, as each pair occurs four times across the sixty triads. The judgements of all students in each sample are combined and represented as a percentage in the Aggregate Similarity Matrix for each university. A score of 1.00 would indicate that the pair of countries is judged most similar every time it appears together in a triad. Conversely, a score of 0.00 would indicate that the pair was never chosen as most similar within any triad. The diagonal of the matrix is 1.00 because each country is assumed to be perfectly similar to itself.
The Cognitive Maps for each university are a visual representation of the results of correspondence analysis applied to each university's Aggregate Similarity Matrix. Correspondence analysis, a type of factor analysis, reduces the complexity of the data into ranked dimensions of similarity and difference. In this case, with ten items (countries), correspondence analysis produces no more than nine dimensions. Each country is given a score (generally between about +1.0 and –1.0) in each dimension, indicating how close (similar) or far (different) the country is from others on that dimension. The lowest (first, second, third) dimensions contain most of the information within the responses, and represent the most common ways in which judgements are being made within the sample. Higher dimensions (ninth, eighth, seventh) account for idiosyncratic (e.g. random or uninformed) answers within the data.
The Cognitive Maps visually represent the first (x-axis) and second (y-axis) dimensions of difference, mapped onto each other. Countries judged most similar appear close to one another; those judged most different appear farther from one another.