The LSAT is not a math test, but it loves to use numbers to set traps. The good news is that the test's numerical toolkit is surprisingly small. Once you learn the core concepts and the specific ways the test exploits them, these questions become far more manageable.

The 10 Core Concepts

About 90% of the LSAT's number-based reasoning draws on these ten ideas:

1. Absolute Number: A raw count. "500 people," "12 incidents," "three studies." No context about proportion or rate.
2. Rate / Likelihood: A frequency relative to something. "1 in 10," "twice as likely," "per capita." Rates tell you how often something happens relative to a group or time period.
3. Total / Aggregate: The sum of all parts. "Total revenue," "combined output," "overall emissions." Totals can go up even if individual contributions go down, and vice versa.
4. Percentage: A proportion expressed out of 100. "40% of respondents," "a 15% increase." Percentages are relative—they depend entirely on what the base number is.
5. Average (Mean): The sum divided by the count. Averages can be pulled dramatically by outliers and tell you nothing about the distribution of individual values.
6. Median: The middle value when all values are ordered. Unlike averages, medians resist outlier distortion, but the LSAT can still trick you by conflating median with mean.
7. Weighted Average: An average where different groups contribute differently based on their size or importance. If you combine two groups with different averages, the combined average is pulled toward the larger group.
8. Rate of Change: How fast something is increasing or decreasing. "Growing by 5% per year," "declining at an accelerating rate." A slowing rate of increase still means the total is going up.
9. Market Share: A proportion of a total market. A company's market share can decrease even if its sales increase, as long as the total market grew faster.
10. Threshold: A cutoff point or minimum requirement. "At least 60%," "more than half," "a majority." Threshold arguments are vulnerable to just-barely-meets-it scenarios.

The Common Traps

The LSAT does not just test whether you know these concepts. It tests whether you can spot when an argument confuses them. Here are the most common traps, along with examples of how they appear.

Trap 1: Absolute vs. Rate

The stimulus gives you an absolute number and draws a conclusion about a rate—or vice versa. For example: "More accidents happen on dry roads than wet roads, so dry roads are more dangerous." The argument ignores that far more driving happens on dry roads. The absolute count is higher, but the rate per mile driven could be lower.

This is one of the most common numerical traps on the test. Whenever you see raw numbers being used to make a comparative claim, check whether a rate would be more appropriate.

Trap 2: Total vs. Average vs. Individual

The stimulus slides between totals, averages, and individual cases as if they are interchangeable. A company's total revenue can increase while the average revenue per employee decreases (because they hired many new employees). An individual's performance can be above average while the group's total output falls.

Watch for arguments that conclude something about individuals from group-level data, or that conclude something about totals from per-unit data.

Trap 3: Weighted Average

Two groups each improve their individual averages, but the combined average goes down. This is Simpson's Paradox, and the LSAT loves it. It happens when the proportion of people in each group shifts. If more people move into the lower-performing group, the overall average can drop even though both groups improved.

Whenever an argument combines groups and draws a conclusion about the overall average, ask whether the group sizes changed.

Trap 4: Overlapping Sets

"60% of employees exercise regularly, and 70% eat healthy diets, so at least 30% do both." That conclusion assumes the groups overlap maximally, but the actual overlap could be anywhere from 30% to 60%. The LSAT exploits the fact that people often assume non-overlapping groups when the numbers could overlap significantly—or assume full overlap when the groups might be mostly separate.

Trap 5: Part vs. Whole

An argument concludes something about the whole based on one part, or about a part based on the whole. "Sales of electric cars rose 20% last year" does not mean total car sales rose 20%. "The average household income in this city is $80,000" does not mean most households earn close to $80,000—a few very high earners could be pulling the average up.

Trap 6: Gambler's Fallacy

The argument assumes that past independent events affect future probabilities. "The coin has landed heads five times in a row, so tails is due." Each flip is independent. The LSAT tests this less often than other traps, but when it appears, it tends to be in questions about probability, risk assessment, or prediction.

Trap 7: Inverse Fallacy

The stimulus confuses the probability of A given B with the probability of B given A. "90% of people with this disease test positive" does not mean "90% of people who test positive have this disease." The base rate of the disease matters enormously. If only 1 in 10,000 people have it, most positive tests are false positives.

This trap often appears in questions about diagnostic tests, screening programs, or any argument that reverses a conditional probability.

Trap 8: Net Effect / Cost Components

The stimulus focuses on one component of a cost or benefit while ignoring others. "The new policy will save $2 million in energy costs" sounds great—until you learn it requires $5 million in new infrastructure. Or: "Emissions from factories dropped 15%" does not mean total emissions dropped, because transportation emissions may have risen more.

Whenever an argument draws a conclusion about an overall outcome based on a single component, check whether other components could offset or reverse the effect.

How to Use This on Test Day

You do not need to memorize a list. You need to build a reflex: whenever a stimulus uses numbers, pause and ask two questions. First, what numerical concept is being used? Second, is the conclusion treating that concept as if it were a different one? If the answer to the second question is yes, you have found the flaw—and the correct answer will almost certainly name it.