Deck 8: Section 2: Accuracy and Efficiency in Social Interference

Question 1

Briefly describe the "law of large numbers." In what ways do people understand, or misunderstand, its implications? How does this relate to sample selection? Briefly discuss some of the implications for psychological research.
a. Small sample sizes can have high variability, and so are not good estimates of population characteristics compared to larger samples).
b. People tend to believe that increasing the size of the predictor sample of events increased prediction power.
c. Without random sampling, the sample population can be biased by selection factors, and the resulting sample population would not be highly representative of the overall population.
d. For research, it is important to pay attention to potential biases in sample selection, otherwise generalizability of the findings is compromised.

Answer 1

Not Answer

Question 2

In a controversial 1977 paper, Nisbett and Wilson made a certain claim about cognitive processes. Describe the claim and the empirical evidence. Do you agree with the authors? Describe why or why not in 2-3 sentences.
a. Claim: people have little or no introspective access to their cognitive processes.
b. Studies: In "consumer preferences" study, participants show serial position effect right-most item more favorable), but explain preference as due to qualities of the garment itself.

Answer 2

Not Answer

Question 3

Are normative models of inference always optimal? Describe a situation in which a normative model and an intuitive model might reach different inferences. Briefly discuss the consequences.
a. Example from book: normative model would suggest not holding a job candidate's nervous gaffes against him, since the behavior is situationally constrained and only a small, momentary sample; however, intuitive model might take into account the need for this person to be composed under pressure, and so the normative model might not be the most helpful for this hiring decision.
b. Many judgments lack objective referents, and so it is not always possible to determine the normatively "best" outcome.

Answer 3

Not Answer

Question 4

Should we allow computers to make social inferences for us? Why or why not or, under which circumstances, and why)?

Question 5

Is it useful to compare how people make social inferences to a normative model of how people should form judgments? If not, then what might be a better approach, and why?

Question 6

Name and describe three reasons why social inference often fails to match normative models.
a. Meeting alternative goals.
b. The social perceiver operates not only under accuracy pressures but also under efficiency pressures, making inferences quickly in a rapidly changing environment with multiple demands.
c. Capacity limitations on short-term memory.