Introduction: Testing Claims About Averages

Have you ever seen a claim like "the average battery life of this phone is 20 hours" and wondered if it was actually true? In statistics, we don't just take people's word for it—we test it! In this chapter, we will learn how to perform a Significance Test for a Population Mean. This is the tool we use to decide if a sample average (\(\bar{x}\)) provides enough evidence to reject a claim about the true population average (\(\mu\)).

Don't worry if the formulas look a bit intimidating at first. The logic is very similar to what you learned with proportions, but now we are dealing with quantitative data (numbers like heights, weights, or time) instead of categories.

1. Defining the Two Types of Tests

In this unit, we focus on two scenarios that use the exact same mathematical process:

  • One-Sample t-Test for a Mean: We are testing a claim about a single population. Example: Is the average weight of a "16-ounce" bag of chips actually 16 ounces?
  • One-Sample t-Test for a Mean Difference (Paired Data): This happens when we have two measurements for the same individual (like "before" and "after") or pairs of similar individuals. We subtract the two values to get a single list of "differences" and then perform a test on those differences. Example: Do students score higher on a second attempt of a quiz?

Quick Tip: If you see "before and after" or "matched pairs," you are looking for a mean difference (\(\mu_d\)). Treat the differences as your data set and proceed with a one-sample test!

2. Setting Up the Hypotheses

Every test starts with two competing statements:

The Null Hypothesis (\(H_0\)): This is the "no change" or "status quo" statement. We assume the claim is true until proven otherwise. It always uses an equal sign.
\(H_0: \mu = \mu_0\) (where \(\mu_0\) is the claimed value)

The Alternative Hypothesis (\(H_a\)): This is what we are actually trying to find evidence for. It uses \(<\), \(>\), or \(\neq\).
\(H_a: \mu > \mu_0\) (One-sided, looking for a value greater than the claim)
\(H_a: \mu < \mu_0\) (One-sided, looking for a value less than the claim)
\(H_a: \mu \neq \mu_0\) (Two-sided, looking for any difference from the claim)

Key Takeaway: Always define your parameter in context! For example: "\(\mu\) = the true mean recovery time (in days) for all patients using the new medicine."

3. The "Rules of the Game": Verifying Conditions

Before we can calculate anything, we must make sure our data is reliable. For a t-test, we check three conditions:

  1. Randomization: The data must come from a random sample or a randomized experiment. This allows us to generalize our results to the population or infer cause-and-effect.
  2. 10% Condition: If we are sampling without replacement from a finite population, our sample size \(n\) must be less than 10% of the total population (\(n < 0.10N\)). This ensures the observations are independent.
  3. Normality (Large Counts): We need to know the sampling distribution of \(\bar{x}\) is approximately Normal. We can satisfy this in one of three ways:
    • The population distribution is stated to be Normal.
    • The sample size is large (\(n \ge 30\)). This is thanks to the Central Limit Theorem (CLT).
    • If \(n < 30\), we look at a graph of the sample data (like a dotplot or boxplot). As long as there is no strong skewness or extreme outliers, we are good to go!

4. Carrying Out the Test: The Math

Since we almost never know the true population standard deviation (\(\sigma\)), we use the sample standard deviation (\(s\)). This means we use the t-distribution instead of the Normal (z) distribution.

A. The Standard Error

The Standard Error of the Mean (\(SE_{\bar{x}}\)) measures how much we expect the sample mean to vary from the population mean just by chance:
\(SE_{\bar{x}} = \frac{s}{\sqrt{n}}\)

B. The Test Statistic (t-score)

The t-statistic tells us how many standard errors our observed sample mean (\(\bar{x}\)) is away from the hypothesized mean (\(\mu_0\)):
\(t = \frac{\bar{x} - \mu_0}{SE_{\bar{x}}}\)

C. Degrees of Freedom

To find the p-value using a t-table or technology, you need the degrees of freedom (\(df\)):
\(df = n - 1\)

D. The p-value

The p-value is the probability of getting a sample mean as extreme as ours (or more extreme) if the null hypothesis were actually true. We use technology (like a graphing calculator) or a t-table to find this.

Did you know? The t-distribution is slightly "fatter" in the tails than the Normal distribution. This extra space accounts for the uncertainty of not knowing the true population standard deviation!

5. Drawing and Justifying Conclusions

This is where you make your final decision. You compare your p-value to your significance level (usually \(\alpha = 0.05\)).

  • If p-value \(\le \alpha\): Reject \(H_0\). We have convincing evidence for \(H_a\).
  • If p-value \(> \alpha\): Fail to reject \(H_0\). We do not have convincing evidence for \(H_a\).
The "Golden Rule" of Interpretation:

Always write your conclusion in context. Use a template like this:
"Because the p-value (\(...\)) is less than \(\alpha = 0.05\), we reject the null hypothesis. There is convincing evidence that the [true mean context] is [greater than/less than/different from] [hypothesized value]."

Important: Never say "we accept the null hypothesis." In statistics, we never "prove" the null is true; we just fail to find enough evidence to throw it away.

Quick Review: Common Pitfalls to Avoid

  • Confusing \(\bar{x}\) and \(\mu\): Hypotheses are always about the population parameter (\(\mu\)), never the sample statistic (\(\bar{x}\)).
  • Forgetting Units: When interpreting the mean, always include the units (inches, grams, etc.).
  • Ignoring Skewness: If your sample size is small (\(n < 30\)), you must check a graph of your data for outliers before proceeding.
  • Paired Data Confusion: If the data is paired, don't do a two-sample test. Subtract the pairs first and do a one-sample test on the differences!

Key Takeaway Summary: To test a mean, check your Random, 10%, and Normal conditions, calculate your t-statistic (\(t = \frac{\text{statistic} - \text{parameter}}{\text{standard error}}\)), find the p-value using \(df = n-1\), and compare it to \(\alpha\) to make your decision.