by Maryam M. Daroodi, PhD Student, School of Design, University of Leeds, UK
The main aim for this research is to investigate the emotions and semiotics of colour. Since colour is an important factor in the design process, this research could aid designers in their colour selection by providing localised knowledge about colour meanings. Most previous psychophysical experiments (e.g. [1-3]) were carried out in the laboratory with controlled stimulus presentation and viewing conditions. One weakness of this approach is that data are generally collected for very few observers who tend to share a common cultural background. For example, Ou’s study was based on 31 observers and Sato’s study was based on 24. How reliable are these data and the models derived from them? The novelty of my research is that we aim to collect data from the internet using a distributed experimental paradigm. The Global Online Colour Survey has been designed to allow any participant to take part, irrespective of their geographical location, age, gender or cultural background. The aim is to collect responses from many thousands of participants. However, there are two issues: firstly, how to promote the survey and drive traffic to it, and secondly, how to motivate participants to take part. The first issue has been addressed by various activities (including direct mailing to international contacts, blogging, youtube video, and other uses of social media). The second issue drove us to present each observer with only one colour (out of the 28 that constitute the study) so that the 10 questions that are asked about each colour can be completed in less than 30 seconds. It is hoped that this not only encourages participants to take part but also it increases the conserved attention for each colour. Therefore, each participant only sees a portion of the whole experiment. However, a controlled laboratory-based version of the same experiment is also running to allow comparison between lab- and web-based experiments. The disadvantage of a web-based experiment would seem to be lack of parameter control (though see ); however, the advantages include collecting responses from a wide variety of people and being able to analyse responses according to gender and/or culture.
The deeper aim of the research is to use the data collected from the web- and lab-based experiments to develop models that encode the relationships between colour and semiotics; parameters of the models could include age, gender and nationality. An interesting issue is that the semiotics of colour is undoubtedly complex and cannot adequately be represented by simple mathematical relationships of the type that have been published. New ways of storing, representing, and displaying the data may need to be developed, especially if the models/data are to be used by the design community. It is also likely that colour semiotics change with time and methods for continual collect of data and assimilation into the database may also be needed.
 Strictly speaking, the Online Global Colour Survey is only available to a subset of the population since it is provided in ten languages. However, these languages cover a substantial portion of the population who have access to the internet.
 We note that since each participant only assesses one out of 28 colours, we will need, on average, about 28 times as many participants to collect the same amount of data as in the laboratory experiment where each observer can assess all 28 colours. Since other researchers have used about 30 observers, this would suggest that the Online Global Colour Survey requires about 1000 participants to be equivalent (as of November 2011, 1600 participants have taken part).