Cab Colours and Cognitive Biases

In the bustling metropolises of the UK, the sight of a black cab is as iconic as Big Ben. However, the world of taxis isn't always so monochromatic. Imagine a scenario where two distinct taxi companies, one predominantly green and the other blue, operate within a city. This seemingly simple setup becomes the cornerstone of a famous psychological experiment that delves into the very nature of our decision-making processes, particularly when faced with probabilities and incomplete information. This isn't just about taxi colours; it's a window into how our brains process risk and likelihood, often with surprising inaccuracies.

The Famous Kahneman and Tversky Experiment

Back in 1981, psychologists Daniel Kahneman and Amos Tversky, renowned for their groundbreaking work on behavioural economics and cognitive biases, published a study that illuminated a common pitfall in human reasoning. Their experiment, involving a hypothetical traffic accident, revealed a significant tendency for people to overlook 'base rates' – the fundamental, prior probabilities of events – when making quick judgments. This oversight can lead to dramatically incorrect conclusions, a phenomenon vividly illustrated by the case of the green and blue cabs.

The Scenario: A Night of Uncertainty

The experiment begins with a narrative: a taxi is involved in a hit-and-run accident one dark night. The city's taxis are operated by two companies: Green and Blue. Crucially, we are given the following statistical information about the taxi fleet:

• 85% of the cabs in the city are Green.
• 15% of the cabs in the city are Blue.

A witness comes forward and identifies the offending cab as Blue. To gauge the reliability of this testimony, the court conducted tests under the specific conditions of the accident night. These tests revealed that the witness correctly identifies the colour of a cab 80% of the time. Conversely, the witness misidentifies the colour 20% of the time.

The Intuitive Trap: What's Your First Guess?

Now, consider yourself in the position of someone asked to assess the likelihood that the cab involved in the accident was indeed Blue. Given the witness's statement and the known accuracy of their identification, what percentage do you intuitively assign to the cab being Blue? Most people, when presented with this scenario without time for detailed calculation, tend to gravitate towards the witness's accuracy. An 80% accuracy rate in identifying the cab as Blue often leads to an intuitive conclusion that there's an 80% chance the cab was Blue.

However, as Kahneman and Tversky demonstrated, this intuitive leap is flawed. The majority of people, in their haste, neglect to consider the 'base rate' – the initial distribution of taxi colours. This is akin to being shown a person wearing a disguise and asked whether they are more likely to be a spy or a librarian. While the disguise might suggest espionage, the sheer abundance of librarians in the population dramatically lowers the probability of them being a spy, regardless of the disguise's effectiveness.

The Correct Answer: A Probabilistic Revelation

The actual probability, when calculated correctly, reveals that the cab involved in the accident was Blue only 41% of the time. This stark difference between the intuitive 80% and the calculated 41% highlights the power of cognitive biases, specifically the neglect of base rates.

Unpacking the Mathematics: The Power of Logic Over Intuition

Arriving at the correct answer without the aid of calculation tools can be challenging. It requires a solid grasp of probability theory or a keen logical faculty. The key to understanding this result lies in incorporating the initial proportion of taxis into the assessment. The fact that 85% of cabs are Green significantly reduces the weight we should give to a single 'Blue' observation, even from a seemingly reliable witness.

For those familiar with probability, this is a classic application of Bayes' Rule. While formulas can be efficient, they can sometimes obscure the underlying logic. To make the reasoning more transparent and easier to follow, we can utilise a tree diagram to visualise the conditional probabilities involved.

Building the Tree Diagram

Let's break down the possibilities step-by-step:

1. Initial State: Taxi Colour

• Probability of the cab being Green (G): P(G) = 0.85
• Probability of the cab being Blue (B): P(B) = 0.15

2. Witness Observation: Correct or Incorrect

We know the witness's accuracy:

• Probability of correctly identifying Blue as Blue: P(Witness says Blue | Cab is Blue) = 0.80
• Probability of incorrectly identifying Blue as Green: P(Witness says Green | Cab is Blue) = 0.20
• Probability of correctly identifying Green as Green: P(Witness says Green | Cab is Green) = 0.80
• Probability of incorrectly identifying Green as Blue: P(Witness says Blue | Cab is Green) = 0.20

3. Combining Probabilities (Multiplying along branches):

Now, let's calculate the probability of each combined event:

• Cab is Green AND Witness says Green: P(G and Witness says Green) = P(G) * P(Witness says Green | Cab is Green) = 0.85 * 0.80 = 0.68
• Cab is Green AND Witness says Blue: P(G and Witness says Blue) = P(G) * P(Witness says Blue | Cab is Green) = 0.85 * 0.20 = 0.17
• Cab is Blue AND Witness says Green: P(B and Witness says Green) = P(B) * P(Witness says Green | Cab is Blue) = 0.15 * 0.20 = 0.03
• Cab is Blue AND Witness says Blue: P(B and Witness says Blue) = P(B) * P(Witness says Blue | Cab is Blue) = 0.15 * 0.80 = 0.12

To verify our calculations, the sum of these probabilities should equal 1:

0.68 + 0.17 + 0.03 + 0.12 = 1.00

This confirms that we have accounted for all possible outcomes.

