Two main ways we access memory: recognition and recall.
- when provided with the information in memory,
- just double check that we’ve seen it before.
- includes: feeling of familiarity, matching, multiple. choice, True/False on exams, recognizing someone you know…
- coming up with the information from memory yourself.
- Includes: coming up with the name for person you recognized, recalling where you were when Kennedy assassinated, fill-in-the-blank on exams.
Recognition involves a process of comparison of info with memory. Recall involves a search of memory and then the comparison process once something is found.
Not just different amounts of the same memory – not just a diff. in strength. Qualit. different. Example: rare words are almost always better recognized than common ones, but usually less well recalled.
Recognition has 2 mechanisms: familiarity and retrieval. We can recognize. an uncommon item just because it seems familiar. For a more common item, will need to retrieve additional information, such as where the item was encountered.
Usually, recognition is better than recall.
Some basic facts about long term memory
Sometimes we fail to retrieve info that we know we have – retrieval failure
One common type of retrieval failure is the Tip-of-the-tongue phenomena – you know you know the meaning of a word, or the answer to a question, but you can’t retrieve it.
But sometimes, we simply don’t encode well. Even a highly familiar item may not have a detailed representation in memory.
Accessing one category can lead to retrieval of others that are linked to it (remembering in clusters; spreading activation from one concept to another, as in Collins & Quillian’s model)
Other times, activating one category member that is active can inhibit retrieval of another – Brown’s (1968) state name experiment. (Fixating on an already active concept….)
Schema consistent vs. Schema inconsistent knowledge
Schemas – generalized knowledge about situations and events. Ex: schema about going out to eat includes that you often wait to be seated; you get shown a menu, and order choices from that; you have a waiter or waitress that takes your order; that things get served in a particular order; that you get a bill at the end; etc. etc.
You can also have schemas about places, such as what is in a typical living room. Generally, we’re likely to later remember things that fit with the schema, but not things that don’t (unless they are so unusual that you made special note of it.)
Ex: experimenter’s office study.
Rock & Engelstein (1959) showed Ss a single meaningless shape – capacity to draw it (recall) declined rapidly, but could recog. them almost perfectly a month later.
Nickerson (1965) showed Ss 600 pics of various scenes and events, and tested retention at intervals from a day to a year later (in a new vs old categorization task). Performance after 1 day [=] 92%, after one year [=] 63%.
Standing, Conezio & Haber (1970) presented 2560 color slides, 10 sec. each; Ss then showed pairs and had to say which was old. Perf. still at 90% a few days later.
Goldstein and Chance (1971) used women’s faces, magnified snowflakes, and inkblots; 14 stim. of each set, 3 sec. each; recall tested immediately and 2 days later by making a new/old decision task mixing the 14 with 70 new ones (so chance = 14%). Perf. at 71% faces, 48% inkblots, 33% snowflakes, both immediatly and 2 days later.
So – when we see pictures of complex scenes, where there are many diff. aspects of the pic, any of which could cause the feeling of familiarity, we can literally remember thousands of pics for days, and lesser numbers for over a year. When the pics have fewer diff. things that can be remembered about them, or are more confusable, performance deteriorates, but is still quite good.
Why do we forget???
- fading (trace decay) over time
- interference (overlaying new information over the old)
- lack of retreival cues.
Evidence for lack of retrieval cues:
- encoding specificity
- state dependent learning
- with more time, context is more likely to chance, so you’ll find it harder to access information in memory on the basis of that cue. Also, as more items are learned in the same context, context becomes less efficient cue (release from PI)