English English Summary Assignment

1

Summary of “[Article Title goes here]”

Student Name

NorQuest College

ENGL 2510: Scientific and Technical Writing

Brooklin Schneider

Assignment 1: Summary

January 7, 2026

2

Step 2: Task Map

3

Step 3: Initial Prompt and Output

4

Step 3: Revised Prompt and Output

Output from GPT-4o

Based on the material provided from "Validity matters more than cheating" by Dawson (2022),

the article argues that the focus within educational assessments should shift from solely

addressing academic cheating to ensuring validity in assessment practices. Dawson highlights

that while academic integrity is often framed around the values of honesty, trust, and

responsibility (Fishman, 2014), the primary concern should be whether assessments accurately

measure student capabilities. The article suggests that tackling cheating is a part of enhancing

assessment validity but warns that some anti-cheating measures may inadvertently reduce the

overall validity by excluding or disadvantaging certain student groups. The paper stresses

employing approaches that ensure graduates are effectively demonstrating the skills and

knowledge they are claimed to possess, rather than just focusing on the absence of dishonest

5

behaviors. Dawson proposes that assessment validity encompasses creating inclusive practices

that enable all students to truly reflect their competencies. In their view, cheating should be seen

as a subset issue within the broader goal of achieving valid assessments, rather than as the

primary focus. This reorientation towards validity over cheating ensures a more comprehensive

and positive educational mission, providing a more accurate assurance of learning outcomes.

"Cheating is addressed without moralising, as part of the broader positive mission of assurance

of learning" (Dawson, 2022). This aligns with the article's broader call to prioritize validity over

moralistic approaches to academic misconduct.

Reference: Dawson, P. (2022). Validity matters more than cheating. Assessment & Evaluation in

Higher Education. doi:10.1080/02602938.2024.2329167

Step 4: My Commentary

When I first asked the AI to “summarize the article,” the results were pretty shallow and

generic across Claude Sonnet and ChatGPT-4o. They sounded like they could fit lots of articles,

not this specific one. Claude was somewhat helpful because it at least gave me a draft I could

work from. On my second try, I used a much more detailed prompt that matched how I actually

summarize, like focusing on the main claim, the key sections, and what the authors want readers

to take away. That version was better, and it felt like the AI followed my steps more closely. To

double-check, I asked it to list what points it included and what it left out, and I also asked it to

explain what “validity” means in Dawson et al. (2024). When I asked what it assumed about

summarizing, it said things like headings and structure help identify key points, the title frames

the argument, and summaries should use an academic tone, which matches my own task pattern.

One problem was citations. Claude gave accurate APA citations, but GPT-4o made some up.

That surprised me because citation formatting feels rule-based, so I expected it to get that right.

6

Step 5: Summary of “Validity Matters More Than Cheating”

In the research article, “Validity Matters More Than Cheating,” Dawson et al. (2024)

argue that higher education has made cheating the headline assessment problem, but this focus is

too narrow to guide good decisions about assessment. They show that cheating is defined

inconsistently across policies and scholarship, so disputes about what counts can obscure the

more useful question of whether an assessment produces evidence of student learning. They also

explain how talk of cheating can invite moralizing about students and legitimizes surveillance

tools like Turnitin, even though the absence of cheating is not the same as evidence of academic

integrity. To reframe the issue, they propose shifting attention from rule-breaking to validity,

treating cheating as only one possible threat among many that can undermine the credibility of

assessments.

The authors define validity as assessing what we intend to assess, then broaden this to

include different forms of validity evidence, especially consequences. They argue that validity is

not only a property of an assessment instrument but of the whole assessment occasion, including

the task, conditions, student responses, and the decisions and actions that follow from

assessment. This lens makes trade-offs visible: strict restrictions might appear to increase

security, but they can also discourage peer learning, feedback seeking, and other practices that

matter for learning, equity, and the quality of evidence produced. They also note that many anti-

cheating approaches are applied broadly even though most students do not engage in the most

egregious forms of cheating, raising the question of whether blanket controls are worth the

benefits.

Applied to generative AI, Dawson et al. (2024) argue that the key question is not whether

AI use is inherently cheating, but whether AI use changes how educators assess learning. They

7

write, “The use of artificial intelligence in assessment is not ‘cheating’, it is a condition to be

attended to alongside other validity matters” (Dawson et al., 2024, p. 1012). If an assessment

depends on students not using AI but cannot realistically prevent its use, they suggest it is not

suitable for high-stakes assurance of learning, and the real challenge is assessment design rather

than student morality. Overall, they argue that recentering validity helps educators weigh pros

and cons more holistically and shift educators away from rule adherence toward what assessment

is supposed to support: confidence that graduates can do what institutions claim they can do.

8

References

Anthropic. (2025). Claude Sonnet (July 7 version) [Large Language Model].

https://claude.ai/new

Dawson, P., Bearman, M., Dollinger, M., & Boud, D. (2024). Validity matters more than

cheating. Assessment & Evaluation in Higher Education, 49(7), 1005–1016.

https://doi.org/10.1080/02602938.2024.2386662

OpenAI. (2025). ChatGPT (July 7 version) [Large Language Model].

https://chat.openai.com/chat