Abstract:

With the advent of the Internet, the possibilities for creating collaborative learning spaces have increased enormously. The changes brought about by this technological development have had a major impact on the way learning tasks are organised and the possibilities for collaboration. In parallel with these changes, a theoretical framework has been developed to understand the dynamics that these new spaces have created. One of the most relevant theoretical constructs in this sense is that of collective intelligence, which suggests that the ability to solve a task in a group is more effective than individually. In order to understand how these processes take place, Woolley (2015) has proposed a model for analysing the different variables involved in these interaction processes. This model distinguishes between so-called bottom-up factors, which are related to the characteristics of the participants, and top-down factors, which correspond to the characteristics of the task and the emerging dynamics of the task. The complexity of the variables included in this approach is such that it is not possible to analyse them in a single study, so it is necessary to isolate some of them for their individual study. Thus, in this paper we propose to analyse the effect of the response format generated in the tasks, one of the top-down variables, on the production of content in the collective intelligence platform Collective Learning.

The Collective Learning platform makes it possible to work together in an anonymous, synchronous and virtual way and to tackle different types of tasks. It is based on a series of collaborative phases that allow high quality answers to be obtained as a result of group consensus. Among the possibilities it offers is the possibility to ask questions in different formats, from multiple choice to open questions. One of the advantages of the former is that they are easy to quantify and check the evolution of change within the platform, but this format could condition the participants' production or create ceiling effects. Open-ended responses, on the other hand, do not have these conditioning factors, but they are more difficult to analyse in terms of content.

In order to make progress in this area, in this study we have tried to compare the evolution of the responses of 245 primary school students aged between 9 and 12 who carried out a working session within the Collective Learning platform. In this session, they analysed a situation in which a joke was created between friends and shared without consent in a Whatsapp group. The students had to answer 5 different questions about the situation, three of them in a closed multiple choice format and two more in an open text format. The results show that there is no statistically significant change in the closed questions during the interaction phases provided on the platform, but that this result changes if we analyse the length of the productions of the open answers.

Keywords:

Collective Intelligence, Collective Learning, Bottom-up factors, multiple choice, open questions, text length.