Press "Enter" to skip to content

Sensemaking distributed: using the mind of the other to give meaning to the web

But we are dedicated not only to seek out that information, but in addition we try to give you a sense, we try to interpret the information to relate the content relevant to our query. This process of giving meaning to information you receive in English the name of sensemaking.

Most of the times we tend not to make explicit this process of sensemaking that we submit the information, so that once finished off our activity, this scheme loses to other persons who may have information needs similar to ours. But, what would happen if the process of sensemaking was accessible to other internet users that have no relation to each other?; what would be of utility to those people and help them to make sense of the information related to your search?

That is what is they asked a group of researchers from Microsoft Research and Carnegie Mellon University, in a study presented at the recent Conference on Human Factors in Computing Systems, in Austin, Texas, where he examines the process of sensemaking, distributed (distributed sensemaking).

To study the potential benefits of the sensemaking distributed, the researchers needed a way to make explicit the process of sensemaking, and that different users may have access. To do this, you chose a very effective tool: the map of knowledge. In this sense, a knowledge map is an external representation and the visual result of the process of sensemaking as applied to the results of a search on the Internet: each map contains a few items of information, and the manner in which these items are related.

The study consists of two phases: in the first phase, we obtained maps of knowledge to be used by the individuals of the second phase; in this second phase, we examined the process of sensemaking distributed under different conditions, monitoring the eye movements of subjects as they viewed maps of knowledge. The methodology is dense, so as not to lose any detail, let’s look at the phases separately.

In the first phase, the researchers recruited 12 individuals of the company Microsoft to develop mental maps, from the results obtained thanks to an engine of one of the six themes follows:

  • 1. Plan a trip of more than a day in a national park
  • 2. Different options to grow a garden in our house
  • 3. Planning a small gala, with food and entertainment, for our team (30 people approx.)
  • 4. Entertaining our friends with a weekend out of the city with tourist activities,..
  • 5. Different options to be presented to a local sports event professional
  • 6. Find resources on how to reshape oneself in the kitchen

After the search, participants had to make screenshots of the results of the search (but not the web sites themselves) that they considered relevant (including images and video), to embed them in a PowerPoint slide. PowerPoint software allowed users to organize screenshots in a very simple way, which can be connected with lines and arrows, adding tags, colors, etc., to obtain a mental model of the relationships between the results received by the subjects with little effort.

Other three subjects different also invited them to perform a search on one of the six themes to develop their mental map, but were encouraged to use as a starting point the maps drawn by the subjects earlier, which could modify both as deemed necessary. This iterative process was repeated two other times in two other groups of subjects, so that the mental maps the end for each topic were prepared by four different individuals , and without contact with each other.

In the second phase of the study, was used to 21 participants, also of the company. Individuals must create mental maps for three of the subjects, but according to three conditions:

1. Start your mind map from scratch (“just a condition”)

2. Use as the starting point of a mind map prepared by another individual from phase one (“other condition”)

3. Use as a starting point one of the mind maps the end of the first phase, that is, one obtained after four iterations (“iterated condition”)

The subject of the conditions 2 and 3 were informed that the maps were to be used as a starting point had been created by “other user” (in condition 2) or by “multiple users” (in condition 3). At the end of the task, participants completed a questionnaire about their experience.

Subsequently, the participants were provided with two additional maps of different themes (created in a case by a single user, and on the other by several) but were not told what: the task of the individuals was to determine the theme of the maps by inspection. While the participants examined the maps to determine the issue, their eye movements were monitored.

The results of the study can be summarized in the following:

The authors were not able to show that the sensemaking distributed would mean a savings in time and effort to organize the information. However, the subject itself that seemed to appreciate the quality and usefulness of the maps obtained after several iterations: using a scale from 1 (low quality) to 7 (high quality), participants gave a higher score to the maps that they had created in the condition 3 (“iterated condicition”), followed in scoring by maps that had been developed from scratch (“single-condition”); in the final position, were the maps that had been created in the condition 2 (“other condition”). Thus, the participants felt that their process of sensemaking was better when they used the content already previously organized by other subjects.

This ranking is reproduced when the subjects were asked to consider how useful they believed that it would be your mind map to another user if it is asked to carry out the task that they had done. Again, the participants found the map most useful to the user would be the one created from the iterations of other users (“iterated condition”), followed by the map created from scratch (“only condition”), and finally by the map created by another user (“other condition”).

The researchers also asked the participants why they considered the maps as useful, or not: well, in most of the times a map was considered useful for their structure (tags), and not for its content (links and groups of images). To verify this idea, the authors analyzed the data obtained by monitoring the eye movements of subjects as they viewed maps: in the maps created by a single user, the subject observed over time in aspects related to the content in the maps that were created by another user; however, in the maps created by various users of the observation times for the elements of structure and the content were not significantly different. According to the authors, these results paracen imply that the iterations allow you to emerge a schema to structure the content, reducing the visual focus on the specific content (in the links).

The maps obtained after several iterations it also had another benefit for the subjects. The participants were asked if they considered that the relationship of the items of the maps was clear: the maps iterados were rated as more clear than the maps created by a single individual (“other condition”).

The study has two implications interesting. In the first place, it seems that participants preferred to start a map from scratch to use a map that had been created by another person. Then, how would it be possible to come to possess maps iterados with this limitation at the beginning of the process? The authors mention some possibilities:

Approaches to doing so include both machine methods, such as automatic alignment of first-round knowledge maps to produce maps that look more like iterated maps; as well as human methods, such as leveraging paid crowds to bootstrap the system or enforcing constraints, such as requiring users to integrate some first-round maps in order to benefit from iterated ones.

In the second place, it appears that, although the utility of the specific content of each map can change between different people, the structure of the content may remain constant and useful. However, in my opinion, the topics chosen for the study are too “procedural”, that is to say, they are focused on processes that could be broken down into steps that are understandable to different individuals in different contexts. It would be therefore interesting to determine if the utility maps are maintained if we consider other kinds of issues.


Image obtained from the blog Design Research