Chapter 7. Inductive Arguments and Statistics
Practice Exercises: Chapter 7
Group 1: Inductive Strength and Cogency
For each argument, identify if it is Strong or Weak. Then, determine if it is Cogent or Uncogent.
-
“90% of the residents in this city support the expansion of the park. Susan is a resident of this city. Therefore, Susan likely supports the expansion.”
-
“Every time I have dropped a glass on this tile floor, it has shattered. I am about to drop this glass. Therefore, it will probably shatter.”
-
“Two people I know who drive that brand of car have had engine trouble. Therefore, that brand of car is generally unreliable.”
-
“Studies show that nearly all cats are carnivores. My pet, Whiskers, is a cat. Therefore, Whiskers is a carnivore.”
-
“A psychic told me it would rain today. Therefore, I should bring an umbrella because it will probably rain.”
Group 2: Generalization and Sampling
Identify the Target Population, the Sample, and any potential Sampling Errors (Hasty Generalization or Biased Sample).
-
To determine if high school students in the state approve of a new cell phone policy, a researcher polls 500 students at a private preparatory academy.
-
A food critic tastes one bite of a new dish and concludes that the entire menu at the restaurant is substandard.
-
A political pollster calls 1,000 randomly selected registered voters from the national database to estimate the outcome of the presidential election.
-
A gamer asks everyone on their specific Discord server what their favorite console is to determine the most popular console among all gamers worldwide.
Group 3: Probability and Cognitive Biases
Identify which concept or fallacy is at work: Frequentist Probability, Bayesian (Subjective) Probability, Gambler’s Fallacy, or the Base Rate Fallacy.
-
“The last four babies born in this hospital were girls. The next one is almost certainly going to be a boy to even things out.”
-
“I initially thought there was a 50% chance the project would succeed, but after seeing the new budget reports, I’m now 80% sure it will work.”
-
“A test for a very rare condition (1 in 10,000) is 99% accurate. Even though you tested positive, a critical thinker knows the actual probability of having the condition is still low because they consider the background frequency.”
-
“Based on data from the last 50 years, there is a 20% chance of a major flood in this valley during any given decade.”
Group 4: Statistical Pitfalls
Evaluate these scenarios using concepts like Simpson’s Paradox, Mean vs. Median, or Correlation vs. Causation.
-
A company claims the “average” salary is $80,000, but the CEO makes $5 million while most employees make $35,000. Which “average” is the company likely using, and which would be more representative?
-
Hospital A has a higher overall survival rate than Hospital B. However, when looking at specific surgery types (heart, lung, etc.), Hospital B actually has higher survival rates in every category. What is this phenomenon called?
-
Statistics show that ice cream sales and shark attacks both increase during the month of July. Does this mean eating ice cream causes shark attacks?
Answer Key
Group 1
-
Strong and Cogent (High statistical probability; assuming the premise is true).
-
Strong and Cogent (Based on consistent physical observation).
-
Weak and Uncogent (Hasty Generalization; the sample size is far too small).
-
Strong and Cogent (Based on biological fact and high probability).
-
Weak and Uncogent (The source—a psychic—does not provide reliable inductive evidence).
Group 2
-
Target: High school students in the state. Sample: 500 students at a private prep academy. Error: Biased Sample (Students at a private academy may not represent the socio-economic diversity of the whole state).
-
Target: The restaurant’s entire menu. Sample: One bite of one dish. Error: Hasty Generalization.
-
Target: The national voting population. Sample: 1,000 randomly selected voters. Error: None (This is a standard representative random sample).
-
Target: Gamers worldwide. Sample: One specific Discord server. Error: Biased Sample (The server likely attracts a specific niche of gamers).
Group 3
-
Gambler’s Fallacy (Treating independent births as if they have a “memory” of previous outcomes).
-
Bayesian (Subjective) Probability (Updating a degree of belief based on new evidence).
-
Base Rate Fallacy (The “Reasonable Person” avoids this by weighing the test result against the rare base rate).
-
Frequentist Probability (Calculating likelihood based on the observed frequency of historical events).
Group 4
-
They are using the Mean. The Median would be more representative of the “typical” employee because it isn’t skewed by the CEO’s extreme salary.
-
Simpson’s Paradox (Likely caused by Hospital B taking on more “high-risk” cases that lower its aggregate score).
-
Correlation vs. Causation (Both are likely caused by a “lurking variable”—warmer weather leading to more people eating ice cream and more people swimming in the ocean).