Descriptive Statistics and Correlation

Summarizing data with measures of centre and spread, and quantifying the relationship between two variables.

By Alex Rivera Last reviewed: 2026-09-29 SL/HL Paper 1Paper 2

Syllabus: Topic 4: Statistics and Probability Covers subtopics 4.24.3

Overview

Before you can do any inferential statistics, you need to describe a data set clearly. This unit covers the standard summary measures, how to read them off a GDC or grouped frequency table, and how to quantify the relationship between two variables with correlation and regression.

Core Content

Measures of Centre and Spread

  • Mean $\bar{x} = \frac{\sum x}{n}$: sensitive to outliers.
  • Median: the middle value; resistant to outliers.
  • Standard deviation $\sigma$: the average distance of data points from the mean; small $\sigma$ means data is tightly clustered.

For grouped data, the IB expects you to use your GDC's statistics mode with frequency and midpoint lists rather than compute $\bar{x}$ and $\sigma$ by hand; know how to enter two-variable lists correctly, since that's where most marks are lost.

Correlation

The correlation coefficient $r$ (Pearson's) measures the strength and direction of a linear relationship, ranging from $-1$ (perfect negative) to $1$ (perfect positive); values near 0 indicate little to no linear relationship.

Linear Regression

The line of best fit, $y = mx+c$, is calculated (by GDC) to minimize the total squared vertical distance from each point to the line. It can be used to interpolate (predict within the data range) but is unreliable for extrapolation (predicting far outside it).

Why This Matters

Correlation only tells you two variables move together; it never proves one causes the other. This distinction resurfaces constantly in Probability and Probability Distributions, where confusing "related" with "caused by" is one of the most common statistical reasoning errors.

Exam Application

Paper 2 statistics questions are almost entirely GDC-driven: you're expected to enter the data, read off $\bar x$, $\sigma$, $r$, and the regression equation directly from calculator output, and interpret them in context; the marks are for correct interpretation, not manual computation. Watch for questions asking you to comment on whether a prediction is reliable (interpolation vs extrapolation).

Misconception Breakdown

Students often say "$r = 0.9$ means $x$ causes $y$." Why? A strong correlation feels like strong evidence of a causal link. But correlation only measures how closely two variables move together; a third, unmeasured factor could be driving both, or the relationship could be coincidental.

Quick Revision

  • Mean is sensitive to outliers; median is not.
  • Standard deviation: measures spread around the mean.
  • $r$: between $-1$ and $1$; measures strength/direction of a linear relationship only.
  • Correlation ≠ causation; extrapolation beyond the data range is unreliable.

Test yourself

Ten random questions from this unit, with an explanation after each answer.

Practice this unit

From the library

No shared notes for this unit yet. Write one.

Back to all Math AA units