Karteikarten
Recognition Memory Part 1
Karten in diesem Stapel (29)
- 1The ability to remember whether or not a presented item has occurred in a particular contextWhat is recognition?
- 2Testing Recogniton in the Lab: Stimuli are usually items subjects have prior experience with:e.g., words, letters, pictures of familiar objects or scenes
- 3Stimuli can also be novel (things that subjects have not experienced prior to the experiment):e.g., fractals, novel objects or scenes
- 4Subject is presented with a list of test (a.k.a. probe) items Targets (“old” items) Items that were studied in the tested context Lures (“new” items) Items that were not studied in the tested contextItem Recognition Test
- 5Test items are presented one at a time Subject responds “old” or “new” (“yes” or “no”) to each one*Old-New (Yes-No) recognition:
- 6Subject must choose which one of two (or more) items were studied*Forced choice recognition
- 7*Subject may also be asked to rate confidence in his/her responseItem Recognition Test
- 8Recognition MeasuresAccuracy/Reaction time/Response time/RT
Response: old/yes- (target-Hit)
New/No- (target/miss)
Old/yes- (Lure/False Alarm)
New/No (Lure/Correct Rejection). - 9Does a High Hit Rate Imply Good Recognition??
- 10Response bias towards "yes" and Response bias towards "no"?
- 11Memory strength of the test probe is assessed Decision is made directly from strength (or from a transformation of strength)Single process theories
- 12Decisions can be made using strength or using a recall-like processDual process theories
- 13Every item in memory has a strength We can assign a numerical value to that strength Items vary in strength When an item is experienced (studied), it’s strength increases A memory decision can be made by comparing the strength of a probe to some criterionStrength Theory: Assumptions
- 14Strengths are randomly distributed in a normal (Gaussian) distribution Studying an item increases its strength At test, the studied items have a higher mean strength than the unstudied items Decision is made by comparing strength of probe to a criterion (threshold)Signal Detection Theory (SDT)
- 15Signal Detection Theory (SDT)Strengths are randomly distributed in a normal distribution with mean μ (mu) and standard deviation σ (sigma)
- 16Signal Detection Theory (SDT)Studying an item increases its strength
- 17—Subjects select a threshold (criterion)
Items whose strength is above the threshold are identified as “old”
Items whose strength is below the threshold are identified as “new” - 18—SDT: Decision rule
- 19SDT: BiasDetermined by where the subject sets the threshold
Symbol: C or β - 20—SDT: Bias
- 21SDT: BiasLow threshold
Few misses (lots of hits)
Lots of false alarms
High threshold
Low chance for false alarms
Lots of misses - 22SDT: SensitivityThe ability to distinguish between targets and lures
Quantified by the degree of overlap between the distributions - 23SDT: SensitivityStandard measure is d’
- 24SDT: Sensitivity
- 25SDT: Confidence RatingsConfidence ratings can be made by setting multiple thresholds
- 26—The probability of making a response at any level of confidence is the area under the curve for that confidence level
- 27—Receiver operating characteristic (ROC) curve shows the relation between the hit rate and the false alarm rate as the criterion varies.
- 28SDT: ROC curvesPlots the cumulative probabilities for hits vs. false alarms for each criterion (confidence level)
- 29SDT: ROC curves
Informationen
- Beschreibung
- Karteikarten am Recognition Memory Part 1, erstellt von sarrowsmith.5 am 29/10/2013.
- Erstellt von
- sarrowsmith.5
- Aufrufe
- 24
- Bewertung