Chapter 23 Inferences about Means
Review One Quantitative Variable Population Mean Value _____ Population Standard Deviation Value ____
Review Estimate ________ Take random sample Calculate sample mean ________ Calculate sample standard deviation _______
Long Term Behavior of Sample Mean Statistic Sampling distribution of sample mean For variables with normal distributions. For variables with non-normal distribution when sample size n is large.
Problem: ____________________ Solution: Replace _______________ with __________________________. Standard error of the sample mean
Sampling distribution of Sample Mean
The t distribution Different t distribution for each value of ________.
Using the t distribution Assumptions Random sample. Independent values. No more than 10% of population sampled. Nearly Normal Population Distribution. __________________________________________
History of t distribution William S. Gosset Head brewer at Guinness brewery in Dublin, Ireland. Field experiments - find better barley and hops. Small samples Unknown σ. Published results under name Student. t distribution also called Student’s t.
The t distribution t distribution _________________________________
t distribution table Row = degrees of freedom. Column One tail probability. Table value = t* where P(T(n-1) > t*) = α Two tail probability. Table value = t* where P(T(n-1) > t*) = α/2 t* = critical value for t distribution.
Inference for μ C% Confidence interval for μ. t* comes from t distribution with (n-1) d.f.
Example Find t* for 95% CI, n = 10 90% CI, n = 15 99% CI, n = 25
Example #1 A medical study finds that in a sample of 27 members of a treatment group, the sample mean systolic blood pressure was with a sample standard deviation of 9.3. Find a 90% CI for the population mean systolic blood pressure.
Example #1 (cont.) d.f. = ___________ t* = __________ Assumption: Blood pressure values must have a fairly symmetric distribution.
Example #1 (cont.)
Example #2 Medical literature states the mean body temperature of adults is In a random sample of 52 adults, the sample mean body temperature was with a sample standard deviation of Find a 95% confidence interval for the population mean body temperature of adults.
Example #2 (cont.) d.f. = ___________ t* = Assumption: ______________________
Example #2 (cont.)
Hypothesis Test for μ H O : _______________ H A : Three possibilities _____________
Hypothesis Test for μ Assumptions
Hypothesis Test for μ Test Statistic
P-value for H A : ___________ P-value = P(t n-1 > t)
P-value for H A : ____________ P-value = P(t n-1 < t)
P-value for H A : _________ P-value = 2*P(t n-1 > |t|)
Hypothesis Test for μ P-value Small ________________________________ _______________________________________ Large ________________________________ _______________________________________ Small and large p-values determined by α.
Hypothesis Test for μ If p-value < α If p-value > α
Hypothesis Test for μ Conclusion: Always stated in terms of problem.
Example #1 A medical study finds that in a sample of 27 members of a treatment group, the sample mean systolic blood pressure was with a sample standard deviation of 9.3. Is this enough evidence to conclude that the mean systolic blood pressure of the population of people taking this treatment is less than 120. Use α = 0.1
Example #1 (cont.) Ho:____________ Ha:____________ Assumptions
Example #1 (cont.)
d.f. = ______________ P-value
Example #1 (cont.) Decision: Conclusion:
Example #2 The manufacturer of a metal TV stand sets a standard for the amount of weight the stand must hold on average. For a particular type of stand, the average is set for 500 pounds. In a random sample of 16 stands, the average weight at which the stands failed was pounds with a standard deviation of 10.4 pounds. Is this evidence that the stands do not hold the standard average weight of 500 pounds? Use α = 0.01
Example #2 (cont.) Ho: ____________ Ha: ____________ Assumptions
Example #2 (cont.)
d.f. = ________ P-value
Example #2 (cont.) Decision: Conclusion:
Example #3 During an angiogram, heart problems can be examined through a small tube threaded into the heart from a vein in the patient’s leg. It is important the tube is manufactured to have a diameter of 2.0mm. In a random sample of 20 tubes, they find the mean diameter of the tubes is 2.01mm with a standard deviation of 0.01mm. Is this evidence that the diameter of the tubes is different from 2.0mm? Use α = 0.01
Example #3 (cont.) Ho:______________ Ha:______________ Assumptions
Example #3 (cont.)
d.f. = ___________ P-value
Example #3 (cont.) Decision: Conclusion: