ShortTerm Memory Laboratory

1
Short-Term Memory Laboratory

• Objectives
• Introduce basic theoretical empirical issues
• Review predictions
• Present methodological basics

• Tools
• Todays presentation
• Next weeks planning session
• References
• Cognitive Psychology lectures

2
ISR and the Articulatory Loop Hypothesis.

• There are severe limits on the number of
  unrelated verbal items that can be immediately
  recalled in their order of presentation.
• Typical paradigm used to study STM Immediate
  Serial Recall or ISR
• Research shows that in most conditions we can
  recall approximately 7 or 2 items or about
  2 seconds worth of material

3
Data Reliability

• Multiple trials per condition
• A trial one memory list to remember
• Multiple participants
• Standard conditions of testing standard
  instructions.

4
Verbal Working Memory LTM...

• What is the relationship of verbal short-term
  memory to LTM ?
• For example, it is well established that lexical
  or high-frequency items lead to superior
  performance in ISR...
• Word frequency / word familiarity
• Lexicality
• Semantic category
• Concreteness

5
Lexicality Frequency effects

• Tehan Humphreys (1988)
• Frequency, word class, suppression
• Gregg, Freedman, Smith (1989)
• Frequency suppression.
• Poirier Saint-Aubin (1996)
• Frequency phonemic similarity

6
LTM contribution ...

• A two-stage account of ISR performance where
• A list sets up a phonological representation in
  working memory that begins to degrade rapidly.
• At recall, degraded representations are
  reconstructed or interpreted by calling upon
  long-term verbal knowledge
• Highly familiar items are more accessible

7
A finer grained hypothesis...

• This leads to the prediction that high
  familiarity, lexical, etc. items will lead to
  better item information recall
• All else being equal, such items will have a
  higher probability of being recalled based on a
  degraded representation.
• Predictions of the redintegration hypothesis when
  high low familiarity items are mixed are not
  supported.

8
Alternative Item Order hypothesis

• In ISR, performance is a function of a trade-off
  between encoding item specific information and
  relational information.
• Low frequency / familiarity items promote item
  information encoding at the expense of order
  information encoding.

• ISR calls upon order encoding for good
  performance - Hence recall is better for highly
  familiar items.
• Recognition better for low-frequency items
• MIXED LISTS ISR
• Order recall more equated
• Differences should disappear or reverse

9
Tests of the item-order hypothesis

• Paradigm 1
• Frequency x list type
• High-low freq
• Pure lists vs. mixed lists.
• Immediate serial recall

• Paradigm 2
• Frequency x list type
• High-low freq
• Pure lists vs mixed lists.
• Missing item task

10
Design Materials P1

• 4 conditions High in pure lists, Low in pure
  lists, High in mixed lists, Low in mixed lists.
• 4 blocks of trials 1 high pure, 1 low pure, 2
  mixed.
• 6 or 7 items per list

11
Design Materials

• 10 to 18 lists per condition
• lists have to be recalled (ISR) or one item
  recalled in missing item task
• Controls
• Word-length, concreteness, imageability, phonemic
  similarity, semantic category

12
Item Order scoring

• pontoon lagoon racoon spoon platoon
  dragoon
• pontoon lagoon spoon XX dragoon platoon
• 1 1 0 0 0 0
• 1 1 1 0 1 1
• Lenient scoring 5, it follow that Item errors
  1
• Strict scoring 2
• Order errors (5-2) 3, but often proportions
  important

13
Research Methods - Lab 1

• The devil is in the details

14
Definition

• To clarify - a trial includes the presentation of
  a to-be-remembered list and the response to that
  list.
• Each condition in your experiment should include
  many trials -- at least 8 to 10.

15
Steps to accomplish

• Selection of the stimuli for inclusion in your
  study.
• How many words do you need?
• What characteristics should they have?
• Word length to avoid floor ceiling effects
• Difference in frequency introduced
• Pre-testing a good idea
• MRC database
• Excel good tool also

16
Steps

• Construction of the lists for each condition
• Probably preferable to use the same words in the
  pure mixed conditions
• Counterbalancing of conditions?
• Preparation of instructions, practice trials, and
  experimental trials.

17
Stimulus presentation

• Most years, stimuli are presented through
  PowerPoint
• Timing capacities of the software can be called
  upon
• Instructions response sheets needed

18
Data scoring
19
Data Input
20
Serial position analyses possible