Cognitive (Reliability of Cognitive Processes) Studies

Question 1

Martin and Halvorson

Answer 1

Aim:
- to determine if pre-existing gender schemas will distort memories that are inconsistent with these schemas

Procedure:
- 24 girls and 24 boys aged 5-6
- shown 16 pictures of a person doing an activity (half gender congruent and half gender incongruent)
- children divided based on their knowledge of gender stereotypes
- each picture shown for about 10 minutes and then children were asked to indicate their confidence of their memories
- a week later they were assessed again and asked which gender was completing the specific activity

Results and Conclusions:
- significantly more reversals of gender in the schema-inconsistent images
- kids had the same confidence in gender swapped images as the accurate answers
- pre-existing stereotypes distorted the children’s memories of the gender incongruent images
- reversed gender to confirm their stereotypes

Strenghts and Limitations:
- extremely controlled
- replicable and reliable
- sample bias means no generalizability outside of American culture

Question 2

Tversky and Kahneman

Answer 2

Aim:
- explore the effects of anchoring bias in a mathematical estimation

Procedure:
- high school students were assigned to two conditions: ascending and descending
- asked to estimate the value of (1x2x3…8) or (8x7x6…1) respective to condition
- since we read from left to right, hypothesis was that the ascending group would use “1” as an anchor and predict a lower value than those that saw “8”
- first number seen would bias the estimate of the value by the participant

Results and Conclusions:
- the median for ascending group was 512
- median for descending group was 2250
- actual value: 40320
- those in the ascending condition estimated a significantly lower number than those in the descending condition, suggest anchoring bias
- quick estimate relied on System 1 thinking, using heuristic (mental short-cut) and estimating very far off from actual value

Strengths and Limitations:
- easily replicated to establish reliability
- high internal validity
- artificial, lacks ecological validity
- independent measures, participant variability may have impacted results