Part 1 Exam Questions (Answers are highlighted)

1. (Points: 5) Which of the variables below is most related to Weight Change (adjusted for baseline weight)? Use Table 1.

a. Months at current weight

b. Self motivation

c. Self esteem

d. Exercise perceived barriers

2. (Points: 5) Which of the variables below is most statistically significantly related to Weight Change (adjusted for baseline weight)? Use Table 1.

a. Acceptable weight loss

b. Impact of weight on quality of life

c. Self motivation

d. Exercise self-efficacy

3. (Points: 5) According to the slopes given in Table 2, which variable of those listed below is negatively related to Change in Weight?

a. Number of diets in past year

b. Weight outcome evaluations

c. Self motivation

d. Body size dissatisfaction

4. (Points: 5) According to Figure 2, is the R2 change for x2 statistically significant?

a. Yes

b. No

c. Can not be determined from this Figure

5. (Points: 5) In Figure 3, an ANOVA was used to determine whether differences existed between the three success groups for the number of diets per year and for self-motivation. If the authors had divided the sample into 2 groups instead of 3, what statistical test would have been appropriate?

a. ANOVA

b. Dependent t-test

c. Post hoc test

d. Independent t-test

6. (Points: 5) Which of the following is correct for self motivation? Use Figure 3.

a. The least successful group had a significantly higher self motivation than the Intermediate group.

b. The Most successful group had a significantly higher self-motivation than BOTH the Intermediate and the Least successful groups.

c. The Least successful group had a self-motivation that was significantly lower than BOTH the Intermediate and the Most successful groups.

d. The Most successful group had a significantly higher self-motivation than the Least successful group but not the Intermediate group.

7. (Points: 5) What type of statistical test was used to determine whether groups differed in the analyses for Figure 3? Refer also to the Methods section.

a. An ANOVA followed by a post hoc test (Tukeys Honestly Significant Difference Test)

b. An independent t-test

c. A mupltiple regression to examine multivariate relationships

d. Last Observation Carried Forward method

8. (Points: 5) Which variable below has a higher level of variability? Use Table 1.

a. Acceptable weight loss

b. Happy weight loss

c. Both variables have the same variability

This is Part 2 of the Final Exam

1. The purpose of this question is to determine your understanding of measures of central tendency and variability.

a. Table 1 in this paper shows descriptive information for several baseline variables. You are given a measure of central tendency and two measures of variability.

b. From Table 1 of this paper, please describe one of the variables included in this table using your knowledge of central tendency and variability. The maximum word count for this question is 125 words.

The variable self-esteem as a pretreatment variable for weight change had a mean of 32.4 and a standard deviation of 3.77. The mean score indicates the statistical average of the self-esteem scores of the respondents in the study. It would mean that on average the score is concentrated in the 30 to 40 score. This is actually supported by the measure of variability which is the standard deviation score, for self-esteem the corresponding SD is 3.77, the lower the SD the less varied or dispersed the scores are. This can also be verified by minimum and maximum score which is 24 and 40 respectively.

2. The purpose of this question is to assess your understanding of correlations as a measure of association. I would like to see if you can identify variables that are “important” in this study as well as those that are not “important”. I would also like to see that you can interpret the number correctly AND do so within the context of what variables were measured in the study.

a. Table 1 presents data for the relationship between weight change over 4 months and several baseline variables that were measured. The type of correlation is the Spearman correlation (which we never specifically talked about but see number 2 under “Important Article Notes” above). This is interpreted in the same manner as the r we’ve talked about; this correlation is abbreviated with a funny looking p (the 4th column of numbers). The 5th column of numbers is labeled p which is the p-value (significance).

b. Pick some of the correlations you find interesting and important and talk about what the data are telling you. You should be looking at the correlation and the p-value. You are also given many variables to look at (all of which are correlated with weight change) which allows you to do some comparing. The maximum word count for this question is 125 words.

Body size dissatisfaction refers to the concept in which a person is dissatisfied with one’s weight and physical shape and size, thus it could be assumed that if a person is dissatisfied with one’s body then he/she will likely take actions to change it. Correlation tests the association of two variables, in this study body size dissatisfaction was correlated to weight change over a 4 month period. The Spearman rank correlation was 0.18 and the p value was 0.038, the result indicate that there is little association or relationship between body size dissatisfaction and weight change, the p value or significance value is also greater than 0.005 indicating also that there is no significant relationship between the two variables.

