Prof. Greg Francis 5/23/081InterferenceIIE 269: Cognitive PsychologyGreg FrancisLecture 18How to take a test.Working memory We spent a good deal of timediscussing working memory Currently dominant theory ofmemory Not the only theory Today we discuss an alternative Not mutually exclusive!CentralexecutiveVisuo-spatialsketchpadPhonologicalloopRetention Revisit: Peterson & Peterson (1959) Brown (1958) Give subjects trigram ask them to count backwards by 3’s andthen recall trigramWRM 782779, 776, 773,...Retention Vary duration of counting backward Test recall00.20.40.60.810 3 6 9 12 15 18Retention interval (seconds)Proportion correctSuggestsmemorieslast only a few seconds!Interference Implications memory performance is a combination ofactive rehearsal and storage in STM (think ofthe phonological loop) if active rehearsal is prevented by thecounting-backward task, the items decay fromSTM Interference is at a process level counting interferes with rehearsalInterference How might things interfere? Block necessary processes prevent learning erase / overwrite exhaust resources prevent recallProf. Greg Francis 5/23/082Interference Theory neutral terms just describe effects Retroactive interference (RI) new information prevents recall ofprevious information e.g., Overwriting a computer file. Proactive interference (PI) prior learning prohibits new learning e.g., Learning new cultural customs.Serial position curve Suggests explanation by interference more interference -> poorer recall compare to STM-LTM explanation00.20.40.60.811 2 3 4 5 6 7Position in listProbability of recallPRIMACY:RIRECENCY:PIBOTH RI AND PIImpliesthat RI isstrongerthan PI.Proactive interference Found in lots of studies subjects do worse in later trials compare to expected practice effects!0204060801000 5 10 15 20 25Number of previous listsPercentage recalledDemonstration of PI Brown-PetersontaskKQR -- 824GHN -- 721JPM -- 632FTN -- 546RBM -- 197KHT -- 54Proactive interference When the experiment also varies retentioninterval Keppel & Underwood (1962)0204060801001200 10 20Trial 1Trial 2Tiral 3Retention interval (seconds)Percentage correctProactive interference Implications STM account of Brown-Peterson result cannotbe completely correct» hypothesized decay of items from STM Not correct because Keppel & Underwoodresult would imply that decay does not occur onthe first trial, but does on later trials» that doesn’t make sense with the Brown-Peterson concept of STM as a passiveholder of informationProf. Greg Francis 5/23/083Proactive interference I took the data from CogLab for our class andanalyzed the first few trials I get a plot like:00.10.20.30.40.50.60.70.80.911 11 211st2nd3rd4Release from PI Proactive interference weakens for differentstimulus types Run two experimentsXJFWRMDBLNRXControlXJFWRMDBL942ExperimentalTrial 1Trial 2Trial 3Trial 4Release from PI Trials 1-3 show build up of PI Experimental group shows release of PI onTrial 40204060801001 2 3 4ControlExperimentTrialPercentage correctRelease from PI Works for all kinds of memory tasks All kinds of stimuli5, 7, 9, 1vs5, 7, 9, HANDNICE, SUNNY, ENJOY, PUPPYvsNICE, SUNNY, ENJOY, KILLD, D, D, DvsD, D, D, FNews storiesMemory system Every memory system must have at least twocomponents/processes Storage retrieval Would like to know if interference prevents items from being stored and thereby makes themunrecallable prevents stored items from being recalledWorking memory Working memory has a storage interferencehypothesis for the phonological loop Working memory suggests that interference canoccur by blocking ACP rehearsal (articulatory suppression,Brown-Peterson task, word length effect) within the PS when items sound similar both of these block the storage of items (items fall out of theloop)Prof. Greg Francis 5/23/084Testing models Test location of PI by changinginstructions after the list is presented Experiment stimuli are names of indoor and outdoorgames subjects usually do not notice that wordon the fourth trial is an indoor game andothers are outdoor games Take two groups of subjects one has traditional PI type experiment one is told of difference on fourth trial, atthe time of testFOOTBALLSOCCERBASEBALLWALLYBALLTrial 1Trial 2Trial 3Trial 4Interference at recall If PI prevented the last item from being stored your telling asubject that the fourth item was an indoor sport, shouldmake no difference (other than guessing) but it makes a big difference, they show release from PI0204060801001 2 3 4Not toldToldTrialPercentage correctHow to take a test Avoid PI Answering successive questions on thesame topic hurts recall after answering unrelated questions go back to questions you cannot answer less proactive interference should recall moreT1, T1, T1, T1vsT1, T1, T1, T2ACTConclusions Interference Retroactive interference Proactive interference Release from PI Strong effects Knowing about can help in everydaytasksNext time Other influences on memory Context Encoding specificity What to do if you are drunk while studyingfor an