The DRM paradigm is used to study false memory. Lenton, Blair, and Hastie (2001) adapted the DRM to study false memories produced by social stereotypes. Lenton et. al. (2001) define a stereotype as the "cognitive representation of ideas, facts, and images associated with a social group." Studies have centered on direct associations, in which there is no reference to another concept (indirect association), as defined by free associations. The dyad meat and potatoes is an example of a direct association. Lenton et al question whether false memories can be produced using indirect stereotypes in the DRM paradigm. False memories have been explained by spreading activation and fuzzy trace theory. Study results have been mixed, as to the "cause" of false memories. One study has shown the false memory effect when study procedure (subliminal presentation of the target) precluded the use of a gist based strategy. The authors speculate that the use of indirect associations provides another avenue for studying false memory when subjects are not able to discern the gist of the items presented. Finally, previous research indicates that stereotype consistent information is remembered more accurately than stereotype inconsistent or neutral information.

Hypothesize:

• More false alarms for stereotypic lures than for unrelated lures.
• Subjects would not be aware of list theme for the indirectly associated list.
• Exposure to stereotypic roles would also increase false alarm rate for roles and traits for that stereotype.

Method:

• 90 subjects (68 female/22 male) undergraduate psychology students participated in class
• Exposure phase: study a list of 75 words for 1 sec per word
• 5 subsets of the list, which are 15 words long
• 4 direct lists taken from Roediger and McDermott (chair, fruit, window, sleep)
• 1 indirect list (either associated with male or female stereotypes), which was the third list presented

• Recognition phase: following a 3 minute distraction task, took a 46-item

• 10 studied (2 per list) and
• 36 non-studied words (4 critical lures for each direct list, 3 roles and 3 traits for each gender stereotype, 20 fillers
• response categories: Confident New, Probably New, Probably Old, Confident Old

• Theme Awareness phase: Asked if noted any themes among words and to list

• Hit Rates: .79 direct list and .88 stereotype lists
• False Alarm Rates: .79 direct list
• .30 role lure and .07 for trait lure
• more likely for stereotype consistent than inconsistent or neutral for roles
• more likely for role lures for the list seen
• more likely for trait lures when the "female" list is seen (not true for men – provides only partial support for spreading activation
• more likely for direct list (.79) than for indirect lists (.43)
• Theme Awareness: direct list themes were identified by participants 76% of the

• indirect list themes were identified 22% of the time, with most subjects identifying occupation as the theme
• indirect theme aware subjects were no more likely to make false alarms than their unaware counterparts

Further examine awareness of gist of indirect list through more direct questioning

Examine possibility that using explicit memory to exclude traits from "new" answers

Method:

• Differences: 100 subjects – 71% female, 79% white
• Removed "virgin" from female role list (WHAT??) and replaced with therapist
• Added list of 15 gender neutral traits (to test if used explicit memory to exclude

• Asked very specific questions about when subjects became aware of list themes and to what extent this information was used during the recognition test
• Implicit gender association task added – prime-target trials comparing response times between stereotype consistent and stereotype inconsistent trials
• Explicit gender association task added – free association task, measure degree to which responses to stereotypic roles and traits
• Sexism measures: Modern Sexism Scale and subscales of Ambivalent Sexism Inventory; were added
• Completed new tests and measures after completing modified version of experiment 1

Hit Rates: .86 for direct list, .84 for indirect list, and .78 for neutral list

False Alarms: .79 of the direct lures

• Significantly more likely for stereotypically consistent roles AND traits than for inconsistent roles and traits (both activated by stereotypic roles)
• Subjects viewing the male list were more likely to false alarm to the male stereotypes than to female stereotypes
• Subjects viewing the female list were no more likely to false alarm for female or male stereotypes
• More likely to false alarm to direct than indirect lures

• Theme Awareness: direct list themes were identified by participants 68% of the
• time, with 17% of subjects identifying all four
• trait list themes were identified 56% of the time
• indirect list themes were identified 8% of the time, when identified subjects predominantly indicated that this information was not used in the recognition test
• Implicit Gender Association Task: subjects responded faster to gender consistent than inconsistent trials, but poor predictor of stereotype memory errors
• Explicit Associations Task: of the 3 association types only secondary associations were found to predict stereotype memory errors, but not in the expected direction. As the number of secondary associations increases, subjects produce fewer stereotype false alarms
• Sexism Measures: neither predicted stereotype false alarms, but they did indicate that the sample was moderately racist

Indirect associations can produce false memories in the DRM. Exposure to stereotypic roles in this manner also led to increased false memories in other indirectly associated conceptual categories (traits).