Why hypothesis testing feels harder than it is
Many candidates do not actually struggle with the arithmetic. They struggle with the order of the logic.
Hypothesis testing asks you to do four things in sequence:
- Start with a claim about the population.
- Measure how far the sample evidence sits from that claim.
- Compare that distance with a decision rule.
- State a conclusion using disciplined language.
Once you stop treating the reading as a pile of formulas and start treating it as a decision process, the topic becomes much more manageable.
flowchart TD
A["State the null hypothesis"] --> B["Choose the alternative hypothesis"]
B --> C["Compute or read the test statistic"]
C --> D["Find the critical value or p-value rule"]
D --> E{"Evidence strong enough?"}
E -->|Yes| F["Reject the null hypothesis"]
E -->|No| G["Fail to reject the null hypothesis"]
F --> H["Discuss Type I error risk"]
G --> I["Discuss Type II error risk"]
The null hypothesis is the starting world
The null hypothesis is the condition you begin by assuming is true. In CFA problems, it usually represents a benchmark, no-change claim, or manager statement that needs testing.
Examples:
- A portfolio manager claims the mean monthly active return is `0%`.
- A credit analyst claims a screening model has no predictive edge.
- A treasury team claims average settlement time is still `2.5 days`.
The alternative hypothesis is what you are looking for evidence to support instead. It could be:
- different from the benchmark
- greater than the benchmark
- less than the benchmark
That choice matters because it determines whether the test is two-tailed or one-tailed.