![www.open.ac.uk Jon Rosewell The Open University Walton Hall Milton Keynes MK7 6AA [email_address]](https://image.slidesharecdn.com/jpr-conf-diversity-eating-04-1226598089198979-9/85/Equitability-and-Dominance-in-Online-Forums-An-Ecological-Approach-69-320.jpg)
More Related Content
Similar to Equitability and Dominance in Online Forums: An Ecological Approach (20)
Equitability and Dominance in Online Forums: An Ecological Approach
- 1. Equitability and dominance in online forums: an ecological approach Jon Rosewell EATING, 13 th November 2008 Dept of Communications and Systems Faculty of Mathematics, Computing and Technology
- 2. The few and the many… Participation in forums is very variable a few students post many messages many students post a few messages
- 3. It seemed a good idea at the time… A relatively painless way of analysing FirstClass forums Looking at patterns of messages, not content quantitative, not qualitative Postings only, not ‘readings’
- 4. What is the point? Analyse past forums make sense of what went on Relate to structural features how to design the perfect forum Dashboard approach monitor the health of forums
- 5. Forums included in study Course-based: moderated, not tutors Large: 100-800 students Mainly 10 pt, 10 week intensive courses Peer support Varying purpose: chat, discussion, help, … Only some presentations analysed 4 courses, 36 forums, 3000 posters, 27000 posts
- 6. The few and the many (again) Robin Mason: ‘Rule of thirds’ 1/3 post many, 1/3 post few, 1/3 lurk read Or should it be: posters, readers, phantoms? Or for posters: post many, post some, post one or two Actually a bit more subtle…
- 8. All the data you want
- 45. Do you want to see it again?
- 47. Ecological analogy Communities often composed of dominant species plus a long tail of rare species Biodiversity a statistic of academic and practical interest
- 48. Diversity indices How to measure diversity? simple species count risky because of long tail dominance should count against diversity Diversity indices try to capture both: richness (species count) equitability (relative abundance of species)
- 49. Shannon’s index where: S total number of species (participants) p i proportion of i th species (proportion of total postings)
- 51. From May (1975) Patterns of species abundance and diversity, pp 81-120 in Cody & Diamond Ecology and Evolution of Communities , Harvard University Press
- 52. Shorrocks Genesis of Diversity after Lack (1947) Darwin’s Finches Pianka (1981) ‘Competition and niche theory’ in May Theoretical Ecology
- 53. From Whittaker (1972) Evolution and Measurement of Species Diversity, Taxon , 21 , pp. 213-251
- 55. Mark Newman Power laws, Pareto distributions and Zipf’s law Contemporary Physics, 46 , 323-351 http://www-personal.umich.edu/~mejn/courses/2006/ cmplxsys899/powerlaws.pdf
- 57. State of play No simple underlying model Not geometric: not straight on log-linear plot Not broken stick or log-normal: not sigmoid Not power law: not straight on log-log plot So use diversity index as summary
- 67. Other possible analyses Thread lengths Starters & finishers Groups Readers etc…
- 68. Is it useful? At least in subset here, patterns of participation seem remarkably consistent, so: doesn’t seem to tell us much about existing forums no clues as to what to do differently next time Dashboard not likely to be much help although simple measures might (rate of posting…) Moderators offer quicker route into feel of conference can intervene more rapidly skilled moderation more useful than statistics!
- 69. www.open.ac.uk Jon Rosewell The Open University Walton Hall Milton Keynes MK7 6AA [email_address]
- 74. The fruitees
- 75. The fruitees – how many readers
Editor's Notes
- #2: Equitability and Dominance in Online Forums: An Ecological Approach Jon Rosewell Communication and Systems Dept, Faculty of Mathematics, Computing and Technology [email_address] It is a common observation that participation in online forums by individuals varies greatly. A few students post many messages, some post a few, and many only read. A rough ‘rule of thirds’ is one way of quantifying this (for example, Mason 1989), but it is likely that this rule of thumb hides interesting structure and differences between contexts. Large scale statistical analysis suggests patterns of participation that are not captured by this rough rule of thirds. However, similar patterns can be seen when analysing the abundance of species in ecological communities, and ecologists have well-established measures to capture these patterns. It therefore seems plausible that indices of ecological diversity could also provide a useful characterisation of an online community. Such indices can unpick both ‘species richness’ (here number of participants) and equitability / dominance. There is also an ecological literature which uses these indices to characterise community types and it is possible that the same indices could provide useful insights into the structure and dynamics of online forums. In particular, it is possible that factors such as group size, topic focus, optional/compulsory activities, etc may control or correlate with measures of participation and equitability. To explore this, 36 forums containing 27,000 messages were analysed to see if an ecological approach to online communities could offer useful insights. In particular: Can simple indices capture useful features such as richness and equitability / dominance of forum contributions? Do such indices vary in a consistent way across different types of forum? Some implications raised are: Can forum organisation be manipulated to produce forums of differing richness and equitability? Are there measures that could usefully be extracted automatically to provide educators and moderators with indicators of the ‘health’ of online communities? Mason, R. (1989) ‘An evaluation of CoSy on an Open University course’, in Mason, R. & Kaye, A. (eds.), Mindweave: Communication, Computers and Education , Pergamon Press, pp. 115-145.