A reading companion for Econ 250 · Bates College · Fall 2026
Hitanshu Pandit · Department of Economics
For reading I recommend Keller, Statistics for Management and Economics, 12th edition, and OpenIntro Statistics, 4th edition. This is not a replacement for either — it is my teaching notes written up, so that what I say in the lecture exists in a form you can read again afterwards.
Use it for three things: to read the week’s material in the order we actually teach it, to work the practice problems, and to drill the multiple-choice banks for each module. Keller is where you go for more exercises and the fuller treatment; this is where you go for how we did it on Monday.
Chapters are published as we reach them.
| Chapter | Covers | |
|---|---|---|
| 1 | What statistics is for | population and sample, data types, sampling designs, the two kinds of error |
| 2 | Centre: mean, median, mode | the three centres, skew, and why we report medians for income |
| 3 | Spread and standardising | variance, why $n-1$, the coefficient of variation, $z$-scores |
| 4 | Covariance and correlation | the deviation table, why units cancel in $r$, $V(X+Y)$ |
| 5 | Probability: the four rules | complement, addition, conditional, multiplication; contingency tables |
| 6 | Conditional probability and Bayes | independence, total probability, base rates |
| 7 | Random variables | $E(X)$, $V(X)$, and the laws for $aX+b$ |
| 8 | Counting successes: the binomial | the four conditions, the formula, the complement trick |
| 9 | Continuous distributions | uniform, normal and exponential; probability as area |
| 10 | $t$, chi-squared and $F$ | three distributions built out of normals, and reading their tables |
| 11 | Sampling: why your answer moves | the sampling distribution, $E(\bar X)=\mu$, the standard error |
| 12 | The Central Limit Theorem | why $\bar X$ goes normal, how large $n$ must be, the binomial case |
| 13 | Estimation and confidence intervals | point vs interval, what “95% confident” means, margin of error, sample size |
| 14 | Hypothesis testing, and what a p-value is | the eight steps, one tail vs two, p-values, Type I and II errors |
| 15 | Comparing two means | independent samples and matched pairs, and why pairing is stronger |
| 16 | Comparing several means: ANOVA | why not three $t$ tests, between vs within variation, the $F$ table |
| 17 | Categorical data | proportions, goodness of fit, and tests of independence |
| 18 | Regression | least squares, $R^2$, testing a slope, and what a line does not license |
Every chapter ends with two sets. The first is five short exercises; the second is a longer More practice block, added in October — 208 further problems across the eighteen chapters, every one with a full worked answer rather than just a letter.
They are not evenly spread. The Central Limit Theorem gets twenty and hypothesis testing gets twenty-five, because those are the two chapters where marks actually go missing. Estimation and sampling get fifteen each. Several problems in those sets have no arithmetic in them at all — they ask you to state hypotheses, or to find the error in somebody else’s conclusion, which is the skill the exam is really testing.
All with full solutions. Work them with the answers covered.
| Set | Covers |
|---|---|
| Describing data | chapters 1–4 — 15 problems |
| Chance | chapters 5–7 — 15 problems |
| Distributions | chapters 8–10 — 15 problems |
| From sample to population | chapters 11–14 — 15 problems |
| Mixed set | the whole course to date — 20 problems |
Eighty problems in these five sets, on top of the 208 in the chapters themselves. The mixed set is the one to use for Checkpoint 2 revision; the four per-Part sets are for working through a single topic.
The 130-question Checkpoint 2 practice set, split so you get only the questions for the chapter you are revising. Every one has a worked answer.
| Set | Chapter | Questions |
|---|---|---|
| Expectation and variance | 7 | 15 |
| The binomial | 8 | 15 |
| Continuous distributions | 9 | 20 |
| Sampling and the CLT | 11–12 | 17 |
| Estimation | 13 | 14 |
| Hypothesis testing | 14 | 49 |
Twenty questions per Part, deliberately hard. Most take more than one step, and every wrong option is the number you get by making one specific mistake — so the answer key tells you which mistake, not just the letter.
| Bank | Covers |
|---|---|
| Describing data | chapters 1–4 |
| Chance | chapters 5–7 |
| Distributions | chapters 8–10 |
| From sample to population | chapters 11–14 |
Chapters 3 to 10 fill in the first half of the course — spread and standardising, covariance and correlation, the probability rules, Bayes, random variables, the binomial, the normal, and the trio of $t$, $\chi^2$ and $F$. Chapters 15 to 18 cover two-sample comparisons, ANOVA, categorical data and regression.
Chapters 11 to 14 were written first because they are what we are doing now.
Read with a pen. Every worked example is set out the way I want an answer written on a problem set, so cover the solution and try it first. Reading a worked example and producing one are different skills, and only the second is any use when there is a clock running.
The examples repeat on purpose. The same commute data runs through Chapters 13 and 14; the same gym example appears in 12 and 14. You are meant to notice that a new tool is answering a different question about data you already understand, rather than spending your attention on a fresh scenario every time.
Stata appears only as it was taught. If you meet a command you have not run in a lab, that is a mistake and I would like to know about it.
Every number in the worked examples is computed by scripts rather than typed, and the Stata figures were produced by running the code. If a number here disagrees with a number on a lecture slide, the book is wrong — tell me and I will fix it.
© 2026 Hitanshu Pandit. Released under CC BY-NC-SA 4.0 — share and adapt for non-commercial purposes, with attribution.