Day 04 — Part 2: Inferential Statistics¶
Descriptive statistics tells you what happened in your sample. Inferential statistics tells you what is likely true about the world. The difference matters enormously in practice: data scientists who confuse statistical significance with practical significance, or who misread a p-value, make expensive decisions. This session addresses those misreadings directly.
Session Overview
Difficulty: Intermediate Reading time: ~2.5 hours | Exercises: ~1 hour Prerequisites: Descriptive statistics · normal distribution (Day 04 Part 1)
What You Will Learn¶
By the end of this session you will be able to:
- State null and alternative hypotheses for a real business scenario
- Correctly explain what a p-value means (and three things it does NOT mean)
- Construct and interpret a 95% confidence interval
- Choose between a t-test, chi-squared test, and Mann-Whitney test for a given situation
- Distinguish statistical significance from practical significance
Session Roadmap¶
| # | Topic | File | Time |
|---|---|---|---|
| 1 | Hypothesis Testing — null/alternative hypotheses, Type I & II errors | 01-hypothesis-testing.md |
30 min |
| 2 | P-Value — what it means, what it does NOT mean, misconceptions | 02-p-value.md |
25 min |
| 3 | Confidence Intervals — construction, interpretation, width vs confidence | 03-confidence-interval.md |
25 min |
| 4 | Correlation — Pearson, Spearman, causation vs correlation | 04-correlation.md |
20 min |
| 5 | Statistical Tests — t-test, chi-squared, Mann-Whitney: when to use each | 05-statistical-tests.md |
25 min |
| 6 | Interview Prep — 18 questions with model answers | 06-interview-prep.md |
30 min |
Total active learning time: ~2.5 hours
How to Use This Session¶
The p-value and hypothesis testing files (01 and 02) are the most conceptually demanding — read them slowly and check every claim against the code examples. File 03 on confidence intervals is a natural continuation; do not skip it.
The interview prep in file 06 is not optional. Inferential statistics is the most common topic in data science phone screens. Work through all 18 questions with the answers hidden first, then check your responses.
Before You Start
Misconceptions about p-values are extremely common, even among experienced practitioners. Go into file 02 ready to have your existing understanding challenged — the most important part of that file is what a p-value does NOT mean.