CHIIR2017 - Tetris Model of Resolving Information Needs
3 likes596 views
The document introduces the "Tetris Model" of resolving information needs. It uses the game of Tetris as a metaphor to show how information needs can vary in complexity, from simple lookups to more involved investigative searches. In the Tetris Model, simple information needs correspond to getting the right piece quickly to clear lines, while complex needs are like having to investigate different pieces to solve a more challenging part of the board. The model helps conceptualize how information needs can increase in depth and complexity over time as knowledge grows. However, it also has limitations and does not capture all aspects of information searching such as stages, actions, or time pressure.
![Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw
Fair Warning
R1: “This is one of the most interesting papers I have read in
some time.
[…]
I haven't felt that way about anything I have read in a long time.”
R2: “This paper reads more like a student's narrative”
but](https://image.slidesharecdn.com/tetristalk-170310092907/85/CHIIR2017-Tetris-Model-of-Resolving-Information-Needs-2-320.jpg)
CHIIR2017 - Tetris Model of Resolving Information Needs
- 1. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw TheTetris Model of Resolving Information Needs (within the Information Seeking Process) Max L.Wilson University of Nottingham, UK @gingdottwit CHIIR2017 Perspectives Paper
- 2. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Fair Warning R1: “This is one of the most interesting papers I have read in some time. […] I haven't felt that way about anything I have read in a long time.” R2: “This paper reads more like a student's narrative” but
- 3. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Main Critique R3: “it tries very hard to relate the goal of a very specific time-based problem-solving game with the concept of information search.” Which is absolutely fair, and worth considering
- 4. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw We ! Models
- 5. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Saracevic - Stratified Model
- 6. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Bates - Stratified Model
- 7. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Ingersen & Jarvelin - Cognitive Model
- 8. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw
- 9. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Dervin - Sensemaking Model
- 10. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Pirolli & Card - Sensemaking Model
- 11. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Kuhlthau - IS Process Model
- 12. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Ellis - IS Process Model (mapped against Kuhlthau’s Model)
- 13. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Marchionini - IS Process Model
- 14. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Bates - Non-Linear ISP Model
- 15. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Spink’s - Non-Linear ISP Model
- 16. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw T.Wilson - Levels Model
- 17. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Jarvelin & Ingwersen - Levels Model
- 18. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Elsweiler, M.L.Wilson & Kirkegaard Lunn - Levels Model (Casual LeisureVersion of Jarvelin & Ingwersen)
- 19. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw White & Roth - Exploratory Search Model
- 20. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw We ! Models
- 21. SDescriptive & conceptual models Explanatory & predictive models Formal (mathematical) models • Describe How people interact with search process • The aim to gain a deep understanding of the users’ ISB and/ or develop theories of such behaviour. • E.g. Bates’ Berry Picking Model • Ingwersen and Järvelin model of information seeking • Provide insight into Why people behave in certain ways and predict How people will behave under different circumstances. • The aim explain & predict search behavior e.g. querying, selecting documents, stopping and marking documents • Such models and theories formalize the relationship between the interactions of the users with the costs and gains of the IRS. • E.g. New Economic model of the Search Process • The interactive Probability Ranking Principle (PRP) model • Try to adapt the ISB models & combine them with the traditional evaluation (Cranfield-styled) measures. • The aim is to translate user models into evaluation measures • E.g. Modeling the interaction of the users with the topic summaries and predict the probability of clicking on a result • Modeling user variance in time-biased gain (Expertly presented by Maram Barifah at CHIIR2017 DC)
- 22. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Models help us to • 1) Conceptually formalise and separate aspects of the model’s focus • 2) Communicate more clearly about these aspects • 3) Create hypotheses for future research and/or interpret research results • 4) Produce implications for future systems (my first assertion)
- 23. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Limitations of Models
- 24. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Limitations of Models
- 25. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Example Problem • Marchionini’s model - good at conveying stages - but - not good at explaining exploratory behaviour
- 26. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Example Problem Its because its showing a sequence of stages And progress is aligned to one dimension
- 27. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Motivation - Describing ES
- 28. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw What I'm going to do now • Not define model, theory, framework, etc • Introduce theTetris Model • Abstraction - with a different focus • Show that it helps us think about many search experiences • And then acknowledge its limitations
- 29. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw First -TheTetris Game By CezaryTomczak, Maxime Lorant - Own work, CC BY-SA 4.0, https:// commons.wikimedia.org/w/index.php?curid=38787773
- 30. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw My Main Position • TheTetris Model captures: - the depth of complexity of an Information Need - the non-linear experience of searchers - IR, InfoSeeking, Exploratory Search in one model - non-searching Information Behaviours • It can complement other e.g. stage based models
- 31. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Tetris Model of Resolving Info Needs Complexity of Info Need
- 32. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Tetris Model of Resolving Info Needs Knowledge increases
- 33. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw So what does that mean?
- 34. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Lookup • The info need is not complex • Quick IR gets you the right piece • Info Need resolved • You have a clear board until you encounter a new Info Need
- 35. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Learn • You thought the Info Need was simple • But answer was more complicated than you expected • You realise there is more to find out • Then something helps you understand the deeper Info Need
- 36. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Learn • You thought the Info Need was simple • But answer was more complicated than you expected • You realise there is more to find out • Then something helps you understand the deeper Info Need
- 37. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Investigate • Your initial Info Need is complex • You need a few pieces to fix this • And those pieces might make the Info Need more complex (!)
- 38. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Investigate • Your initial Info Need is complex • You need a few pieces to fix this • And those pieces might make the Info Need more complex (!)
- 39. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Taking it Further • Realistically - for any domain (the game space) • We probably learned a few extra things along the way • That we maybe leave for later
- 40. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Life-Long Learning (In certain knowledge areas)
- 41. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Life-Long Learning (My board on Foreign Languages, politics, and history)
- 42. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Limitations
- 43. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Critique:Time Pressure • BUT - its interesting to think about it - we sometimes are working to a deadline - time limits in user studies DO affect behaviour - lots of research into the negative effect of time-delays etc Image By Wyatt915 - Own work, Public Domain, https:// commons.wikimedia.org/w/index.php?curid=4603015
- 44. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Critique: Random Pieces • BUT - its interesting to think about it - Information Encountering - even to encounter lots of information in SERP Image: http://ilikethesepixels.com/real-world-tetris-by-remi-gaillard/
- 45. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Critique: Resolving top down • BUT - its interesting to think about - Sometimes we DO have to figure things out in an order Image from: http://alyjuma.com/curiosity/
- 46. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Formalising & Communicating • Because progress isn't tied to a dimension - We can conceptualise the complexity of the Info Need - And discuss the idea of progress - resolving • We are all grappling with the Exploratory Search agenda -TheTetris Model helps us to communicate about it • (but you cant, of course, e.g. communicate about stages)
- 47. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Create Hypotheses for Research & Implications for Systems • We have already asked questions about time, encountering, etc • What happens to people who’s Info Need keeps getting deeper? • Can systems track Info Need depth, rather than searcher stage? - e.g. displaying results to resolve a predicted session - highlighting results that relate to previously seen info - can we highlight “the piece they need”? • Is ‘encountering new info’ the reason that Query Suggestions can be disruptive to searchers?
- 48. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Conclusions • Introduced theTetris Model of Resolving Information Needs • Captures: Info Need Complexity - ties this to Complexity of Search Experience • Everything from Look Ups to Exploratory Search • Has limitations (like all models) in what it captures - doesn’t capture stages, or user actions
- 49. Dr Max L.Wilson http://cs.nott.ac.uk/~pszmw Questions? https://xkcd.com/888/