Judgment and Decision Making (PSYC10) - Lecture Notes

About Dr. Steph

Research interests include:
- Technology attitudes.
- Sacred Values.
- Moral Judgment and Decision-Making.
Teaching:
- C02 - Sci Communication in Psych
- C71 - Social Psych Laboratory
- D15 - The Psychology of Morality
- C10 - Judgment and Decision-Making

Attendance and participation in class discussions are expected (no Web Option).
Punctuality is expected.
No screens (computers, tablets, phones, etc.) during lecture without permission.
Outline slides are posted before lectures.
Full lecture slides are posted after the lectures.
Sit on the side of the room with the screen preference that matches yours!

More applied than most psychology courses at UTSC.
No textbook. All readings are posted on Quercus. Most are written for a lay audience and focus on real-world applications of the research.

Learn how to evaluate and accumulate evidence: how to draw valid inferences from data, and how to identify what kinds of data would best help you make a decision.
Learn how to avoid common decision-making errors and biases.
Learn how to predict and influence the decisions and behavior of others.

IHME's April 8, 2020 Forecast of COVID-19 Deaths in the U.S.
U.S. Coronavirus Rates Are Rising Fast Among Children
Covid may raise the risk of diabetes in children, C.D.C. researchers reported.
Over Two Million Americans Aren't Working Due to Long Covid (Brookings Institution report says the loss of work translates into roughly $170 billion a year in lost wages).

Decisions are forecasts. Launching an album in March vs. July.
Much of decision making is about predicting the future. If I do X, Y will happen.
Let's consider how to make forecasts.

Should we release our album in March or July?
Look for examples of what is done. For example, maybe most albums of this genre are released in the summer.
Rely on salient examples/stories/experience. For example, maybe you know of a similar album that was released in July and did well.
Look for data on consumer demand. For example, maybe our target audience is more likely to buy and listen to music in July than in March.
Look for data on effectiveness of different album release times. For example, maybe most albums that get released in July outperform most albums that get released in March .

Experiments have three defining features:
- Independent variable: The thing you change across conditions.
- Dependent variable: The thing you are are trying to change.
- Random assignment: Randomly assign units to treatments.

Imagine you’ve made two movies – a horror and a comedy – but can only release one of them. We want to release the one that people will like more.
Recruit 2,000 people to watch one of the movies, and then to rate how much they like it (1 = not at all; 9 = extremely).
There are so many individual differences that might affect people’s ratings, like their sex, age, education, income, which other movies they are thinking about when making their ratings, how much they like horrors & comedies , how they use rating scales, etc.

Random assignment helps ensure that the treatment(s) vs. control differ only in the treatment.
All else equal, changing the independent variable in this way changes the dependent variable this much.

Running 5 Minutes A Day Has Long-Lasting Benefits (New York Times, July 30, 2014).
How Much Alcohol Is Safe to Drink? None, Say These Researchers (New York Times, August 27, 2018).
Dating Apps Are Making Marriages Stronger (The Wall Street Journal. August 29, 2019).
Henry Ford: Treatment with Hydroxychloroquine Cut Death Rate Significantly in COVID-19 Patients, Henry Ford Health System Study Shows (July 02, 2020).
A Randomized Trial of Hydroxychloroquine as Postexposure Prophylaxis for Covid-19 (The NEW ENGLAND JOURNAL of MEDICINE).
NIH halts clinical trial of hydroxychloroquine.
Effect of Hydroxychloroquinein Hospitalized Patients with Covid-19 (The NEW ENGLAND JOURNAL of MEDICINE).
Low-Fat Diet Does Not Cut Health Risks, Study Finds (New York Times February 6, 2006).

There are a small number of randomly assigned units.
There is a problem of attrition (people dropping out of the study before being measured).

Even If Random Assignment Works…
Does the observed relationship generalize to the situation I am interested in?
The best way to address this is to conduct multiple experiments under different circumstances, with different materials, participants, etc. If all experiments suggest the same result, you can be more confident that it generalizes.

Random assignment: randomly assign participants to different conditions
- Random assignment allows you to establish causality
Random sampling: randomly sample units to include in your study
- Random sampling allows you to establish generalizability

If the next federal election were held tomorrow, what percentage of the vote would the Liberal Party receive? To know, you’d have to test a random sample of voters.
Are emotional or factual political advertisements more persuasive to voters?
- To really know, you need:
1. Random sample of voters
2. Random sample of emotional ads
3. Random sample of factual ads

If the experiment will be too costly, either in $$ or time.
If the benefit of experimentation is low (either because the variable will likely have no effect or because the potential payoff is small).
If it is physically, legally, or ethically impossible for you to randomly assign units to treatments.