Psychology 318Winter 2009
A researcher is interested in the effects of tourism on a population of endangered monkeys. Specifically, she would like to determine whether chronic exposure to humans tends to increase stress levels among the monkeys, as this may have serious effects on their health and reproductive success. She identifies 2 large social groups, one that interacts with tourists on a daily basis (High exposure group) and one that does not (Low exposure group). She randomly samples individual monkeys from each group and tests their stress hormone levels using a fecal assay. She obtains the following data:
Low exposure group (ng/g)High exposure group (ng/g)
49.5163.19
46.6342.88
35.0572.75
66.5761.48
59.7271.89
55.0852.95
84.4276.73
82.9462.69
77.0678.62
46.9463.83
48.8259.97
49.2574.01
36.8353.01
45.2453.17
67.7943.34
63.95
n = 15n = 16
mean = 56.79mean = 62.15
stdev = 15.73stdev = 11.01
1. What statistical test (be specific) should she use to determine whether the groups differ in their stress hormone levels?
independent groups t-test using a pooled variance estimate
What alternative is available that does not require the homogeneity of variance assumption?
Independent groups t test using separate variance estimates – aka the Welch-Satterthwaite t test
2. State the null and alternative hypotheses (in statistical terms) for a non-directional test.
H0: L - H = 0
H1: L - H 0
3. Calculate the appropriate test statistic, assuming homogeneity of variance..
4. What can she conclude (α = .05)?
fail to reject H0
5. Construct a 95% confidence interval.
95% CI:
6. Calculate the effect size.
(small effect)
7. Recalculate the appropriate test statistic for which the homogeneity of variance assumption is not required:
-1.09
A researcher working at the University of Washington observed that Seattle traffic seems to suffer tremendously from snowy conditions. He reasons that these troubles may be due to Seattleites’ poor snow-driving skill. However, he also wonders whether Seattle’s streets present more treacherous conditions because of the hilly landscape and the lack of snowplows. He decides to conduct an experiment to test the hypothesis that Seattleites differ in snow-driving proficiency from people in other, snowier regions. The researcher collects a random sample from each of two groups: people who have lived their entire life in Seattle, and people who have lived and driven a car for at least 10 years in an area that receives at least 40 inches of snow annually. (Seattle receives an average of 11.4 inches of snow per year; Boston, for example, receives 42.2 inches/year.)
The researcher creates a challenging driving course at a ski area parking lot in the Cascade Mountains, and measures the number of “accidents” (cones knocked over) as each participant navigates the course. He obtains the following data.
non-Seattle driversSeattle drivers
25
64
27
04
26
41
34
55
17
69
25
33
6
4
n = 12n = 14
mean = 3mean = 5
variance = 3.65variance = 3.84
1. What statistical test (be specific) should she use to determine whether the groups differ in their driving skill?
independent groups t-test
2. State the null and alternative hypotheses (in statistical terms) for a non-directional test.
H0: N - S = 0
H1: N - S 0
3. Calculate the appropriate test statistic, assuming homogeneity of variance.
4. What can he conclude (α = .01)?
()
fail to reject H0
5. Construct a 99% confidence interval.
95% CI:
6. Calculate the effect size.
(large effect)