1. Chapter 9
Hypothesis Testing

2. GED111/CDS111 Statistics in Modern Society
Hypothesis Testing
A hypothesis is a claim about a population parameter, such as a population proportion, P, H0, gives a specific value for a population parameter. When testing a claim about a population mean or population proportion, we (usually) write the null hypotheses in the form
H0 : population parameter = claimed value
GED111/CDS111 Statistics in Modern Society

3. Hypothesis Testing (cont.)
The alternative hypothesis, or Ha, is a statement that the population parameter has a value that somehow differs from the value claimed in the null hypothesis.
There are always 2 possible outcomes of a hypothesis test:
Reject the null hypothesis, which lends support to the alternative hypothesis
Do not reject the null hypothesis, note that this does not prove that the null hypothesis is true
GED111/CDS111 Statistics in Modern Society

4. Formulating Hypotheses
The manufacturer of a new fuel-conserving car advertises that the car averages 38 miles per gallon on the highway. A consumer group claims that the true mean is less than 38 miles/gal.
H0 : µ = 38 miles per gallon
Ha : µ < 38 miles per gallon
GED111/CDS111 Statistics in Modern Society

5. Statistical Significance
Rare Event Rule
If, under a given assumption (such as the null hypothesis), the probability of a particular event at least as extreme as the observed event is very small (such as 0.05 or less), we conclude that the assumption is probably not correct.
GED111/CDS111 Statistics in Modern Society

6. GED111/CDS111 Statistics in Modern Society
p-value
The p-value for a hypothesis test of a claim about a population parameter is the probability of selecting a sample at least as extreme as the observed sample, assuming that the null hypothesis is true.
GED111/CDS111 Statistics in Modern Society

7. Example : Significance and Birth Weight
A county health official believes that the mean birth weight of male babies at a local hospital is greater than the national average of 3.39 kg. A random sample of 145 male babies born at that hospital has a mean birth weight of 3.61kg, a calculation shows that the probability of selecting a sample with a mean birth weight of 3.61kg or more is
Formulate the null and alternative hypotheses
What is the P-value for this sample?
Is the difference between the population mean (3.39kg) and the observed mean (3.61kg) significant at the 0.05 level?
GED111/CDS111 Statistics in Modern Society

8. Example : Significance and Birth Weight (cont.)
The null hypothesis is the claim that the mean birth weight of all male babies born at this hospital is the national average of 3.39kg ; H0 : µ = 3.39kg
The alternative hypothesis (formulated before the sample is selected) is the claim of the health official; Ha : µ>3.39kg
The p-value is 0.032; it is the probability of randomly selecting a sample with a mean of at least 3.61 kg (assumption that the population mean is really 3.39kg
GED111/CDS111 Statistics in Modern Society

9. Example : Significance and Birth Weight (cont.)
Because the p-value is less than 0.05, the difference is significant at the 0.05 level. Based on the sample, there is sufficient evident to reject the null hypothesis, thereby supporting the alternative hypothesis that the mean weight for all male babies born at the hospital is greater than 3.39kg.
GED111/CDS111 Statistics in Modern Society

10. Legal Analogies of Hypothesis Testing
In courts of law, the fundamental principle is that a defendant is presumed innocent until proven guilty.
H0 : the defendant is innocent
Ha : the defendant is guilty
A defendant is found guilty or NOT guilty, but they are never found to be innocent.
A verdict of not guilty means the evidence is not sufficient to establish guilt, but it does not prove innocence.
GED111/CDS111 Statistics in Modern Society

11. Medical Analogies of Hypothesis Testing
A physician generally starts with the assumption of normal health (no disease) and then looks for evidence that a disease is actually present.
H0 : the disease is absent.
Ha : the disease is present
The aim of a physician is to collect enough evidence (the sample data) to reject the null hypothesis and conclude that disease is present.
GED111/CDS111 Statistics in Modern Society

12. GED111/CDS111 Statistics in Modern Society
Exercise
Q13 p378
Q15-18 p379
GED111/CDS111 Statistics in Modern Society

13. Setting Up hypothesis Tests
Forms of the Null and Alternative Hypotheses
The null hypothesis gives a specific value for a population parameter. Thus, it has the form that includes equality
H0 : population parameter = claimed value
The alternative hypothesis has one of the following forms
(left-tailed) Ha : population parameter < claimed value
(right-tailed) Ha : population parameter > claimed value
(two-tailed) Ha : population parameter ≠ claimed value
The hypotheses should be formulated before sample data are analyzed, and one should never test a hypothesis using the same data that suggested the hypothesis.
GED111/CDS111 Statistics in Modern Society

