Bluman A.G. Elementary Statistics- A Step By Step Approach

548 Chapter 10 Correlation and RegressionSpeaking ofStatisticsIn correlation andregression studies, it isdifficult to control allvariables. This studyshows some of theconsequences whenresearchers overlookcertain aspects in studies.Suggest ways that theextraneous variablesmight be controlled infuture studies.Coffee Not Disease Culprit, Study SaysNEW YORK (AP)—Two new studies suggestthat coffee drinking, even up to 5 1 2 cupsper day, does not increase the risk of heartdisease, and other studies that claim to havefound increased risks might have missed thetrue culprits, a researcher says.“It might not be the coffee cup in onehand, it might be the cigarette or coffee rollin the other,” said Dr. Peter W. F. Wilson, theauthor of one of the new studies.He noted in a telephone interview Thursdaythat many coffee drinkers, particularlyheavy coffee drinkers, are smokers. And oneof the new studies found that coffee drinkershad excess fat in their diets.The findings of the new studies conflictsharply with a study reported in November1985 by Johns Hopkins University scientistsin Baltimore.The Hopkins scientists found that coffeedrinkers who consumed five or more cups ofcoffee per day had three times the heartdiseaserisk of non-coffee drinkers.The reason for the discrepancy appears tobe that many of the coffee drinkers in theHopkins study also smoked—and it was thesmoking that increased their heart-diseaserisk, said Wilson.Wilson, director of laboratories for theFramingham Heart Study in Framingham,Mass., said Thursday at a conference sponsoredby the American Heart Association inCharleston, S.C., that he had examinedthe coffee intake of 3,937 participants inthe Framingham study during 1956–66 andan additional 2,277 during the years1972–1982.In contrast to the subjects in the Hopkinsstudy, most of these coffee drinkers consumedtwo or three cups per day, Wilsonsaid. Only 10 percent drank six or more cupsper day.He then looked at blood cholesterol levelsand heart and blood vessel disease in the twogroups. “We ran these analyses for coronaryheart disease, heart attack, sudden death andstroke and in absolutely every analysis, wefound no link with coffee,” Wilson said.He found that coffee consumption waslinked to a significant decrease in total bloodcholesterol in men, and to a moderate increasein total cholesterol in women.Source: Reprinted with permission of the Associated Press.Exercises 10–11. What is meant by the statement that two variables arerelated? Two variables are related when a discernible patternexists between them.2. How is a linear relationship between two variablesmeasured in statistics? Explain.3. What is the symbol for the sample correlation coefficient?The population correlation coefficient? r, r (rho)4. What is the range of values for the correlationcoefficient? The range of r is from 1 to 1.5. What is meant when the relationship between the twovariables is called positive? Negative?6. Give examples of two variables that are positivelycorrelated and two that are negatively correlated.Answers will vary.7. Give an example of a correlation study, and identify theindependent and dependent variables. Answers will vary.8. What is the diagram of the independent and dependentvariables called? Why is drawing this diagramimportant? The diagram is called a scatter plot. It shows thenature of the relationship.9. What is the name of the correlation coefficient used inthis section? Pearson product moment correlation coefficient10. What statistical test is used to test the significance of thecorrelation coefficient? t test11. When two variables are correlated, can the researcherbe sure that one variable causes the other? Why orwhy not? There are many other possibilities, such as chance orrelationship to a third variable.For Exercises 12 through 27, perform the followingsteps.a. Draw the scatter plot for the variables.b. Compute the value of the correlation coefficient.10–16