Why Check the Normal/Large Sample Condition?
Before using a t procedure to make an inference about a population mean, check whether the procedure’s model is reasonable. The Normal/Large Sample condition addresses the shape of the sampling distribution of the sample mean, \(\bar{x}\). As discussed in “Sampling Distribution When the Population Is Normal” and “Central Limit Theorem Explained,” that sampling distribution is Normal, or approximately Normal, when the population is approximately Normal or the sample is sufficiently large.
This condition is separate from the random condition and the 10% condition. In “Checking the Random Condition,” we considered how the data were collected. In “Checking the 10% Condition for Independence,” we checked whether a sample drawn without replacement was small enough relative to its population to support approximate independence. Passing either check does not tell us whether the Normal/Large Sample condition is met.
For AP Statistics mean inference, the practical rule is to check whether the population is approximately Normal or the sample size is at least 30. When \(n<30\), the population’s shape matters: you need evidence that it is approximately Normal. When \(n\geq 30\), the Large Sample rule supports using an approximately Normal model for \(\bar{x}\).
How to Apply the Rule
First identify the sample size \(n\). If \(n\geq 30\), the Large Sample route is met under the AP rule. If \(n<30\), do not conclude that the condition fails automatically. Instead, consider whether the population distribution is approximately Normal. In practice, a graph of the sample data—such as a dotplot, histogram, or boxplot—can help assess the shape when the population distribution itself is not given.
For a small sample, look for a roughly symmetric, unimodal pattern without strong skewness or outliers. These features are evidence that an approximately Normal model may be reasonable. A graph cannot prove that the entire population is Normal, but it can help you decide whether the data are consistent with that assumption. If the sample is too small to give a useful picture, and no information about the population’s shape is provided, say that the condition cannot be verified from the available information.
The condition is about the distribution of \(\bar{x}\), not a claim that the individual observations themselves become Normal when \(n\geq 30\). For a small sample from an approximately Normal population, \(\bar{x}\) is approximately Normal because of the population’s shape. For a sufficiently large sample, the Central Limit Theorem supports an approximately Normal sampling distribution for \(\bar{x}\), even if individual observations are not Normal.
This is a practical AP rule, not a statement that the sample size changes the population distribution. A large \(n\) supports the model for sample means; it does not make the individual measurements Normal. Also, checking this condition alone does not establish that a t procedure is appropriate. Randomness and independence still require their own checks.
A Step-by-Step Check
A clear response identifies the route used and gives the evidence for it. If you use the large-sample route, state the sample size and compare it with 30. If you use the population-shape route, describe the information that supports an approximately Normal population. If neither route can be checked, explain what information is missing rather than guessing.
Find \(n\), the number of observations used to calculate the sample mean.
If \(n\geq 30\), state that the sample meets the Large Sample rule for this condition.
Use information about the population or inspect a graph of the sample data for evidence of an approximately Normal shape.
Connect the evidence to whether an approximately Normal model for \(\bar{x}\) is reasonable. Keep this conclusion separate from the random and independence checks.
Worked Examples
Worked Example: A Small Sample From an Approximately Normal Population
A beverage plant measures the fill volume of 15 randomly selected bottles from a production run of 500 bottles. The production process is known to produce fill volumes that are approximately Normal. The sample is taken without replacement. Is the Normal/Large Sample condition supported?
State. Let \(\mu\) be the mean fill volume of bottles in this production run. We are checking whether the Normal/Large Sample condition supports a t procedure for \(\mu\).
Plan. The sample size is \(n=15\), which is less than 30, so the Large Sample route does not apply. We can instead use the information that the population distribution of fill volumes is approximately Normal. Because the sampling is without replacement, we should also check the 10% condition for independence.
Do. The population has \(N=500\) bottles, and the sample is \(n=15\). The sample’s share of the population is
The sample is no more than 10% of the population, so the 10% condition is met. For the condition considered here, the population is described as approximately Normal, which supports an approximately Normal sampling distribution of \(\bar{x}\) even though \(n=15\).
Conclude. The Normal/Large Sample condition is supported because the population distribution of fill volumes is approximately Normal. The 10% condition is also met, supporting approximate independence for this sample. The random selection supports the random condition. These checks together support using a one-sample t procedure, subject to the stated sampling details.
Worked Example: Using a Sample Graph When \(n<30\)
A school selects a simple random sample of 24 students from 720 students and records each student’s weekly time spent practicing a musical instrument. A dotplot of the sample shows one main cluster, roughly balanced on either side of its center, with no clear outliers. Does the sample provide evidence for the Normal/Large Sample condition?
State. Let \(\mu\) represent the mean weekly practice time for students at the school. We need to assess whether the Normal/Large Sample condition is supported for inference about \(\mu\).
Plan. The sample size is \(n=24\), so it does not meet the \(n\geq 30\) rule. We will use the sample’s shape as evidence about the population distribution. For a small sample, a roughly symmetric pattern without strong skewness or outliers is consistent with an approximately Normal population. We will separately check randomness and the 10% condition.
Do. The dotplot has one main cluster, is roughly balanced around its center, and shows no clear outliers. These features support treating the population distribution as approximately Normal for this inference. The sample is a simple random sample, supporting the random condition. For independence, compare the sample size with the population size:
Since \(3.3\%\) is less than 10%, the 10% condition is met.
Conclude. Although \(n=24\) is less than 30, the sample graph is roughly unimodal and symmetric with no clear outliers, giving evidence that the population distribution is approximately Normal. Thus the Normal/Large Sample condition is reasonably supported. The random and 10% checks also support using a one-sample t procedure for the school’s mean weekly practice time.
Worked Example: Applying the Large Sample Route
A community garden coordinator takes a random sample of 42 garden plots from a group of 480 plots and records the mass of produce harvested from each plot during one week. The distribution of individual harvest masses is described as right-skewed. Is the Normal/Large Sample condition met by the AP rule?
Identify the sample size. There are \(n=42\) plots in the sample. Since \(42\geq 30\), the sample meets the Large Sample rule.
Keep the distributions distinct. The individual harvest masses are right-skewed, but the condition concerns whether the sampling distribution of the sample mean is approximately Normal. The Large Sample route supports an approximately Normal sampling distribution for \(\bar{x}\); it does not claim that individual plot harvests are Normal.
Check independence. The plots were sampled without replacement from 480 plots. Their sampled fraction is
Because \(8.75\%\) is less than 10%, the 10% condition is met. The description says the sample is random, which supports the random condition.
Conclusion. The Normal/Large Sample condition is supported by \(n=42\), which is at least 30. The right-skewed shape of individual harvest masses does not by itself negate the Large Sample route in this AP rule. The random and 10% checks are also supported, so the stated conditions for a one-sample t procedure are met.
Worked Example: When the Shape Information Is Missing
A researcher measures the time required to complete a short computer task for 17 randomly selected volunteers. The description gives no information about the population distribution and provides no graph or list of sample values. Can the Normal/Large Sample condition be checked?
The sample size is \(n=17\), so \(n<30\) and the Large Sample route is not available. Because the population shape is not described and no sample graph or observations are given, there is no evidence for deciding whether the population is approximately Normal. The word “random” supports the random condition but does not provide information about shape.
Conclusion. The Normal/Large Sample condition cannot be verified from the information given. A careful response should request information about the population distribution or the sample data’s shape. It should not assume that the condition holds simply because the sample was random, nor should it claim the condition fails solely because \(n<30\).
Common Mistakes and AP Exam Tips
- Treating 30 as a minimum for every situation. A sample smaller than 30 can still meet the condition when the population is approximately Normal. State which route applies rather than stopping at the sample-size comparison.
- Claiming a small sample is fine without shape evidence. If \(n<30\), give evidence about the population shape, often from a graph of the sample data. “The data look acceptable” is too vague; describe features such as rough symmetry, one main cluster, and no clear outliers.
- Confusing individual observations with sample means. The Large Sample route supports an approximately Normal sampling distribution of \(\bar{x}\). It does not say that individual values become Normal.
- Confusing randomness with Normality. Random selection concerns how the sample was obtained. It does not guarantee an approximately Normal population or a particular sample shape.
- Using the condition to cover every assumption. A supported Normal/Large Sample condition does not replace the random condition or, for sampling without replacement, the 10% condition. As “Why Conditions Matter in Mean Inference” emphasizes, each condition addresses a different concern.
- Saying a graph proves the population is Normal. A sample graph is evidence, not proof. For a small sample, describe what the graph suggests and make a suitably cautious conclusion.
A strong AP response connects the evidence to the specific condition. For example: “Because the sample size is 42, which is at least 30, the Large Sample rule is met, so an approximately Normal model for the sampling distribution of \(\bar{x}\) is reasonable.” For a smaller sample, name the observed shape features and say they provide evidence that the population is approximately Normal.
Check Your Understanding
For each situation, decide whether the Normal/Large Sample condition is supported, not supported, or cannot be checked from the information given. Explain your reasoning.
- A random sample of 14 patients is drawn from a population described as approximately Normal. Which route supports the condition?
- A sample of 35 randomly selected measurements comes from a population described as strongly right-skewed. Does the sample meet the Large Sample rule? What distribution does that rule concern?
- A sample of 21 delivery times has a roughly symmetric, unimodal dotplot with no clear outliers. What evidence can you cite, and what should you avoid claiming?
- A sample of 18 randomly selected workers is described, but neither the population shape nor a graph of sample values is provided. Can you verify the condition?
- In your own words, explain why meeting the Normal/Large Sample condition does not by itself establish the random or 10% condition.