One of the most revealing results from psychology is that we are intuitively horrible statisticians. In this article, let's explore how poorly we calculate likelihoods of events and why we could derive some benefit from knowing how probability works.
1) You go to see a doctor and she runs a test to see if you have cancer. If you have cancer, there is a 95% chance that your result will be positive. If you don't, there is only a 2% chance that your result will be positive. You are shocked to learn that the result was positive. How worried should you be?
If you're like most people, you would be terrified because you'd think that the odds of you having cancer is about 95%. This is a classic example of ignoring the base rate which in this case is the percentage of people in the U.S. who get cancer each year. It turns out that percentage is roughly 0.5%. Given this information, you can calculate how likely it is for you to get a positive test result whether you do or don't have cancer. Despite getting the positive result, you have a probability less than 19.3% of actually having cancer (after applying Bayes' Theorem from elementary probability theory).
2) Here's a simpler one that is deceptively tricky. There's a couple who has two children and you are told that one of them is a girl. What are the odds that the couple has two daughters? Again, most people think 50% since the other child has a 50/50 chance of being male or female.
Let's go through this problem step by step. Note that there are four possibilities:
OLDER CHILD YOUNGER CHILD
Girl Girl
Girl Boy
Boy Girl
Boy Boy
You know that they have at least one girl so you must get rid of the last possibility; all the others are fair game and they all have equal likelihoods. Thus, the probability of having two daughters is actually only one-third, not one-half.
3) Ashley majored in English literature at a small liberal arts school where she was active in theater, music, and political activism concerning women's and minorities' rights in America. It has been over a year since she graduated from college. Given five possible post-graduation outcomes, place a probability next to each possibility:
A) She is pursuing her dream to become an actress.
B) She is working as a banker.
C) She is working as a banker and auditioning for acting roles at night.
D) She is attending law school.
E) She is unemployed.
Given what you know about Ashley, take the time to actually place a probability for each outcome.
Okay, done? Great! First, make sure that your numbers don't add up to over 100% since that wouldn't make any sense. Second, what probability did you assign to B? How does that compare to the probability you assigned to C? Many people would assign C with a higher probability than to B since Ashley seems like she's very artistically inclined. However, this is actually impossible. The probability of her working as a banker AND auditioning for acting roles must be lower than the probability of her working as a banker. One way to think about it is that the probability of her working as a banker is equal to the probability of her working as a banker and auditioning for roles PLUS the probability of her working as a banker and NOT auditioning for roles. Thus, the probability of C cannot be greater than the probability of B.
These types of puzzles unveil some of the imperfections of our "rational" minds. If you are interested in exploring the faulty statistics that go on in our heads, I highly recommend Daniel Kahneman's Thinking: Fast and Slow and Leonard Mladinow's The Drunkard's Walk: How Randomness Rules Our Lives. I may also post more about these issues in the future.