+1 (315) 557-6473 

Testing for Evidence of Learning in Statistics

June 25, 2024
Richard Davis
Richard Davis
United States
Statistics
Richard Davis is the Howard Levene Professor of Statistics at Columbia University. He is renowned for his expertise in statistical methodologies and has made significant contributions to the field. His academic qualifications and research work highlight his profound impact on statistics education and research.

As the Howard Levene Professor of Statistics at Columbia University, I have dedicated my career to advancing the field of statistics through both teaching and research. My work spans various domains, from Time Series Analysis and Econometrics to Business Analytics and Biostatistics. In this blog, I will discuss effective strategies for assistance with your statistics assignment, leveraging tools like R, SAS, STATA, and SPSS to enhance the assessment process. Assessments, whether exams or assignments, should provide clear evidence of student learning. In statistics, it is crucial to define learning targets that encapsulate specific goals, including Knowledge Targets, Reasoning Targets, Skill Targets, and Product Targets. For example, students may be required to understand Probability principles, execute Panel Data Analysis, or use software tools like MegaSTAT, MyMathLab, and XLSTAT for various statistical applications. By structuring assessments with clear, measurable outcomes and utilizing appropriate assessment methods such as selected response, written response, performance assessment, and personal communication, educators can ensure a comprehensive evaluation of student learning. This approach not only aids in accurately reflecting student knowledge and skills but also prepares them for practical applications in fields like Psychology, Quantitative Methods, and Operations Research.

What do you want your students to learn?

Testing for Evidence of Learning in Statistics

Assessments, whether exams or assignments, should provide clear evidence of student learning. In statistics, it is crucial to define learning targets that encapsulate specific goals:

  1. Knowledge Targets: These assess factual or procedural knowledge. For example, using Excel or XLSTAT to calculate statistical measures.
  2. Reasoning Targets: These involve solving problems, forming judgments, and drawing conclusions, often seen in Econometrics or Operations Research.
  3. Skill Targets: These test the ability to complete a process, such as performing data analysis using PowerBI or Tableau.
  4. Product Targets: These evaluate the capacity to produce or create something, like developing a predictive model with XLMINER.

Example: You might want students to understand Probability principles or execute Panel Data Analysis. Tasks could include forecasting economic data with MegaSTAT or solving probability problems using MyMathLab.

Write out specific outcomes for your assessment.

Consider the evidence you want from students to prove their understanding of the material. Outcomes should be clear and measurable:

Good Example: Students will conduct a Multivariate Analysis using R and interpret the results.

This example uses specific, actionable verbs and is measurable. Students know they need to perform an analysis and provide an interpretation, which can be evaluated via an exam or assignment.

Bad Example: Students will appreciate the usefulness of statistics in business.

This is vague and unmeasurable. Instead, specify: "Students will utilize Business Analytics to optimize business decisions using Panel Data Analysis."

Starting points for creating measurable outcomes.

Review your course syllabus to identify overarching themes and ideas. Break these down into specific, actionable outcomes. For instance, in a course covering Biostatistics, an outcome might be: "Students will perform hypothesis testing on clinical trial data using JAMOVI."

Determining what type of evidence your outcome will yield.

Categorize measurable outcomes into specific learning target categories:

  • Outcome 1: Students will identify statistical patterns in data using STATA.
    • Learning Target: Knowledge.
    • Why? The verb "identify" suggests recognizing patterns without deeper analysis.
  • Outcome 2: Students will solve complex linear programming problems using GRETL.
    • Learning Target: Reasoning.
    • Why? Solving problems involves using knowledge, process, and understanding.
  • Outcome 3: Students will create interactive dashboards in Tableau.
    • Learning Target: Skill and Product.
    • Why? This requires following a process and producing a tangible outcome.

How are the students going to demonstrate they know what you want them to know?

Choose appropriate assessment methods for your learning outcomes:

  1. Selected Response: Suitable for knowledge and reasoning targets. Use multiple-choice or true/false questions to assess understanding of Probability or Hypothesis Testing.
  2. Written Response: Useful for both short and extended responses. Ask students to explain a statistical concept or describe the methodology for a Time Series Analysis.
  3. Performance Assessment: Ideal for skill and product targets. Have students create a predictive model using Connect Math or perform an Econometrics analysis using MySTATLab.
  4. Personal Communication: Assess through journals or blogs. For instance, students could reflect on their learning process in Operations Research.

Determine how much evidence you require for each outcome.

Blueprints or assessment plans help organize outcomes, learning targets, assessment methods, and evidence required for each learning outcome. Here's an adapted example:

Outcome Learning Target Category Assessment Method Importance (%)
Students will define statisticalterms used in Business Analytics. Knowledge Selected Response 10
Students will compare and contrastmethods in Panel Data Analysis. Reasoning Written Response 25
Students will identify and applystatistical theories using MegaSTAT. Knowledge, Reasoning Selected Response 30
Students will demonstrate datavisualization skills using PowerBI. Performance, Product Performance 20
Students will synthesize resultsfrom Multivariate Analysis using R. Knowledge, Reasoning Written Response 15

By structuring your assessments in this manner, you ensure a fair and balanced exam that accurately reflects student learning across various dimensions of statistics. This approach also prepares students for practical applications in fields like Psychology and Quantitative Methods.

Conclusion

Assessing student learning in statistics requires a thoughtful and structured approach. By defining clear and measurable learning outcomes, selecting appropriate assessment methods, and integrating relevant software tools, educators can create assessments that not only evaluate knowledge but also develop critical thinking, problem-solving, and practical skills. This comprehensive approach to assessment prepares students for real-world applications and enhances their overall learning experience.


Comments
No comments yet be the first one to post a comment!
Post a comment