Answering the Crucial Question

The question is: Given that the witness said the cab was Blue, what is the probability that the cab was actually Blue?

We are interested in the cases where the witness identified the cab as Blue. These are the two scenarios where the witness said 'Blue':

• Cab is Green AND Witness says Blue (Probability = 0.17)
• Cab is Blue AND Witness says Blue (Probability = 0.12)

The total probability of the witness saying 'Blue' is the sum of these two probabilities: 0.17 + 0.12 = 0.29.

Now, we want to find the proportion of these 'Blue' identifications that were actually due to a Blue cab. This is calculated as:

P(Cab is Blue | Witness says Blue) = P(Cab is Blue AND Witness says Blue) / P(Witness says Blue)

P(Cab is Blue | Witness says Blue) = 0.12 / 0.29

Calculating this gives us approximately 0.41379, which rounds to 41%. This is the correct probability.

The Influence of Base Rates: A Deeper Dive

The dramatic difference between the intuitive 80% and the calculated 41% is a powerful demonstration of the 'base rate fallacy'. Our brains, when faced with new, specific information (the witness's testimony), tend to overweight it and underweight the pre-existing, general information (the proportion of green and blue cabs). The 85% of green cabs acts as a significant dampener on the likelihood of the cab being blue, even when the witness claims it is.

Exploring Other Scenarios

The beauty of this probabilistic framework is its versatility. We can use the same data and logic to answer other questions. For instance, what if the witness had said they saw a Green cab?

In this case, we would look at the scenarios where the witness said 'Green':

• Cab is Green AND Witness says Green (Probability = 0.68)
• Cab is Blue AND Witness says Green (Probability = 0.03)

The total probability of the witness saying 'Green' is 0.68 + 0.03 = 0.71.

The probability that the cab was actually Blue, given the witness said Green, would be:

P(Cab is Blue | Witness says Green) = P(Cab is Blue AND Witness says Green) / P(Witness says Green)

P(Cab is Blue | Witness says Green) = 0.03 / 0.71

This results in a probability of approximately 0.04225, or about 4%. This illustrates how the base rate heavily influences outcomes, making a 'Blue' identification highly improbable if the witness claims to have seen 'Green', given the overwhelming prevalence of green cabs.

Why This Matters in Everyday Life

The lesson from the green and blue cabs extends far beyond probability puzzles. It influences our judgments in medical diagnoses (overweighting symptoms, underweighting disease prevalence), legal proceedings (focusing on specific evidence while ignoring statistical likelihoods), and even personal relationships (judging someone based on a single action rather than their general character). Understanding and consciously applying base rates can lead to more rational and accurate decision-making.

Frequently Asked Questions

Q1: What is a 'base rate' in this context?

A base rate refers to the initial, prior probability of an event or characteristic before any new evidence is considered. In the taxi experiment, the base rate is the overall percentage of green and blue cabs in the city (85% green, 15% blue).

Q2: Why do people often get this experiment wrong?

People tend to fall prey to the 'base rate fallacy' or 'neglect of probability' by focusing too heavily on specific, vivid information (like the witness's testimony) and ignoring the more general, statistical information (the actual proportion of cabs).

Q3: Is the witness's testimony irrelevant?

No, the witness's testimony is crucial, but its impact is tempered by the base rate. The 80% accuracy is important, but it must be weighed against the fact that most cabs are green.

Q4: How does Bayes' Rule help?

Bayes' Rule provides a mathematical framework to update the probability of a hypothesis (the cab being blue) in light of new evidence (the witness's statement), properly incorporating prior probabilities (base rates) and likelihoods (witness accuracy).

Q5: Can you give a real-world example of the base rate fallacy?

If a rare disease has a 1 in 10,000 prevalence (base rate), and a test for it is 99% accurate, a positive test result for a random person still has a very low probability of actually having the disease, because the vast majority of positive results will come from the 9,999 healthy people who get a false positive. Most people intuitively think a positive test means a 99% chance of having the disease, ignoring the low base rate.

Conclusion

The tale of the green and blue cabs serves as a powerful reminder that our intuition, while often a useful tool, can be a misleading guide when dealing with probabilities. By understanding and consciously applying the principles of base rates and conditional probability, we can sharpen our judgment and make more informed decisions in a world saturated with data and uncertainty. The next time you're faced with a situation involving likelihood, remember the cabs – and don't forget to check the base rate!