Prof. Greg Francis 1/2/12 PSY 200: Intro. to Cognitive Psychology 1 Purdue University Two store model PSY 200 Greg Francis Lecture 15 Why it is difficult to win a pizza at Little Caesar’s. Purdue University Memory  Simple view  memory is a container of past impressions and knowledge  memories can leak-out, decay away  Not very realistic  need to explain why memories disappear Purdue University Simple view  The container theory of memory does not explain, for example,  why some memories are very long lasting (my childhood car trips to Utah)  why some memories are very brief (my wife asks me to take out the trash)  We are not going to get a full theory of memory, but we can start to get an outline  and identify some misconceptions about memory Purdue University Ebbinghaus’ experiments  First memory experiment (1885)  Measure how long it takes to learn a list of nonsense syllables perfectly  NOF, QAP, HOS, LEQ, FIK, MEC, KIJ, HOM, NEM, MOJ  How long does the memory last?  In what form does the memory last?  How does it affect future behavior?  Does it help relearn the list at a later time? Purdue University Ebbinghaus  Relearn the list at later points in time  a different list each time  Measure how long it takes to relearn the list  Calculate savings Savings = Time TimeTimeoriginal relearnoriginal−Purdue University Forgetting curve  Savings =1  subjects do not need to relearn, perfect memory  Savings=0  subjects show no evidence of earlier learning 00.20.40.60.811.20 5 10 15 20 25 30Time (days) Savings Still not 0!Prof. Greg Francis 1/2/12 PSY 200: Intro. to Cognitive Psychology 2 Purdue University Significance  Ebbinghaus’ results suggest that memories can last a very long time, in some form  Memories were believed to be “stored” in a memory system and did not just fade away (otherwise, the curve should not asymptote above zero)  Memory loss was believed to be due to interference of other memories  Other experiments challenge this view Purdue University Memory task  See (or hear) a trigram of consonants  Report it back in order  Ebbinghaus’ results suggest good memory until other letters are also memorized WRM Purdue University Retention  Peterson & Peterson (1959)  Brown (1958)  Give subjects trigram  ask them to count backwards by 3’s and then recall trigram WRM 782 779, 776, 773,... Purdue University Retention  Vary duration of counting backward  Numbers are different from letters, you might not expect any interference  but they can have very strong interference 00.20.40.60.810 3 6 9 12 15 18Retention interval (seconds) Proportion correct Suggests some memories last only a few seconds! Purdue University Retention  The results of the Brown-Peterson study suggest that some aspects of forgetting are process driven  keeping a memory “active” requires effort  if you are distracted by another task, you cannot apply the effort to keep the memory  similar to our observations about attention and processing Purdue University Retention  The results of the Brown-Peterson study also suggest that some aspects of forgetting are passive  even if you are distracted, you can recall the trigram if only a short time has passed  if many seconds have passed, while you are distracted, you cannot recall the trigram  memory has “decayed”, or something like decay, while you were doing the distracting taskProf. Greg Francis 1/2/12 PSY 200: Intro. to Cognitive Psychology 3 Purdue University Retention  CogLab data  193 subjects  Not everyone shows interference on this task Purdue University Another experiment  Memory span  how many items can you correctly recall immediately after exposure?  “The magic number 7+/-2:…”  Miller (1956) 9 5 7 8 6 3 4 9 2 7 5 9 0 1 3 6 8 2 9 3 4 7 2 5 1 6 9 3 9 1 0 3 5 7 3 2 3 9 0 5 7 4 2 1 6 Purdue University Interpretation  There exist two types of memory systems  Long Term Memory (LTM)  high capacity (no limit)  long duration (forever)  Ebbinghaus’ experiment  Short Term Memory (STM)  small capacity (~7 items)  short duration (seconds)  Memory span, Brown-Peterson Purdue University Pizza  There used to be a Little Caesar’s in W. Lafayette with a game where you could win a pizza  must repeat a sequence of flashing lights  The sequence gets longer until you make a mistake  need a sequence length >7 to win much » nearly impossible with STM properties  http://www.freegames.ws/games/kidsgames/simon/simon.htm Purdue University Modal Model of Memory  Atkinson & Shiffrin (1968)  Multiple stages of memory  STM plays a dominant role in active memory  Requires transfer between STM (STS) and LTM (LTS) Purdue University Modal Model of Memory  When something is memorized  Items are first held in STM (temporary store)  Items may transfer to LTM (permanent store)  Takes time to transferProf. Greg Francis 1/2/12 PSY 200: Intro. to Cognitive Psychology 4 Purdue University Free Recall Serial Position Curve  Given almost any list of items  Subjects remember the first and last few items best (free recall, not immediate serial recall) 00.20.40.60.811 2 3 4 5 6 7Position in list Probability of recall PRIMACY RECENCY Purdue University Serial position curve  The effect of position is robust across many types of lists  words  letters  numbers  pictures…  Here’s the CogLab data  (196 subjects)  Demo Position in list"Percent correct"Purdue University Serial position curve  In some situations the serial position curve can be explained by different properties of STM and LTM 00.20.40.60.811 2 3 4 5 6 7Position in list Probability of recall PRIMACY: Use LTM RECENCY: Use STM NEITHER LTM NOR STM Purdue University Conclusions  Short Term Memory (STM)  Long Term Memory (LTM)  STM / LTM distinction is one of the strongest conclusions of cognitive psychology  Accounts for quite a bit of data  Many details are unresolved Purdue University Next time  Expansion of STM into  Working