case study

1

CASE STUDY ASSESSMENT RUBRIC (60 points available) Criteria Level of Achievement1

5 4* 3 2** 1 Identification of Scientific Question(s) and Data Exploration

(15 points available)

Report includes the following:

1. Identifies scientific question(s) clearly.

2. Uses appropriate descriptive statistics to display the main features of the case study data.

3. Uses appropriate charts to display the main features of the case study data.

4. Clearly explores the distribution of the target (response) variable in relation to the potential explanatory variables in the data.

(15 points) (12 points)

Report has TWO of the following issues:

1. Scientific question(s) is(are) not clearly identified.

2. Uses only descriptive statistics but no charts.

3. Uses only charts but no descriptive statistics.

4. Did not explore the distribution of the target (response) variable in relation to the potential explanatory variables in the data.

(9 points) (6 points)

Report has THREE of the following issues:

1. Scientific question(s) is(are) not clearly identified.

2. Uses only descriptive statistics but no charts.

3. Uses only charts but no descriptive statistics.

4. Did not explore the distribution of the target (response) variable in relation to the potential explanatory variables in the data.

(3 points)

Estimation of Population Parameters and Testing Research Hypotheses

(20 points available)

Report includes the following:

1. Used the sample data to compute point estimate(s) and construct confidence interval(s) for the parameter(s) of interest.

2. Also conducted hypothesis test(s) to compare the average of the target variable among different population groups.

3. The assumption(s) needed to construct the confidence interval/perform the statistical tests are clearly stated and checked.

4. Made clear comments about the relevance of these estimates/hypothesis tests to answer the scientific question(s).

(20 points) (16 points)

Report has TWO of the following issues:

1. Did not compute point estimate(s) and/or construct confidence interval(s) for the parameter(s) of interest.

2. Did not conduct appropriate hypothesis test(s).

3. The assumption(s) needed to construct the confidence interval/perform the statistical tests are NOT clearly stated and/or checked.

4. Failed to make clear comments about the relevance of these estimates/hypothesis tests to answer the scientific question(s).

(12 points) (8 points)

Report has THREE of the following issues:

1. Did not compute point estimate(s) and/or construct confidence interval(s) for the parameter(s) of interest.

2. Did not conduct appropriate hypothesis test(s)

3. The assumption(s) needed to construct the confidence interval/perform the statistical tests are NOT clearly stated and/or checked.

4. Failed to make clear comments about the relevance of these estimates/hypothesis tests to answer the scientific question(s).

(4 points)

2

Predictive Models (25 points available)

Report includes the following: 1. Studied the correlation between the

response variable and each of the potential explanatory variables (using correlation coefficient and/or scatter plots).

2. Tried different regression models to explain and predict the response variable based on the explanatory variables.

3. Chose the best model using appropriate model selection criteria.

4. Made clear and relevant interpretations of the results of the chosen model (significance of the overall model, significance of explanatory variables, the amount of variation in the response variable explained by the model, interpretation of regression coefficients).

5. Used appropriate plots to check the assumptions for regression analysis and commented on the reliability of the regression results/predictions.

6. Used the regression results to answer the scientific question (made necessary predictions).

(25 points)

(20 points)

Report has TWO or THREE of the following issues:*** 1. The correlation between the

response variable and some potential explanatory variables is explored (uses only correlation coefficient or only scatter plots).

2. Fitted one single regression model (instead of trying several models) to explain and predict the response variable based on the explanatory variables.

3. Wrong choice for the best model (or didn’t use appropriate model selection criteria).

4. Made unclear and irrelevant interpretations of the results of the chosen model or didn’t interpret some of the regression results (significance of the overall model, significance of explanatory variables, the amount of variation in the response variable explained by the model, interpretation of regression coefficients).

5. Failed to use appropriate plots to check the assumptions for regression analysis and/or didn’t comment on the reliability of the regression results/predictions.

6. The regression results are not utilized to answer the scientific question (or didn’t make necessary predictions).

(15 points)

(10 points)

Report has FOUR or FIVE of the following issues:*** 1. The correlation between the

response variable and some potential explanatory variables is explored (uses only correlation coefficient or only scatter plots).

2. Fitted one single regression model (instead of trying several models) to explain and predict the response variable based on the explanatory variables.

3. Wrong choice for the best model (or didn’t use appropriate model selection criteria).

4. Made unclear and irrelevant interpretations of the results of the chosen model or didn’t interpret some of the regression results (significance of the overall model, significance of explanatory variables, the amount of variation in the response variable explained by the model, interpretation of regression coefficients).

5. Failed to use appropriate plots to check the assumptions for regression analysis and/or didn’t comment on the reliability of the regression results/predictions.

6. The regression results are not utilized to answer the scientific question (or didn’t make necessary predictions).

(5 points)

4* Exhibits some characteristics of “5” and some of “3” 2** Exhibits some characteristics that fall somewhere between “3” and “1” *** TWO or THREE (FOUR or FIVE) issues is determined based on the seriousness of the issue