Web navigation design with respect to cognitive science
Download as PPTX, PDF2 likes1,578 views
This document summarizes an honors thesis project that examined how website navigation design can impact cognitive load and information intake. The project involved designing an online game to test navigation of a mock university website between an experimental and control group. Key findings included the experimental group performing better on fact and menu recall questions, indicating the navigation design supported lower cognitive load. However, the study also highlighted lessons for improving the design of the online game and survey instruments to better test the hypotheses.
Web navigation design with respect to cognitive science
- 1. Web Site Navigation Design with Respect to Cognitive LoadAimee Maree ForsstromSupervisor Dr Graham CooperHonours Presentation Southern Cross University
- 2. Interdisciplinary researchComputer ScienceEducational PsychologyCognitive Load and Human Computer InteractionDoes web site navigation design effect how we take in informationHow can we apply Cognitive Load Theory to web site navigation design to assist the intake of informationThe story so far…
- 3. Timeline - Key DatesHonours Thesis started in February 2008Original plan to run cognitive load tests on Australian Learning and Teaching Council June 2008First Honours presentation July 2008Honours update presentation October 2008Plan changed to be run on a mock up University Faculty website January 2009Plan changed to be run on a online game presenting random facts to the user May 2009Ethics received in August 2009Experiment ran first two weeks of September 2009Final Honours presentation October 2009
- 4. Research in the Social Network eraParticipants informed onlineTweets placed on www.twitter.comMessage sent to friends on www.facebook.comLink posted on www.facebook.com profile pageEmail sent to all Southern Cross University staff/studentsEmail sent to all Programming Society members (UTS)202 participants played the game124 meet needed requirements118 participants in the final clean data set 59 in control group59 in treatment groupParticipants in both groups balanced on age/gender/comp use
- 5. What we learntSocial networking is a great way to reach large number of participantsParticipants numbers coming to the site will peak and then slowly die offOnline game was a great tool to drive people through website navigationPeople don’t mind playing a game and answering some short questionsExpect a percentage of drop out’s and possible technical glitches (40% error rate encountered)Track mouse paths through all stages of the site to ensure proper human activity (not a robot)Data must be thoroughly cleaned and results checked to ensure there are no anomalies from technical issues or cheating
- 7. Playing a Game can be funOnline fact learning game design chosen15 questions presented in random fashionAnswers where found by navigating through the web site menu5 top levels of categories in menu 4 lower level menu options per categoryUsing a game we can drive the participant through the menu maximum use of menu 15 clicks with short time frame 10minEngage participant in web site material and appeal to wide audience
- 8. Experiment ComponentsResearch online checkRandom fact game play 10min, 15 questionsEffort survey Fill in facts, participant to fill in missing words from 10 facts presented in gameMenu rebuild, participant asked to rebuild the menu to show if they acquired the menu schemaUsability Survey
- 9. Random Game start page presentedParticipant asked to scroll to bottom of page and roll diceAfter the dice is rolled participant is given a question and asked to locate the section in the menu that will hold the answerThe menu sections are named after key words in the question askedIf the participant clicks on the wrong menu they are asked to re-read the question and try againOnce the final answer is read the participant is directed to answer questions about the gameExperiment Steps
- 11. StatisticsNon Parametric and parametric statistics usedDistribution was heavily skewed so Parametric statistics could not be used on most questionsQualitative and Quantitative data analysed
- 12. Survey asked on EffortTwo questionsQuestions answered by 7 point likert scale Mann Whitney U TestDescriptive Statistics
- 13. Survey asked on UsabilityTwo questionsQuestions answered by 7 point likert scale Mann Whitney U TestDescriptive Statistics Three open Ended questionsLinear regression on positive and negative word countLinear regression and on word categories
- 14. Fill in Facts Questions10 Fill in fact questions grouped by easy, medium, hardMann Whitney U test on overall score averageANOVA Question type score and group interactionVariables Easy, Medium, Hard questions, Factor group idChi-Square test for each individual question Group comparison of the number of participants who got the question rightWilcoxon Ranked Sum Test Difference in individual question answers between groupsGroup Average for Fill in facts question completion time
- 15. Menu Schema RebuildMenu Rebuild, 18 menu answers capturedMann Whitney U test on overall score averageANOVA Menu top, lower menu rebuild and group interactionVariables top, lower level menu sections, Factor group idChi-Square test for each individual menu element Group comparison of the number of participants who got the question rightWilcoxon Ranked Sum Test difference in individual menu section answers between groupsGroup Average for Menu rebuild time
- 16. Mouse path collected for each group (screen click point captured)Comparison of mouse click path used in gameOverall random facts game timeT-test of game time meansRandom Facts navigation time between each of the 15 questionsT-test of individual navigation stepsMouse Paths and Game Time
- 18. Effort Survey Likert Question OneEffort one question displays slight backward trendHow easy was it to use the menu to play the gameTreatment group mean 2.05 Control group mean 1.851 = very easy2 = easy
- 19. Statistical Significance found for Effort question twoHow distracting did you find the scrollingEffort Survey Likert Question TwoTreatment group mean 2.71 Control group mean 3.512 = very slightly distracting3 = slightly distracting4 = distracting
- 20. Usability Survey Likert QuestionsNo difference in likert scale questionsAfter rolling the dice for each question locating the navigation menu was easyTreatment group mean = 5.37Control group mean = 5.34I enjoyed playing the random facts gameTreatment group mean = 4.97Control group mean = 4.974 = neither agree nor disagree5 = slightly agree6 = agree
- 21. Usability Survey Open EndedQualitative data was analyzed for positive and negative words mentioned Experiment Group had higher positive scoreTreatment Group had higher negative score45% of the treatment group asked to change the scrolling in the game 12% of the experiment group asked to change the scrolling in the game
- 22. Fill in Facts QuestionsStatistical significance between groups for overall question scoreTreatment group outperformed the control group on overall test scoresTreatment group mean 13.56Control group mean 12.05Maximum Score for facts = 17
- 23. Fill in Facts QuestionsStatistical difference found for overall mark on question type and group idTreatment group outperformed the control group on the medium to harder questions
- 24. Fill in Facts QuestionsStatistical difference found for individual question marks between groupTreatment group outperformed control group
- 25. Menu Schema RebuildStatistical significance not found on overall scoresDifference in means 14.03 for treatment group and 12.02 for control group not significant Treatment group mean 14.03 Control group mean 12.02Maximum Score for facts = 33
- 26. Statistical significance found on number of correct answers between groupsExperiment group outperformed control groupMenu Schema Rebuild
- 27. Lessons LearntNeed twice as many questions 30 instead of 15keep participants engaged longer game play average was 5min not 10minMore elements in menu schema rebuild 30 instead of 18Need for more medium and heavy questions10/15 instead of 6/10 easy/hard question ratioEnsure wording of questions crosses bordersInclude more survey questions Perform more technical checks and pilot
- 28. Thank you