Deck 9: Review the Evidence

Question 1

Suppose a researcher received responses to a 5-point value inventory (1 = strongly disagree; 2 = disagree; 3 = neutral; 4 = agree; 5 = strongly agree) in which the difference between the mean for women (3.8) and that for men (4.1) was statistically significant. Would that be a substantial (professional) difference?

Answer 1

The response to this item should provide evidence that the student has a grasp of the difference between statistical significance and substantial (educational or professional) significance. The averages provided are 0.3 units apart and are very close to the same-scale score value. One could say with confidence that, while statistically different, both groups basically "agreed" with the item.

Question 2

In a research report on salaries, if the mean value (average) reported was $145,000 and the median (midpoint of the range of values) was $35,000, what conclusion could you reach regarding the sample?

Answer 2

Answers should show knowledge of the median as midpoint and the mean value as average, and responses should also indicate that there must be outliers on the upper end of the scale. This is evidenced by the fact that the mean is higher than the median. Outliers are irrelevant for the median, so the $35,000 figure alone does not say anything about outliers. The $145,000 number by itself also does not say anything about outliers. However, once you compare the two values, you can quickly tell that there is a value in the distribution on the high end that is pulling the value of the mean well above the value of the median.

Question 3

If a data set contains "yes" and "no" answers from 50 men and 50 women, what would be the expected frequencies in the cells of a chi-square test? Why would a researcher use a chi-square test on this data set?

Answer 3

Chi-square tests the assumption that all cells have an equal number; therefore, in this example, the answer is 25 in each cell. The test would tell the researcher if there was a statistical difference from an equal distribution between men and women. The raw data (frequency counts) would need to be examined in order to reveal the direction and magnitude of the difference.

Question 4

What is triangulation of the data, and why is it important?

Question 5

Why is member checking an important tool for the ethnographer?