Name That Difference Test

= NAME THAT SIGNIFICANCE TEST!

Note: these ar= e all difference tests

Scroll down to the bottom f= or the answers:

1. Are there significantly more cars on freeways displ= aying “Barack Obama= ” bumper stickers compared to “John McCain” bumper stickers? A social scientists selects 10 freeway corridors at random and tallies the average number of bumper stickers for each candidate What difference test should be used?

2. Do car buyers prefer some car colors significantly = more than others? A researcher selects 200 new car buyers, at random, and records the colors of the cars they purchased.&nbs= p; Six car colors are preferred; black, blue, green, red, silver, and white. How can the researcher determine i= f any differences in car color preferences are significant?

3. An organizational consultant wants to know whether = female managers use more pro-social strategies than male managers. One hundred female and male manage= rs, selected at random respond to a series of hypothetical situations, such as chronic lateness, or the use of offensive language. The strategies the managers report= they would use are recorded as either “positive/proso= cial,” “neutral/informational,” or “negative/antisocial” in nature. How can the researcher determine whether one gender uses significan= tly more of one strategy than the other gender?

4. A college dean compares the GPAs of students in thr= ee different majors; accounting, marketing, and finance. How can the dean tell if there are significant differences in GPA across the th= ree groups?

5. An employee has an opportunity to be promoted from = an hourly wage to a guaranteed salary. However, the employee wonders whether she will have to work significantly more hours as a salaried employee. She compares the average weekly wo= rk hours for 36 hourly employees and 36 salaried employees to see if there is a significant difference between the two.&nb= sp; What statistical test would= tell her the answer?

6. A pharmaceutical firm compares the effectiveness of= a non-prescription herbal remedy for depression (St. John’s W= ort) with a prescription medication (Zoloft) with two levels (mild versus severe= ) of depression. The data on the t= wo remedies’ effectiveness consists of scores on “Beck’s depression inventory” a scale that measures how depressed a person is. What statistical test would reveal any differences in the two medications’ effectiveness for different levels of depression?

7. A talk show host asks her audience to raise their h= ands to vote for whichever of two movie stars they think is most sexy; Orlando B= loom or Johnny Depp. The vote is 84 for Orla= ndo and 79 for Johnny. A social scientist in the audience wonders whether that sort of difference could be attributed to chance. What test should the social scientist employ?

8. At a chili cook-off, five recipes are selected for = the final taste test. Ten judges = taste each of the chile recipes in random order and r= ate each one on a 10 point scale. What statistical test would reveal if= one recipe tastes significantly better than the others?

9. A researcher wants to know if male versus female te= ens, and younger (13-15) versus older teens (17-18) are more likely to engage in shoplifting. If shoplifting behavior is measured using nominal data (arrested/detained, or not arrested/detained) what type of da= ta analysis should be performed?

10. A researcher designs an experiment to determine wh= ether novice computer users master computer skills faster on a PC versus a Mac. 200 participants with no computer background are recruited for the study.&nb= sp; Half are randomly assigned to learn how to create a document on a PC; the other half perform the same task on a Mac. The average time it takes each participant to master the task is measured. The average times for the two grou= ps are then compared. What statistical test should be used?

Answers: 1. t-test (two group comparison w= ith ratio data), 2. chi-square (one variable, car color, an= d six conditions or colors), 3. chi-square, two groups= with nominal data. 4. one-way ANOVA (three majors, GPAs are scale data, 5= . a simple t-test, 6. 2-Way ANOVA, a 2 X 2 design, 7. chi square, two nominal categories, 8. one-way ANOVA, one variable—type of chile, with five conditions, 9. crosstabs, nominal categories and nominal data, 10. t-test.