14. Hypothesis Test Requirements
The claimed value of the population parameter. This value may be either a population mean, µ, or a population proportion, p.
The sample mean x, or the sample proportion, p.
The sample size, n.
In the case of a population mean, we also need the population standard deviation, σ, but for large samples we can approximate it by the sample standard deviation, s. _ ^
GED111/CDS111 Statistics in Modern Society

15. Hypothesis Test Requirements (cont.)
We use the above information to determine the probability of observing the sample statistics found in the study (p-value), assuming the null hypothesis is true. Based on this probability, we decide whether to reject the null hypothesis.
GED111/CDS111 Statistics in Modern Society

16. GED111/CDS111 Statistics in Modern Society
Example : Mean Rental Car Mileage pp (1st ed.)
Example : Pre-Election Polls pp (1st ed.)
Exercise : Q5-14 pp (1st ed.)
GED111/CDS111 Statistics in Modern Society

17. Stating the Results : p-values
Significance levels and standard scores (normal distribution) allow us to make a yes/no decision about rejecting the null hypothesis. However, it is common practice to be more specific about the strength of the evidence for rejecting the null hypothesis.
p-value is used to state the significance of the sample (observed) data.
GED111/CDS111 Statistics in Modern Society

18. Interpretation of p-values
Less than 0.01
Test is highly statistically significant and offers strong evidence against H0
0.01 to 0.05
Test is statistically significant and offers moderate evidence against H0
Greater than 0.05
Test is not statistically significant and does not offer sufficient evidence against H0
GED111/CDS111 Statistics in Modern Society

19. Statistical Significance and Practical Significance
Consider a weight lost program that guarantees weight loss after 2 days in the program. Suppose a random sample of thousands of people in the program has a mean weight loss of 0.37 pound after 2 days and this mean weight loss proves to be significant at the 0.05 significant level (in another word, the null hypothesis of no weight loss is rejected).
Despite this statistical significance, the mean weight loss of 0.37 pound is so small that it has virtually no practical significance.
GED111/CDS111 Statistics in Modern Society

20. Statistical Significance and Practical Significance (cont.)
Consider a large company in which the employees claim they are underpaid compared to the national average for workers in the same position. Suppose a sample of 50 employees has a mean monthly salary of US$2,500, compare to a national mean salary of US$2,950. A significance test give a p-value of 0.072, which is not significant at the 0.05 level. Because of factors such as large variation and a small sample, the difference of US$450 might not be statistically significant, but it could have practical significance for the employees.
GED111/CDS111 Statistics in Modern Society

21. GED111/CDS111 Statistics in Modern Society
Two-tailed Test
Ha : µ ≠ claimed value
Consider a drug company that seeks to be sure that its 500mg aspirin tablets really contain 500mg of aspirin. If the tablets contain less than 500mg, consumers are not getting the advertised dose. If the tablets contain more than 500mg, consumers are getting too much of the drug.
Null hypothesis : the population mean of the aspirin content is 500mg; H0 : µ = 500mg
Alternative hypothesis : the population mean content is either less than or greater than 500mg; Ha : µ ≠ 500mg
GED111/CDS111 Statistics in Modern Society

22. Errors in Hypothesis Testing
Type I
H0 is wrongly rejected
Type II
Wrongly fail to reject H0
Significance level is the probability of making a type I error.
GED111/CDS111 Statistics in Modern Society

23. Errors in Hypothesis Testing
Decision Table for H0 and Ha
Reality
H0 True
Ha True
Decision
Reject H0
Type I error
Correct Decision
Do not reject H0
Type II error
GED111/CDS111 Statistics in Modern Society

24. GED111/CDS111 Statistics in Modern Society
Exercise
Q5-7 p390(1st ed.)
GED111/CDS111 Statistics in Modern Society

25. Hypothesis Testing for Proportion
Formulate H0 in the form p = claimed value. Formulate Ha in the form p< claimed value (left-tail), p>claimed value (right-tailed), or p≠claimed value (two-tailed).
Identify the relevant sample statistics : the sample size, n, and the sample proportion, p.
Determine the outcome of the test either by comparing the standard score with the critical value or by computing a p-value. If the p-value is less than or equal to 0.05, then the test is significant at the 0.05 level; there is sufficient evidence to reject H0. If the p-value is greater than 0.05, H0 cannot be reject. ^
GED111/CDS111 Statistics in Modern Society

26. GED111/CDS111 Statistics in Modern Society
Exercise
Chapter Review Exercise
Q1 p401
GED111/CDS111 Statistics in Modern Society

27. GED111/CDS111 Statistics in Modern Society
Focus on Agriculture
Are Genetically Modified Foods Safe? Pp
GED111/CDS111 Statistics in Modern Society