The research problem
Tagging, or a more academic term: folksonomy, is essentially an information organization and retrieval system. Much has been said about the user behavior and social significance of tagging, but few studies have been conducted to directly compare tagging with the traditional categorization, in the context of information management.This study compared the performance of tagging and categorization. By performance we meant mental workload, satisfaction, and learning effect.
Tagging is different from categorization for its non-exclusiveness and non-hierarchy. Rashmi Sinha, the CEO of slideshare, claimed that tagging imposes less mental workload to users than categorization since it lets users annotate something with all the concepts generated for the object, while categorization forces users to pick one of those concepts. However, this claim is never verified by experiment.
Source: Rashmi Sinha: A cognitive analysis of tagging
Experiment design
Independent variables
- Using tagging / categorization to perform informatin organization and retrieval tasks
Dependent variables
- mental workload
- user subjective satisfaction
- learning effect
mental workload was assessed by two different methods. One was NASA-TLX workload measurement, which is a widely used subjective measurement of mental workload. Other one was time-estimation method, a secondary task technique of mental workload. In this experiment we asked the subject to step down a gaming wheel pedal every 10 seconds, and used the accuracy of time estimation as the mental workload measure.
Result
To our surprise, the experimental result shows that tagging imposes higher mental workload than categorization in both information organization and retrieval. This result is contrary to previous dominating viewpoint of Sinha, as well as our hypotheses.
A detailed discussion of cognitive activities and paths involved in tagging and categorization was presented to review and modify the current model. Several characteristics contribute to higher workload in tagging include the mental engagement needed to generate context-dependent tags, the need to filter and modify tags, and heavier memory usage in tagging.
Below is the revised cognitive processes for tagging and categorization.
