Inference for the Difference in Two Population Means — AP Statistics
1. Core Concepts and Notation ★★☆☆☆ ⏱ 2 min
Inference for the difference in two population means (also called two-sample t-inference for means) is a collection of statistical methods used to compare the true mean of a quantitative variable across two distinct populations. This topic makes up approximately 4-5% of your total AP Statistics exam score, appearing in both multiple-choice and free-response questions.
This framework only applies when the two samples are independent, meaning no observation in one group is linked to any specific observation in the other group. We almost always conduct inference for $\mu_1 - \mu_2$, the true difference between the two population means.
2. Conditions for Two-Sample Inference ★★☆☆☆ ⏱ 3 min
Before conducting any inference, you must verify three core conditions to ensure your results are statistically valid. All three must be explicitly checked to earn full credit on AP free-response questions.
- **Random**: Both groups are independent random samples from their populations, or come from a randomized comparative experiment with two treatment groups. This ensures an unbiased sampling distribution.
- **Independence**: Individual observations within each sample are independent. When sampling without replacement, this requires the 10% condition: each sample size is less than 10% of its total population.
- **Normal/Large Sample**: The sampling distribution of $\bar{x}_1 - \bar{x}_2$ is approximately normal. This is satisfied if: (1) both samples are size ≥ 30 (Central Limit Theorem), or (2) for smaller samples, each sample distribution has no extreme outliers or strong skew.
Exam tip: Never just write 'conditions are met' on an FRQ. Explicitly verify each condition with reference to the problem context to earn full credit.
3. Confidence Intervals for $\mu_1 - \mu_2$ ★★★☆☆ ⏱ 4 min
A confidence interval for $\mu_1 - \mu_2$ gives a range of plausible values for the true difference between the two population means. It follows the standard confidence interval structure: $\text{point estimate} \pm \text{margin of error}$.
The point estimate for $\mu_1 - \mu_2$ is $\bar{x}_1 - \bar{x}_2$, the difference of the two sample means. The unpooled standard error (the default for AP Statistics) is:
SE = \sqrt{\frac{s_1^2}{n_1} + \frac{s_2^2}{n_2}}
This formula comes from the rule that the variance of the difference of two independent random variables equals the sum of their variances, so we always add the variance terms (never subtract). AP accepts two methods for degrees of freedom: the conservative method $df = \min(n_1 - 1, n_2 - 1)$ or the Welch-Satterthwaite approximation from calculators; both get full credit. The final interval is:
(\bar{x}_1 - \bar{x}_2) \pm t^* \times SE
Exam tip: If 0 is not in your confidence interval, you can reject the null hypothesis of no difference at significance level $\alpha = 1 - \text{confidence level}$, a common connection tested on multi-part FRQs.
4. Two-Sample t-Hypothesis Tests ★★★☆☆ ⏱ 3 min
Two-sample t-tests test a claim about the difference between two population means. The null hypothesis is almost always $H_0: \mu_1 - \mu_2 = 0$, meaning there is no difference between the two population means. The alternative hypothesis can be two-sided, left-sided, or right-sided, depending on the research question.
The test statistic follows the standard structure for hypothesis tests:
t = \frac{(\bar{x}_1 - \bar{x}_2) - \Delta_0}{SE}
Where $\Delta_0$ is the hypothesized difference from the null hypothesis, which is almost always 0. Standard error and degrees of freedom are calculated the same way as for confidence intervals, and you always use the unpooled approach unless explicitly told to pool.
Exam tip: Always define your parameters in words when writing hypotheses for FRQs. AP exam graders require this to earn full credit for the hypotheses step.
5. Key Tested Distinctions ★★★★☆ ⏱ 2 min
Two key distinctions are frequently tested on the AP Statistics exam, and are common sources of lost points. We cover both below.
**Pooled vs Unpooled Procedures**: Pooled t-procedures assume the two populations have equal variance, which is almost never known to be true in practice. The AP exam *never requires pooled procedures* unless explicitly instructed to use them. Unpooled is always the default.
**Independent Two-Sample vs Matched Pairs**: Two-sample inference for the difference in two population means is only for independent groups, with no linkage between observations across groups. Matched pairs data occurs when observations are paired: for example, the same subject measured before and after a treatment, or subjects matched on confounding variables like age and gender. For matched pairs, you calculate the difference within each pair and use a one-sample t-procedure, not two-sample inference.
Exam tip: AP exam questions often include a distractor that looks like two-sample data but is actually matched pairs. Always check for a pairing structure first before selecting your inference procedure.
Common Pitfalls
Why: Students confuse the difference in means with the difference in variances, forgetting variance adds for any independent random variables.
Why: Students default to two-sample because there are two groups, and miss the pairing structure that makes the groups dependent.
Why: Students forget that *both* sampling distributions need to be approximately normal, not just one.
Why: Students know t approximates z for large $n$, but forget that population standard deviations are still unknown.
Why: The true difference is a fixed value, not a random quantity, so this interpretation is incorrect.
Why: Some introductory courses teach pooled first, leading students to assume it is the standard approach.