# Assessment of Statistics Programs

## BS in Statistics

### Student Learning Outcomes

Desired outcomes for the Undergraduate Degree in Statistics and Data Science include mastery of underlying principles, computational methods, and applications of the following concepts: calculus-based probability, statistical inference, applied linear models, and statistical design. Calculus-based probability is studied in STAT 510, statistical inference is studied in STAT 511, applied linear models is studied in STAT 705, and statistical design is studied in STAT 710, STAT 720, and STAT 722. Statistics and Data Science majors are required to take STAT 510, 511, 705, and one of 710, 720, and 722.

For each of the aforementioned classes, the instructor of the class assesses each undergraduate major of Statistics and Data Science on the following criteria:

1. Calculus-based probability:

1. Knowledge of probability laws and conditional probability.
2. Ability to set up and solve distributional problems that involve calculus.
3. Knowledge of properties of well-known distributions such as normal and binomial.
4. Knowledge of the basic theorems for linear combination of independent random variables.
2. Statistical inference:
1. Knowledge of the role of the Central Limit Theorem in constructing large sample confidence intervals and tests of hypotheses for means and proportions.
2. Use of the t-distribution including assumptions under which its use is appropriate.
3. Frequentist interpretation of confidence intervals, tests of hypothesis, and p-values.
4. Knowledge of methods of estimation including maximum likelihood and method of moments, mean square error.
3. Applied linear models:
1. Ability to set up basic ANOVA models for single and multiple factors.
2. Ability to set up models for simple linear, multiple regression, and variants of the multiple regression model such as polynomial regression in which the parameters are linear.
3. Construction of prediction intervals and confidence intervals for E(Y|X), assess model fit or lack of fit.
4. Ability to carry out analysis using SAS and interpret output in a way non-statisticians would understand.
4. Statistical design:
1. Selection of appropriate experimental designs or sampling plans for a scientific study.
2. Identification of standard study designs
3. Selection of appropriate models and analyses for standard study designs.
4. Ability to report results using tabular and graphical methods as appropriate and explain results in a way non-statisticians would understand.

### Summary of Undergraduate Assessment Activities

In the era of “Big Data,” the demand for people with the skills to collect and analyze data has never been greater. As a consequence, the Statistics and Data Science major has been in a constant process of evolution to meet these demands. Changes to the major include updated requirements and course offerings and the inclusion of the phrase “Data Science” to the major. There are currently a record high of 31 majors and roughly 58 minors. For majors, student learning outcomes currently appear to be satisfactory—neither outstanding nor deficient. However, data is currently insufficient to make stronger claims about the effectiveness of the changes made to our program at this time. Potential improvements that may help monitor the success of our program have emerged through the assessment process. These include: 1) implementation of an exit survey for majors to help measure post-graduation employment outcomes and to help facilitate communication with our alumni and 2) making changes to our data collection and analysis process to improve our ability to assess student outcomes over time given the new emphases in our program.

## Graduate Certificate in Applied Statistics

### Student Learning Outcomes

1. Student will understand the basics of applied statistics.
2. Student will understand the major issues in design of experiments.
3. Student will demonstrate capability with a major software package.

## Master's Degree in Statistics

### Student Learning Outcomes

A student in the Masters Program should demonstrate mastery in the following areas:

1. Communication Skills
1. Public speaking and presentation skills
2. Report-writing skills
3. The ability to use graphical methods to display and interpret information
4. The ability to work and communicate with researchers in other disciplines
2. Computer Skills
1. The ability to use spreadsheets, word processors, and graphics packages
2. Statistical computing skills / ability to use at least one statistical software package
3. Fundamental Statistical Knowledge
1. An understanding of fundamental ideas of statistical theory
2. An understanding of fundamental ideas of linear model theory
3. An understanding of fundamental ideas of study design
4. Applications of Statistics
1. An understanding of practical application of statistics to real problems
2. An understanding of the research process / scientific method
5. Extension of Basic Statistics to Complex Problems
1. The ability to link theory and applications
2. The ability to draw parallels between different kinds of statistical methods
3. The ability to investigate and implement new statistical methods

### Summary of Activities for Assessment of the Masters Degree

The following assessment activities have been completed (as of Spring 2009):

• Student Learning Outcomes (SLOs) and the alignment matrix have been posted on the Department of Statistics website.
• Assessment forms have been created for Statistics 713, 710, 720, 722, 770, 771, and 860.
• Assessment rubrics have been created for Statistics 945 (Consulting Seminar) and the Masters Defense.

Current assessment activities (as of Spring 2009):

• Data is being collected from faculty instructors in Statistics 713, 710, 720, 722, 770, 771, and 860.
• Data is being collected upon a student’s completion of a consulting project for Statistics 945.
• Data is being collected from faculty observers of Masters Defenses.

## PhD in Statistics

### Student Learning Outcomes

1. The ability to carry out research in statistical science.
2. The ability to access and read statistical literature.
3. An understanding of the major modes of statistical inference: frequentist, decision theoretic, Bayesian, and likelihood.
4. Knowledge of statistical procedures based on computation and simulation.
5. The ability to communicate with applied scientists and the statistical community.

### Summary of Progress on PhD Assessment

Student learning outcomes (SLOs) and an alignment matrix have been developed and posted on the Statistic’s Department Website. The alignment matrix lists courses, qualifying exams, and a dissertation as tangible criteria that will be used to assess how the PhD program aligns with SLOs. Preliminary exams, a research update, and a final defense are also listed, as is an exit interview. To supplement evaluation of the PhD program, survey forms for the preliminary exam and final exam have been drafted for the collection of quantitative data from faculty and students for program assessment. The survey form for the preliminary exam has been “piloted” for several cases and some initial data collected. Based on feedback from faculty that have completed this survey, some minor revisions may be made to improve the tracking of assessment criteria in the surveys to departmental and university SLOs.