3. The purpose of this question is to determine your ability to interpret data from a regression analysis.

a. In the analysis presented in Table 2, several variables (in the body of the table) were entered in an analysis to see how they, all together, related to weight change. The “squared semi-partial correlation” is the unique contribution of each variable to the weight change variable. You are given the B (slope), the p-value, and the r-squared (which is the semi-partial correlation) for each variable. Remember that B is not a correlation! In the footnote, you are given the r-squared for the entire equation and the SEE for the equation. I DO NOT expect you to talk about the t in the table or the F or adjusted r-squared in the footnote.

b. Please describe what this table is telling us above what Table 1 is telling us. Notice that the variables are the same in Table 2 as are in Table 1, with the exception of Baseline Weight in Table 2. (The inclusion of baseline weight in Table 2 should NOT be the focus of your answer because it was added mostly as a covariate, something that may affect how the other variables are related to weight change.) To answer this question, it may be useful to look at what regression gives us that correlation does not. Also, consider why the authors included the variables that they did. The maximum word count for this question is 150 words.

Table 1 presented the correlation values for the measured variables and weight change, which only told us how each variable, was associated with weight change. Table 2 on the other hand presented the regression analysis scores of each variable in terms of how much they affected weight change. Table 2 variables included number of diets in past year, weight outcome evaluations, self-motivation and body size dissatisfaction which was ran through a regression analysis using the forced model. The variables wee forced into the regression analysis because the said variables had a higher probability to have a significant relationship to weight change. The regression analysis provided a deeper analysis into how the almost significant variables contributed to weight change which the correlation analysis did not verify.

4. The purpose of this question is to determine your ability to interpret methods to test for mean differences.

a. In Figure 3, the sample was divided into tertiles (three equal groups) based on weight loss success. Two variables were presented: number of previous diets and self-motivation.

b. Describe the main findings being presented in Figure 3. What can you conclude based on these results? To answer this correctly, you must be able to interpret the p-value presented for each variable and the way the authors identified specific group differences between the three groups. Specifically, which groups are different from which? The maximum word count for this question is 100 words.

F score for number of previous diets was 4.77 and p value was 0.010 at 95th confidence interval, the figure tells us that there is not much difference between the most successful group and the intermediate group since the number of previous diets for both group was less than or equal to 1. For self-motivation, there is also significant difference between the most successful group and the least successful group but not the intermediate group. The F score is 4.03 and the p value is at 0.020 which is still less than the 0.05 level of significance. The figure also indicates that the most successful group was highly motivated as compared to the least successful group.

5. The authors displayed an inference when they discussed the results relative to the “Goals and Relative Weights Questionnaires and subsequent weight loss” (fifth paragraph of Discussion section). Briefly discuss what inferences were made by the authors. In answering this question, I would like you to show me that you understand inferential statistics. The maximum word count for this question is 75 words.

The authors inferred that when a large population is considered, it would seem likely that the expectation-outcome relationship is curvilinear. This would mean that the higher the expectation or the more stringent the beliefs are the smaller the outcome would be while the lower the expectation the greater the outcome would be. They also inferred that American women were likely to have more stringent expectations while Portuguese women had conservative expectations.

6. A Type II error is made when we conclude that there is no relationship (or mean difference) when there actually is a relationship (or a difference). Find and briefly describe one example in this paper where the authors risk making a Type II error. (Remember that we don’t know if an error was made until someone else replicates the study and corroborates (or doesn’t) the findings. So, I’m not saying that the authors made a Type II error, I’m saying they risk making one.) The maximum word count for this question is 75 words.

Good Luck!!

The researchers reported that self-esteem had no significant relationship to weight change as indicated by the zero correlation coefficient in Table 1. The authors risk committing Type II error in this case because we know that self-esteem is actually tied to one’s personal evaluation of one’s self. People who have high self-esteem usually are not dissatisfied with their bodies and therefore do not need to lose weight which can be an inverse relationship, not the absence of a relationship.