*Contributing author Nicole Dalzell is an Assistant Teaching Professor of Statistics at Wake Forest University.*

I often hear variations of this statement at the beginning of each semester. Writing is not something students tend to associate with statistics, nor is it something that most stats faculty members have been formally trained to teach. However, the ability to create and critique written communication involving data and statistics is becoming increasingly important. Students who will be using statistics and data in future careers need to be able to communicate their results and processes to get a job. Other students read statements involving statistics on social media or news sources, and then must decide whether that information is correct or reliable. In this post, I will share an activity that can be used to begin to teach statistical writing.

The activity focuses on reproducibility, as this lends itself well to discussion of other key writing skills. The activity is designed to minimize grading for the faculty, and to work well with a variety of courses.

## Teaching Reproducibility

When I introduce reproducibility, I pose the following questions: “Have you written out your process in such a way that someone else can follow your steps and get the same result? How can we make sure our writing achieves this quality, and why does it matter?”

Writing in a reproducible manner requires thinking about order, word choice, and audience, all important writing skills for communicating in statistics. Discussing these topics helps students think critically about not only what they write, but also what they read and hear.

To help students practice and develop these skills, I use a three-stage activity that I call write/feedback/revise. The stages of this activity are outlined below:

– Students create a rough draft explaining step by step how to do something, like create a confidence interval.*Writing Stage*– Each student receives the (anonymous) rough draft another student created. The student follows the given steps and provides feedback.*Feedback Stage*– Students receive feedback on their rough draft from an anonymous classmate and use this feedback to revise their original submission.*Revise Stage*

In each stage of the activity, the student takes on a different role in the writing process: the creator, the user and reviewer, and then the reviser. This means that each stage of the assignment gives students a different perspective on writing, as well as allowing them to interact with the writing of their classmates.

I use a 10-point scale for the entire assignment, and because students provide feedback on each other’s work, grading only requires faculty feedback at the Revise Stage. I detail the stages and grading processes below.

### Writing Stage

The assignment begins with the Writing Stage where each student is given one of two possible tasks. For example, students may be asked to (A) clean data set X or (B) construct a confidence interval for the mean of Y. You can choose any tasks you like that are relevant for your course, but make sure there are two different tasks (Task A and Task B). Both tasks should require roughly the same amount of work from the students.

Let’s say Student A is given Task B: build a confidence interval. To complete the Writing Stage, Student A will build the confidence interval and submit written step by step instructions for how they did so. Typically, I choose tasks whose steps can be well explained in a paragraph. This is important because this rough draft will be reviewed by another student, and a paragraph is a manageable length for revision.

This first stage is just a rough draft so I grade on completion using a 2-point scale.

0 points | Assignment not submitted |

1 point | Assignment submitted but incomplete |

2 points | Assignment submitted and complete |

### Feedback Stage

In the Feedback Stage, feedback on the rough draft is provided by another student, specifically one who did not complete the same task as the student whose rough draft they are reviewing. In other words, a student who completed Task A will be reviewing the draft of a student who completed Task B. This allows students to become the audience for the writing and to reflect on the writing from that perspective. Are the steps clear? Correct?

It is important to make sure that all rough drafts and feedback are anonymous to the students in the class. Anonymous feedback can often be set up in LMS systems, like Canvas. It is also important to provide students with some guidance on what it means to provide constructive feedback. This is a skill in and of itself that students practice with this activity. I generally provide the following guidance:

Your assignment is to provide at least three clear, constructive comments about your classmate’s work by answering the following questions: 1) Were the steps provided in a logical order? If not, provide suggestions for how the order could be changed, and why these changes would be helpful. If so, explain how the order of the steps helped you complete the task. 2) Were the steps clear, correct, and easy to follow? If not, provide specific suggestions for how the steps could be improved, and why these changes would be helpful. If so, explain how the choice of words or writing style helped you complete the task. 3) Provide any other constructive comments on your classmates’ writing. |

In addition to practicing giving feedback and taking on the role of an audience, students in the Feedback Stage are also exposed to the writing style of a classmate. This provides students the opportunity to see examples of how others approached the rough draft, which may help them reflect on their own writing when they revise in the next stage.

As in the Writing Stage, I use a 2-point grading scale for the Feedback Stage. Typically, I have the pairs review each other. This is another reason why anonymity is important in this activity. I will note that I allow students to appeal if they disagree with the score they receive.

0 points | No feedback or feedback provided is not constructive. |

1 point | Limited / Not very helpful feedback. Feedback may be confusing or difficult to follow. |

2 points | Helpful feedback. Feedback is clear and relatively easy to follow. |

### Revise Stage

Students are now given back their original submission from the Writing Stage, along with anonymous comments that a classmate provided in the Feedback Stage. Students are then asked to use this feedback to edit their work.

I grade this final submission on a 6-point scale, to come up with a grand total of 10 points for the assignment.

0 points | Missing or unchanged from initial draft. |

2 points | Incomplete, difficult to follow, or somewhat incorrect. Feedback was not well incorporated. |

4 points | Complete and mostly clear, though some explanations may need clarifying. Feedback was mostly incorporated, as needed. |

6 points | Complete and clear. Steps are easy to follow and clearly justified. Feedback was incorporated, as needed. |

## Summary

This activity is just one way to incorporate statistical writing into your courses. I have found that starting with one activity, or one specific writing skill, is a great place to start. It makes the process a little less daunting and can help you brainstorm other ways to incorporate statistical writing into your courses. Each skill we incorporate helps better equip our students to engage with data and statistics outside of the classroom.

I encourage you to look for small, unexpected opportunities to discuss statistical writing. When I share an article with a class, can I use that opportunity for a brief discussion of word choice and how that impacts the readability of the article? When we complete a problem or an analysis, can we talk about ways the conclusion might be misunderstood and work together to improve our phrasing? When I’m coding with students in class, can we talk about how we would explain the coding steps we have done so far, and practice writing them out? What small step can I take today to help improve the statistical writing ability of my students?

Thank you very much, I really like the idea.

I would like to ask a short question: do you think this be scaled to >100 students? And how many different tasks would you recommend for 100 students?

It can be scaled, yes! Since each student is only reviewing one other student’s work, this scales well. I would recommend 2-3 tasks, primarily because it will make it easier for you to grade if you only have a few unique tasks you need to provide feedback on. I hope this works well for you!!