Guidelines for Assessment and Instruction in Statistics
[Paragraph introducing College GAISE]
Links to previous College GAISE reports:
Recommendations for Statistics and Data Science
Teach statistics and data science as iterative processes of gleaning insights from data to inform evidence-based decisions.
Data summarize the world—indeed, the universe—around us. Data therefore encapsulate information that is essential for understanding what surrounds us and for making decisions. Statistics and data science are, collectively, the science of extracting meaningful information from data to produce insights and analyses. Students learning this science in introductory statistics and data science courses should engage in the process of working with data.
Working with data involves studying the context in which data occur, interrogating and summarizing data, formulating research questions that can be answered with data, using appropriate methodology to analyze data, and effectively communicating results involving data. These steps need not proceed in this order, and analyses or summaries may suggest new questions; statistics and data science, like science itself, are often iterative.
For more information and resources, click here. Emphasize effective written and oral communication of results from data, with attention to the scope and limitations of conclusions.
Producing results is only one step in learning from data; students must also communicate their findings clearly and accurately, both orally and in writing. This requires responsiveness to both the communication genre and the audience background.
Effective communication starts with understanding and conveying the lineage of the data—their source, incompleteness, and currency—emphasizing their effect on conclusions. Inferences, visualizations, and analyses must be presented as evidence, not definitive proof, of conclusions. Conclusions drawn from data can yield a better understanding, and iterations of the process can continue to improve that understanding.
Ethical communication requires transparently acknowledging limitations and contextualizing results at an audience-appropriate technical level—i.e., keeping the focus on conceptual understanding.
Effective communication requires clarity of purpose. For example, this purpose could be to inform, influence, problem-solve, motivate, or facilitate dialogue. To help students craft the purpose of their statistical communication, we offer the mnemonic “genre, audience impact, statistical effectiveness.” There is a spectrum of audience background knowledge as well as one of genre formality level. We suggest a range of sample activities to provide students with experience communicating along both spectra.
Audiences also differ—and can be heterogeneous—in ways other than statistical background knowledge. Their goals and power relations inform decisions such as the degree of detail to communicate and the location of results in the broader story. Students should practice navigating these nuances as they complete communication activities that span the spectra of audience and genre.
For more information and resources, click here. Focus on conceptual understanding rather than algebraic manipulation and formulas.
Given limited instructional time, it is essential to prioritize ideas that promote conceptual understanding over procedural skills. Understanding how and why statistics works is more valuable than the ability to perform manual calculations.
This shift in focus changes how we present topics and assess student learning. Manually calculating statistics like standard deviation or correlation coefficient has little pedagogical value; trained statisticians rely on technology, and so should students. While some basic computation is necessary, complex formulas and distribution tables should be de-emphasized or eliminated. Instead, instructors should introduce only formulas that support deeper conceptual understanding and use technology to automate calculations. Dynamic tools like applets can illustrate statistical concepts effectively, such as how sample size affects standard error, without requiring students to memorize formulas.
Procedural knowledge still has a place—but only when it supports conceptual learning. For example, identifying variable types, exploring relationships, and interpreting results in context reinforce the statistical process better than memorizing formulas. Asking students to manually calculate test statistics and look up p-values may shift focus away from understanding the question and context.
A conceptually strong foundation helps students make sense of statistics in the real world and prepares them for more advanced study. While technical steps matter less in an introductory course, strong conceptual grounding often leads to procedural fluency over time.
For more information and resources, click here. Integrate real data with a context and purpose throughout the course. Select data that are meaningful and engaging to the students.
Real data comes from genuine, practical situations and reflects the complexity of data encountered in everyday life. Integrating real data into a statistics course bridges abstract concepts and real-world applications, making learning more engaging and relevant.
Using data related to topics students care about—like health, sports, or social issues—helps students ask meaningful questions while building statistical skills. Real data supports authentic problem-solving and critical thinking by providing context, such as its origin and purpose, and exposing students to varied variable types and data sources (e.g., surveys, experiments, business metrics).
Working with real data supports both curricular and pedagogical goals. It emphasizes inquiry over rote learning and exposes students to real-world challenges like missing values or non-standard distributions. Realistic data–simulated datasets designed to mimic patterns and challenges of authentic data–can provide controlled yet realistic learning experiences in some situations. Instructors should choose real or realistic data carefully to align with students’ learning needs and instructional goals.
Some potential sources of real datasets include textbooks, online repositories, classroom-generated activities, and collaborations with colleagues in other disciplines.. Educational data portals also offer valuable, curated resources.
Real data might include:
Authenticity: Collected for actual research or decision-making purposes.
Relevance: Tied to topics meaningful and engaging to students.
Complexity: Includes real-world features like outliers or missing data.
Context: Comes with background information to support interpretation and application.
For more information and resources, click here. Encourage multivariable thinking.
Multivariable thinking (MVT) is essential for navigating today’s complex, data-rich world. Real-world phenomena are influenced by many factors, and datasets with numerous variables are now common. MVT is no longer just for statisticians—it is a necessary skill for all students to interpret data critically and draw informed conclusions.
MVT has always had a place in statistics education, but students need consistent, structured practice with it. Rather than treating MVT as an add-on, we should integrate it throughout the course—from data exploration to modeling and inference.
