Long Term Memory

Endel Tulving's Distinctions in LTM

Episodic Memory
        memory for a particular time or event
today we will consider:
        memory for a particular list of words (one time/episode)
Ebbinghaus's Pioneering Work
Modern studies follow a tradition begun by Ebbinghaus, who studied memory and forgetting using objective methods at
the time of Wundt, but was outside the field and working largely on his own. 

Ebbinghaus studied his memory for a list of "nonsense syllables" CVC triads like DAR, BEL, FOT, MUK, LIM

He would learn a list until he knew it perfectly and count the number of times through the list that it took for him to
master it.

Then after a delay he would try to remember the list in order. If he forgot the list he would go through it until he
remembered the whole list perfectly again. The number of times it took him the second time would be less than the
number of times he had to go through the list the first time. This was his "Savings". Ebbinghaus actually plotted savings
as a percentage -- so if it took him 6 trials the first time but only 3 the second time -- his savings was 50%. If he
remembered the whole list the second time without any new practice, his savings was 100%.

Thus, Ebbinghaus had an objective DEPENDENT measure -- a measure that depended on his performance as a subject --
and it was the percent of savings after a delay. 

He also had an INDEPENDENT variable -- a variable that he manipulated as the experimenter -- and that was the amount
of delay between the first and second attempts.

Ebbinghaus looked at savings across a wide range of times and what he found was this:

                         delay          savings   
                         immediate      100%
                         20 minutes     60%
                         1 hr           45%
                         9 hr           35%
                         1 day          30%
                         2 days         25%
                         6 days         22%
                         30 days        20%

to graph this, we put the DEPENDENT variable (the one that "depends" on the subjects behavior or the measure of
performance that we collect in the experiment) on the Y-AXIS

and the INDEPENDENT variable (the one that the experimenter manipulates or the one that defines the groups or subjects
or stimuli that we want to compare) on the X-axis

so we get a graph like this

which is called the "forgetting curve"

Ebbinghaus discovered an important principle in his research -- that memory decays as a function of time, with the most
dramatic forgetting happening soon after the original learning -- but perhaps more importantly, Ebbinghaus provided
modern cognitive psychology with a model for the empirical investigation of thought.
What gives us teh best memory for episodes?
When do we remember lists better?

1. Elaboration (Depth of Processing, Craik & Lockhart)
     Surface, Rhyme, Semantic tasks   best recall for semantic task
2. Transfer Appropriate Processing
     Effectiveness of processing DEPENDS ON TEST
     (Morris, Bransford & Franks, 1977)
3. Encoding Specificity (Tulving & Thompson)
People will be able to recall the most under conditions that match the
way information was encoded
State dependent learning
     Context of the PERSON also plays a role! 

5. Organization 
Organized lists are easier to learn and remember
6. Distinctiveness            von Restorff effect

7. Imagery (Paivio)

Dual coding hypothesis
Words capable of evoking a vivid and stable image will be easier
to remember

imagining words gives us TWO possible ways of encoding
information -- which in turn means two sources of retrieval cues