【導(dǎo)讀】cognitive. users?andusers?focusonusers?

　　

【正文】 cal cluster analysis allowed us to identify clusters corresponding to the following colours: green, magenta, blue, orange and grey (see Fig. 4a and b). The inspection of Fig. 4a showed that, for the users, the most clearly discernible clusters are the ones containing the greens (100– 160176。), the magentas (280176。, 340176。), the blues (200176。, 260176。) and the greys (3176。, 4176。). The orangereds (0– 40176。) also form a cluster. Although there are other clusters, such as the one containing green (80176。) yellow (60176。) and blue (180176。), they are not homogeneous, and were therefore excluded from the analysis. The inspection of Fig. 4b showed that, for the Website designers, the results are relatively similar to those in the previous analysis, despite the different cluster boundaries. The main clusters contain the greens (80– 160176。), magentas (280– 340176。) and oranges (20176。 and 40176。). Two of the blues are in one cluster (200176。 and 220176。), and three in the other (180176。, 240176。 and 260176。). The achromatic colours (greys, black and white) also form a cluster. Second, we defined the precise hue values for these colours with regard to the representativeness of the colour (., its central position within the cluster) and a distance from the following cluster to get an interval of at least 100176。 between each colour. In addition,we added a value corresponding to the grey. As the group supplying user data was larger than the group supplying designer data, the precise values of the colours were based on the users? data, as follows: orange (20176。), green (120176。), blue (220176。), magenta (320176。) and grey (3176。). The judgments expressed by the users and designers about these selected colours (see Fig. 5) were pared by means of Mann– Whitney tests (between users and designers), and of Friedman ANOVAs (within colours for users and for designers) followed by nonparametric pairwise parisons (Wilcoxon paired sample test). The mean scores given to homepage colours by the professional designers were lower than those given by the users for the green (Mann– Whitney test, U = , p .01), the blue (U = , p .01), and the magenta (U = , p .001), and higher for the grey (U = , p .01). The difference was not significant for the orange (U = 683, p .10). The analysis of user data revealed a main effect of colour, Friedman chi2(4) = , p .001. Blue was preferred to all other colours (all z = .001). By contrast, grey was significantly less appreciated than all the other colours (all z .01), except for green (marginal significance。 z = .054). Orange appeared to be an intermediate colour, as it was appreciated more than green (z .05) but did not differ significantly from magenta. Lastly, green was not significantly preferred to magenta. 9 The analysis of Website designer data also revealed a main effect of colour, Friedman chi2(4) = , p .001. Grey, blue and orange were preferred to green and magenta (all p .001). The differences between grey, blue and orange, and between green and magenta were not significant. In short, the homepages that obtained the highest scores (or the least low scores), whatever the participants? status, were the ones presented in blue and in orange. Contrary to users, professional designers also gave similar scores to homepages presented in grey. 5 Discussion First, the professional designers appeared to be more critical of the Webpages than the users. However, blue and orange were regarded as the most appealing colours by all participants, whatever their status. Website designers also liked (or at least did not dislike) grey, unlike the users. The choice of these three colours can partly be explained by the reported associations between these colours and emotions[43,47,41,18]. The result for blue, in particular, was in accordance with previous research, insofar as blue is frequently regarded as users? favourite colour, whatever their age and cultural context ., [63– 66]. Findings for orange contrasted with previous results suggesting that cooler colours are viewed more favourably than hotter ones. However, the warm effect conveyed by the orange homepage may have explained its high score. Finally, although grey is a neutral colour, it may also be regarded, to some extent, as an innovative feature by Web designers, allowing such Websites to be perceived as trendy. This relatively discreet colour may also have enhanced the way in which the information was set out on the homepage, thereby generating a semantic effect (accessibility of information structure) rather than an aesthetic one. To take our research one step further, a second, plementary experiment was conducted in order to determine the impact of these three colours (blue, orange and grey) on users? behaviours and cognitive processes. 6 Study 2: impact of colours on users’ behaviours and cognitive processes Hypotheses Our general hypothesis was that Website users would react differently according to the colour of the site they were exploring. For this experiment, three versions of the same site, corresponding to the different colours for which participants in the previous study had expressed a preference, were shown to new participants. In line with results of the first study, we expected that the three selected colours (blue, orange and grey) would have an impact on the users? judgements. Moreover, we hypothesized that these Website colours would also exert an influence on users? 10 behaviours and cognitive processes. More precisely, we expected that these colours would have an impact on both the users? searches for information and the way they memorized it. Therefore, the main originality of this experiment lays in the fact that the impact of the colours was analyzed with regard to both objective and subjective measurements. Three types of data were analyzed: (1) data relating