Simple ways to begin introducing MVT include encouraging students to consider the broader context of a dataset, having students explore multivariable datasets, using visualizations with three or more variables, and having students brainstorm confounding variables. Additional ideas for introducing MVT in statistics and data science courses are provided in the link below.
For those already including MVT at a beginning level, the link below also includes suggestions for deepening MVT exposure and suggestions for aspirational activities.
For more information and resources, click here. Incorporate software/apps to explore concepts and work with data.
Technology should play a central role in statistics and data science courses by reducing the algebraic burden of analysis, enabling the creation of data visualizations, and supporting a deeper understanding of data. Effective tools allow students to shift their focus from manual calculations to interpreting results and grasping key concepts. Technology also provides access to real-world, often large-scale datasets that might otherwise be too complex to explore manually. Additionally, appropriate and vetted use of artificial intelligence (AI) can enhance data understanding and simplify complex calculations.
Using statistical software or applets with capabilities beyond those of graphing calculators is essential in modern statistics and data science courses. While no specific tools are mandated, software and applets that support data wrangling, visualization, numerical summaries, inference, and modeling help instructors focus on interpretation. These technologies also enable students to engage with advanced topics and promote equity by reducing barriers linked to algebraic and computational skills.
For more information and resources, click here. Emphasize responsible and ethical conduct in the collection and use of data and in their analysis.
With growing access to real-world datasets and powerful tools, ethics education is more important than ever. Instruction must equip students to responsibly manage data privacy, consent, and transparency in an evolving landscape. Society benefits from informed judgments supported by ethical statistical practice.
Statistical practitioners must commit to making decisions ethically to ensure trustworthy outcomes and protect against harm. The statistical practitioner must understand the provenance of the data—including origins, revisions, and any restrictions on usage—and fitness for use prior to conducting statistical practices. As technology evolves and real-world data becomes more accessible, instruction must address the ethical considerations of data collection, analysis, and communication1.
The American Statistical Association’s Ethical Guidelines for Statistical Practice (2022) emphasize transparency, reproducibility, and valid interpretation. Ethical reflection should be woven into each stage of statistical practice, including formulating questions, data collection and preparation, exploratory data analysis (EDA), analysis, communicating results, and adapting to change.
For more information and resources, click here. Employ evidence-based pedagogies that actively engage students in the learning process.
Research consistently shows that students learn better when they are actively involved in the learning process rather than passively receiving informatione.g. , 2,3,4. Active learning allows students to construct their own understanding, test ideas, and engage in meaningful reflection and dialogue. In statistics and data science, this can be achieved through prediction exercises, real data projects, or the use of manipulatives and technology to explore abstract concepts. Effective instruction centers the student experience, blending brief active strategies with lecture, encouraging both conceptual understanding and deeper engagement. Even small changes—such as think-pair-share, interactive polling, or group reflections—can make a significant difference in learning outcomes.
Successful active learning also requires thoughtful facilitation. Instructors should tailor activities to student backgrounds, build in flexibility for different comfort levels, and foster respectful peer interactions. Incorporating collaborative work, offering students choices in assignments or data sets, and creating space for uncertainty and exploration help students take ownership of their learning. Online courses, too, can support active engagement through tools like discussion boards, video responses, and collaborative tasks. With intentional design and student-centered practices, active learning can enrich statistics and data science education across formats and student populations.
For more information and resources, click here. Use a variety of formative and summative assessments to improve teaching and learning.
Effective teachers use assessment to guide instruction, motivate students, provide feedback, and evaluate performance. Assessments are generally categorized as formative (to support learning) or summative (to evaluate learning). Because the purposes differ, the approach to administration and feedback should also differ. Formative assessments provide ongoing, actionable feedback that helps students adjust their understanding and instructors refine instruction. Summative assessments, on the other hand, are used for evaluation and grading.
Formative assessment should occur during key points in the learning process and provide timely, specific, and non-evaluative feedback. It can take many forms—written comments, class discussions, clicker questions, or think-pair-share activities. These strategies not only support student understanding but also offer instructors insight into where students may be struggling, such as with interpreting p-values. By responding to these insights, instructors can adjust instruction in real time, improving outcomes. Using varied formative strategies also increases student engagement and helps foster a sense of belonging.
For more information and resources, click here. Implement inclusive strategies that create learning environments where all students are valued.
All students deserve the opportunity to succeed in statistics and data science and be prepared for the challenges of our technological society4. Creating these opportunities is both a moral and professional obligation5.
However, many students face barriers that limit their ability to succeed. Inclusive pedagogy can help address these inequities by fostering equal opportunity and a stronger sense of inclusivity6,7. A supportive classroom community influences how students see themselves in the field and affects their academic and career choices8.
Instructors can start promoting inclusivity during course design by considering accessibility needs and anticipating barriers—for example, students with disabilities, with limited internet access, or who are hesitant to speak in class. These efforts should extend throughout all interactions with students. Our table of practices provides actionable strategies to help dismantle systemic barriers4.
Collecting student feedback early in the course is vital. Clearly communicating course design intentions and inviting feedback shows students their experiences are valued9.
Beyond our resources, numerous tools online can support inclusive teaching across all types of courses. Even small steps toward valuing all learners contribute to a more inclusive and statistically literate society.
For more information and resources, click here.