Jeffrey T. Vietri , Gretchen B. Chapman & Janet Schwartz ACTOR vs OBSERVER Flashcards
What are the aims of the paper?
○ Explore why people often have unrealistic expectations about the usage of items
○ Explore how the process + accuracy of our predictions about other’s behaviours are influenced by the information they provide, esp self-predictions
What does it mean by self-predictions are optimistically biased, but informative?
○ When people are making predictions about themselves they are both bias and show good discrimination
§ Bias = eg overestimation of time needed to complete something = reliably optimistic
§ Good discrimination = they are able to finish their work quickly relative to others but not as fast as they predicted
• How has the accuracy of predictions been operationalised?
○ Discrimination accuracy = how strongly our prediction correlates with what we are predicting
○ Bias = whether the mean prediction is systematically higher or lower than what is being predicted
What studies have shown people often have an optimistic bias when predicting their own behaviour?
○ Buehler, Griffen + Ross 1994
§ Undergrads took 50% longer than they had predicted it would take them to complete senior thesis = Planning fallacy
§ Also overly optimistic when asked to explicitly recall their own xp with previous assignments
• Other than academic tasks, where has the optimistic bias been found?
○ Epley + Dunning, 2006 Self-prediction for voting
○ Epley + Dunning, 2000 donation
○ Epley + Dunning 2006, how long a romantic relationship will last
• What evidence is there to suggest people rarely show optimistic bias when predicting for others?
○ Epley + Dunning, 2000 § Found predictions for the avg student in class showed no significant bias
• What can change the predicting process of other people (showing no optimistic bias)?
○ Individuating person whom you are making predictions about
○ Base-rate info are ignored (Epley + Dunning, 200)
- Significantly overestimated how long someone else would take to complete an assignment (Beuhler et al, 1994)
• For observer judgements, what does good discrimination require?
Relevant individuation info
What could be an explanation for the differences in discrimination and bias between actors + observers?
○ Taking two different approaches (Kahneman + Tversky, 1979)
§ Internal - focus on aspect of situation that are specific to the particular case at hand
□ “how much I can cannibalise from my previous work”
§ External - distributional info
□ How long it takes people to write papers (population-base rate)
□ How long has it taken me in the past to complete (personal-base-rate)
• Neglect of what type of info/ approach makes bias more likely (internal vs external)?
Distributional information
• What was the purpose of the longitudinal study?
○ Assess accuracy of a common prediction: How frequently a person will use an item received as gift over the winter holidays
• What was the procedure of the longitudinal study?
○ Actors at the start of fall were asked to make a list of what they would like to receive over the holidays
§ Filled questionnaire asking if what they wanted was a replacement, how often they think they will use it and why they predicted that.
○ Observers asked to predict after a year
§ 2 observer for each actor - one knows the actors prediction + other info
§ Predicted how often the actor used the item + avg college student (subjective base rate) + themselves
• What were the findings of the longitudinal study about the actors?
○ People expect to use a longed-for object more than they actually will
○ Actors are optimistically biased but show discrimination - higher prediction coincide with greater use
What were the findings of the longitudinal study, observers?
○ Both accuracy + prediction process used by observers is dependent on knowing the actor’s prediction for the behaviour
§ Informed observers = unbiased estimates + discriminated among actors = avoiding optimistic bias
□ WHICH MEANS informed observers actually make better predictions than the actors themselves since…no optimistic bias + still show discrimination like the actor
§ Blind observers = bad estimates –> overestimated item use + lacked discrimination
○ Informed observers gave lower estimates of how often they would use the item = in line with how often the actors actually used their items
• What did they conclude from their findings?
○ Access to actor’s estimate determined the strategy used by the observers to produce an estimate
§ Blind observers = similar to actors process
□ They used themselves as a reference (planning fallacy)
□ Blind observers’ estimate for how often they themselves would use the item were inflated vs actor’s reported use
§ Informed observers = use actor’s prediction as starting point then adjust down to correct for the bias present in actor’s predictions
