Hypothesis Testing with Z-Tests

1.1 The Logic of Hypothesis Testing

In descriptive statistics, we summarize data. In inferential statistics, we use sample data to draw conclusions about a larger population. The most structured way to do this is through hypothesis testing — a formal procedure for deciding whether sample evidence supports or contradicts a claim about a population parameter.

Every hypothesis test begins with two competing statements:

• Null hypothesis (H₀): A statement of “no effect” or “no difference.” It represents the status quo — the claim we assume to be true unless strong evidence contradicts it.
• Alternative hypothesis (H_a): The claim we are trying to find evidence for. It contradicts the null hypothesis.

The burden of proof lies with H_a. Just as a defendant is presumed innocent until proven guilty, H₀ is presumed true until the data provide convincing evidence against it. We never “prove” H₀ — we either reject it or fail to reject it.

Type I and Type II Errors

Because we make decisions based on samples (not the entire population), we can make mistakes:

• Type I error (α): Rejecting H₀ when it is actually true. This is a “false positive.” The probability of a Type I error is the significance level α, typically set at 0.05.
• Type II error (β): Failing to reject H₀ when it is actually false. This is a “false negative.”

Lowering α reduces the chance of a Type I error but increases the chance of a Type II error. Analysts must balance these risks based on the business context.

1.2 The Z-Test for a Population Mean

The Z-test is used to test a hypothesis about a population mean when two conditions are met: (1) the population standard deviation σ is known, and (2) the sample size is n ≥ 30 (or the population is normally distributed). Under these conditions, the sampling distribution of x is approximately normal, and we can use the standard normal (Z) distribution to calculate our test statistic.

The Z-Test Statistic

The test statistic measures how many standard errors the sample mean is away from the hypothesized population mean. A large absolute value of Z indicates that the sample mean is far from what we would expect if H₀ were true.

You can also use Excel's STANDARDIZE function: =STANDARDIZE(x, μ0, σ/SQRT(n)), which computes the same result.

1.3 One-Tailed vs Two-Tailed Tests

A two-tailed test checks for a difference in either direction (H_a: μ ≠ μ₀). A one-tailed test checks for a difference in a specific direction:

• Right-tailed: H_a: μ > μ₀ (the parameter is greater than claimed)
• Left-tailed: H_a: μ < μ₀ (the parameter is less than claimed)

The choice between one-tailed and two-tailed depends on the research question. If you only care about one direction of departure, a one-tailed test is more powerful because all of α is concentrated in one tail.

Effect on Critical Values

For α = 0.05:

• Two-tailed: Critical values are ±1.96 (splitting α/2 = 0.025 in each tail)
• One-tailed: Critical value is 1.645 (right-tail) or −1.645 (left-tail), placing all of α in one tail

Worked Example: One-Tailed Z-Test

Using the same GreatLakes data (Z = 2.83), suppose the engineer specifically tests whether the diameter has increased:

• H₀: μ = 10.00   H_a: μ > 10.00 (right-tailed)
• Critical value at α = 0.05: Z = 1.645
• Since 2.83 > 1.645, reject H₀
• The one-tailed p-value is approximately 0.0023 (half the two-tailed p-value)

1.4 The P-Value Approach

The p-value is the probability of observing a test statistic as extreme as (or more extreme than) the one calculated, assuming H₀ is true. It quantifies the strength of evidence against H₀:

• If p-value ≤ α, reject H₀
• If p-value > α, fail to reject H₀

Common Misinterpretations

The p-value is frequently misunderstood. It is not:

• The probability that H₀ is true
• The probability that H_a is true
• A measure of the size of the effect

A small p-value means the observed data would be unlikely if H₀ were true. It does not tell us the probability that any hypothesis is true or false — it tells us about the data, given an assumption.

Calculating P-Values in Excel

For a two-tailed Z-test:

• =2*(1-NORM.S.DIST(ABS(Z),TRUE))

For a one-tailed test (right tail):

• =1-NORM.S.DIST(Z,TRUE)

Chapter Summary

In this chapter, we established the logical framework of hypothesis testing and applied it using the Z-test for a population mean. Here is what you should take away: