AP Unit 7 – Inference for Means

Inference for Quantitative Data: Means

A. t-distribution:
1. When s is used instead of σ to calculate a test statistic, the corresponding
distribution, known as the t-distribution, varies from the normal distribution in
shape, in that more of the area is allocated to the tails of the density curve than
in a normal distribution. (still mound shaped, but wider than a normal curve.)
2. Degrees of freedom (d.f.) = 𝑛𝑛 − 1
3. As the degrees of freedom increase, the area in the tails of a t-distribution
decreases.
4. At 𝑑𝑑𝑑𝑑 = ∞, the critical values for t are the same as for z.

B. one-sample t-interval for a mean

1. Verify the conditions for calculating confidence intervals for a population mean:
a. To check that the sampling distribution of 𝑥𝑥̅ is approximately normal (shape):
𝒏𝒏 ≥ 𝟑𝟑𝟑𝟑 𝒐𝒐𝒐𝒐 𝒕𝒕𝒕𝒕𝒕𝒕 𝒑𝒑𝒑𝒑𝒑𝒑𝒑𝒑𝒑𝒑𝒑𝒑𝒑𝒑𝒑𝒑𝒑𝒑𝒑𝒑 𝒊𝒊𝒊𝒊 𝒏𝒏𝒏𝒏𝒏𝒏𝒏𝒏𝒏𝒏𝒏𝒏
1. Check if 𝒏𝒏 ≥ 𝟑𝟑𝟑𝟑
2. If not, see if problem stated the population is known to be normal.
3. Finally, the distribution of the sample data should be free from strong
skewness and outliers. You must graph the data in order to discuss this.
b. sample should be SRS from population of interest (discuss IN CONTEXT!!)
c. independence: If sampling without replacement, n < 10% of population.

2. Make Confidence Interval:

a. On formula sheet: 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 ± (𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣)(𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑜𝑜𝑜𝑜 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑡𝑡𝑖𝑖𝑖𝑖)
𝑠𝑠
𝑥𝑥̅ ± 𝑡𝑡 ∗
√𝑛𝑛
b. Where 𝑥𝑥̅ is the midpoint of the interval, also called the ‘point estimate’
c. 𝑡𝑡 ∗ is the critical value, determined by the confidence level, the boundaries that
have _____% between them, determined by degrees of freedom on the t-table.
If the d.f. are not on the table, use the smaller d.f.
𝑠𝑠
d. is the standard error (estimate of the CLT standard deviation)
√𝑛𝑛
𝑠𝑠
e. 𝑡𝑡 ∗ is the margin-of-error or half-width of the interval.
√𝑛𝑛

3. Interpret a Confidence Interval:

I am ____% confident that the actual mean ____(variable)______ for
____(population)____ is captured in the interval from _____ to _____.

Interpret a Confidence Level: (only if asked)

In repeated random sampling with the same sample size, approximately C% of
confidence intervals created will capture the difference of population means.
 AP Unit 7 – Inference for Means

Note: A confidence interval for a population mean either contains the population
mean or it does not, because each interval is based on data from a random sample,
which varies from sample to sample.

4. Identify what affects the margin of error:

a. When all other things remain the same, the width of the interval decreases as
the sample size increases (smaller margin of error).
b. The width of the interval increases as the confidence level increases (larger
margin of error).

C. One sample t-test for means:

1. The null hypothesis for a one-sample t-test for a population mean is 𝐻𝐻𝑜𝑜 : 𝜇𝜇 = 𝜇𝜇𝑜𝑜 ,
where 𝜇𝜇𝑜𝑜 is the hypothesized value. Depending upon the situation, the alternative
hypothesis is 𝐻𝐻𝑎𝑎 : 𝜇𝜇 < 𝜇𝜇𝑜𝑜 , 𝐻𝐻𝑎𝑎 : 𝜇𝜇 > 𝜇𝜇𝑜𝑜 , or 𝐻𝐻𝑎𝑎 : 𝜇𝜇 ≠ 𝜇𝜇𝑜𝑜 .
2. Verify conditions are met: same as for confidence interval.
3. test statistic is a t-score since we usually don’t know 𝜎𝜎 and are substituting 𝑠𝑠.
𝑥𝑥̅ −𝜇𝜇
𝑡𝑡 = 𝑠𝑠
√𝑛𝑛
4. the p-value is found using the t-table and the d.f. Look for the t-score on the d.f.
row, and the p-values are at the top of the table. You will usually be reporting that
the p-value is between two numbers.

D. Two-Sample t-interval for the difference in two population means:

1. Verify the conditions:
a. To check that the sampling distribution of 𝑥𝑥̅1 − 𝑥𝑥̅ 2 is approximately normal:
1. Check if 𝒏𝒏𝟏𝟏 𝒂𝒂𝒂𝒂𝒂𝒂 𝒏𝒏𝟐𝟐 ≥ 𝟑𝟑𝟑𝟑
2. If not, see if problem stated the populations are known to be normal.
3. Finally, the distribution of the sample data should be free from strong
skewness and outliers. You must graph the data in order to discuss this. This
should be checked for both sets of data if necessary.
b. Data should be collected using two independent, random samples or an
experiment with random assignment.
c. independence: If sampling without replacement, n < 10% of population for both
samples.

2. Make Confidence Interval:

a. On formula sheet: 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 ± (𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣)(𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑜𝑜𝑜𝑜 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠)
𝑠𝑠2 𝑠𝑠2
(𝑥𝑥̅1 − 𝑥𝑥̅2 ) ± 𝑡𝑡 ∗ �𝑛𝑛1 + 𝑛𝑛2
1 2

b. Where 𝑥𝑥̅1 − 𝑥𝑥̅2 is the midpoint of the interval, also called the ‘point estimate’
 AP Unit 7 – Inference for Means

c. 𝑡𝑡 ∗ is the critical value, determined by the confidence level, the boundaries that
have _____% between them, determined by degrees of freedom on the t-table.
If you use the table, then use the smaller n-1 as degrees of freedom and hand
calculate your confidence interval.
If using the calculator, report the df given by the calculator along with the
interval.
𝑠𝑠2 𝑠𝑠2
d. �𝑛𝑛1 + 𝑛𝑛2 is the standard error (estimate of the CLT standard deviation)
1 2

𝑠𝑠12 𝑠𝑠2
e. 𝑡𝑡 ∗ �𝑛𝑛 + 𝑛𝑛2 is the margin-of-error or half-width of the interval
1 2

3. Interpret a Confidence Interval:

I am ____% confident that the actual difference in the mean ____(variable)______
for ____(population 1)____ and ____(population 2)____ is captured in the interval
from _____ to _____.

Note: If 0 is included in the interval, it is plausible that there is no difference

between the population means.
If both endpoints are positive, it’s likely that the first population has a greater mean
than the second, and opposite for if both endpoints are negative.

E. Two sample t-test for the difference between population means:

1. The null hypothesis is 𝐻𝐻𝑜𝑜 : 𝜇𝜇1 = 𝜇𝜇2 or 𝐻𝐻𝑜𝑜 : 𝜇𝜇1 − 𝜇𝜇2 = 0. Depending upon the
situation, the alternative hypothesis is 𝐻𝐻𝑎𝑎 : 𝜇𝜇1 < 𝜇𝜇2, 𝐻𝐻𝑎𝑎 : 𝜇𝜇1 > 𝜇𝜇2 , or 𝐻𝐻𝑎𝑎 : 𝜇𝜇1 ≠ 𝜇𝜇2 .
2. Verify conditions are met: same as for confidence interval.
3. test statistic is a t-score since we usually don’t know 𝜎𝜎 and are substituting 𝑠𝑠.
(𝑥𝑥̅ 1 −𝑥𝑥̅2 )−0
𝑡𝑡 =
𝑠𝑠 2
𝑠𝑠 2
� 1+ 2
𝑛𝑛1 𝑛𝑛2

4. The p-value is found using the t-table and the d.f. Look for the t-score on the d.f.
row, and the p-values are at the top of the table. You will usually be reporting that
the p-value is between two numbers.
If you use the table, then use the smaller n-1 as degrees of freedom and hand
calculate your confidence interval.
If using the calculator, report the df given by the calculator along with the given p-
value.
 AP Unit 7 – Inference for Means

F. Matched Pairs
1. Matched pairs can be thought of as one sample of pairs. They are not two
independent random samples.
2. Find the differences of the pairs of data, then calculate the mean and standard
deviation of the differences (𝑥𝑥̅𝑑𝑑 𝑎𝑎𝑎𝑎𝑎𝑎 𝑠𝑠𝑑𝑑).
3. Once differences between pairs are found, finding a confidence interval proceeds as
for a one sample t interval, but for the population mean difference.
3. The null hypothesis for a matched pairs t-test for a population mean difference is
𝐻𝐻𝑜𝑜 : 𝜇𝜇𝑑𝑑 = 0. Depending upon the situation, the alternative hypothesis is 𝐻𝐻𝑎𝑎 : 𝜇𝜇𝑑𝑑 < 0,
𝐻𝐻𝑎𝑎 : 𝜇𝜇𝑑𝑑 > 0, or 𝐻𝐻𝑎𝑎 : 𝜇𝜇𝑑𝑑 ≠ 0.
4. The rest of the significance test proceeds as for a one sample t-test, but for the
population mean difference.

**Interpretations, conclusions, types of errors, what makes intervals wider/narrower, how

power is affected, etc. are all the same for means as for proportions. Refer back to your Unit
6 notes for this information.