Evidence · Advanced · About 6 min
Test causal claims
Did it cause it, or just come before it?
Both sides · Before the roundRead the case: The four-day workweek
Resolution A four-day, 32-hour standard work week should replace the current five-day, 40-hour model
24 volunteer workplaces, covering offices, retail and distribution, joined a six-month comparison.
- Half were assigned by lottery to 32 hours at unchanged weekly pay; half kept 40 hours.
- Both groups maintained an output index of 100.
- Absence averaged three days per worker in the shorter-week group and five in the comparison group; these counts were not adjusted for scheduled days.
- Customer coverage was maintained, but payroll in the shorter-week retail and distribution sites was 8% higher; office payroll was unchanged.
- Healthcare and other continuous-service industries were not included.
For this practice round, the affirmative defends replacing the normal 40-hour model with a 32-hour standard at unchanged weekly pay; it does not claim every shift must be identical. The negative challenges whether the benefits justify the costs of that replacement.
Something happening after a change does not prove the change caused it. Ask how the cause would work, what you are comparing against and what else could explain the result.
Good study design makes one explanation more believable. Randomly assigning groups tells you about cause. Who got into the study in the first place tells you how far the result reaches. You can use causal reasoning to build your case as well as to attack theirs.
Explore an example before trying
Inspect the source
Random allocation makes some starting differences less likely to explain the group difference, though a small trial still has limits.
Use the group comparison as support. Keep clear which scheduling package was tested. These are days per worker, not a rate adjusted for scheduled days or a direct health measure.
Random allocation within a selected trial does not make the workplaces representative of all industries.
Explain how it works
Say how the cause would produce the result.
Check the comparison
Ask what would have happened without the change.
Test the alternative
Name another explanation and what evidence would rule it out.
Your response
Negative · RebuttalChallenge the causal forecast from the employer review without denying the recorded gains. Explain how the challenge affects your case.
Read the case: AI and the future of work
Resolution Advanced AI will on net improve outcomes for workers
a two-year review of 30 employers records 300 retained workers receiving 4% higher inflation-adjusted pay and working 37 rather than 40 hours per week after AI adoption.
- Sixty entry-level roles were removed and 40 technical roles added.
- Of the 60 displaced workers, 18 entered the new roles, 24 found work elsewhere at 8% lower pay, and 18 were still unemployed six months later.
- The other 22 new roles went to different applicants.
- The review does not isolate AI from every other change or forecast the whole economy.
The affirmative argues that productive tools can improve pay and time for many workers. The negative argues that concentrated losses and barriers to new work can outweigh those gains.
You can type or use your keyboard’s microphone to dictate. Your text stays editable. Practice is optional.
Loading your progress…
Help me get started
The review records… It does not separate… That changes… My remaining argument is…
Try it aloud
Use the timer for a spoken attempt if it helps. Practice without recording, then note what you would improve.
Speaking is self-reviewed. No microphone or recording is required.
Allocation and generalization do different work
Resolution A four-day, 32-hour standard work week should replace the current five-day, 40-hour model
24 volunteer workplaces, covering offices, retail and distribution, joined a six-month comparison.
- Half were assigned by lottery to 32 hours at unchanged weekly pay; half kept 40 hours.
- Both groups maintained an output index of 100.
- Absence averaged three days per worker in the shorter-week group and five in the comparison group; these counts were not adjusted for scheduled days.
- Customer coverage was maintained, but payroll in the shorter-week retail and distribution sites was 8% higher; office payroll was unchanged.
- Healthcare and other continuous-service industries were not included.
For this practice round, the affirmative defends replacing the normal 40-hour model with a 32-hour standard at unchanged weekly pay; it does not claim every shift must be identical. The negative challenges whether the benefits justify the costs of that replacement.
- Comparison
Volunteer workplaces were assigned by lottery to the shorter-week package or the 40-hour model.
Assigning workplaces by lottery makes the two groups fair to compare.
- Affirmative · Supported inference
Lower absence in the shorter-week group gives more support to the scheduling package than a before-and-after increase alone would.
That rules out some other explanations for the difference.
- Negative · Remaining limit
Volunteer selection and the missing healthcare sites still matter when extending the result to a general standard.
It still doesn’t tell you whether these volunteer workplaces look like everyone else’s.
Use a study’s strength for what it actually proves.
Test the reasoning
Read the practice case
Resolution A four-day, 32-hour standard work week should replace the current five-day, 40-hour model
24 volunteer workplaces, covering offices, retail and distribution, joined a six-month comparison.
- Half were assigned by lottery to 32 hours at unchanged weekly pay; half kept 40 hours.
- Both groups maintained an output index of 100.
- Absence averaged three days per worker in the shorter-week group and five in the comparison group; these counts were not adjusted for scheduled days.
- Customer coverage was maintained, but payroll in the shorter-week retail and distribution sites was 8% higher; office payroll was unchanged.
- Healthcare and other continuous-service industries were not included.
For this practice round, the affirmative defends replacing the normal 40-hour model with a 32-hour standard at unchanged weekly pay; it does not claim every shift must be identical. The negative challenges whether the benefits justify the costs of that replacement.
Review and continue
Resolution Advanced AI will on net improve outcomes for workers
Read the practice task and case
Challenge the causal forecast from the employer review without denying the recorded gains. Explain how the challenge affects your case.
a two-year review of 30 employers records 300 retained workers receiving 4% higher inflation-adjusted pay and working 37 rather than 40 hours per week after AI adoption.
- Sixty entry-level roles were removed and 40 technical roles added.
- Of the 60 displaced workers, 18 entered the new roles, 24 found work elsewhere at 8% lower pay, and 18 were still unemployed six months later.
- The other 22 new roles went to different applicants.
- The review does not isolate AI from every other change or forecast the whole economy.
The affirmative argues that productive tools can improve pay and time for many workers. The negative argues that concentrated losses and barriers to new work can outweigh those gains.
Your response
You can return to the practice whenever you want.
One model response
The review records higher real pay and shorter hours for retained workers after AI adoption. It does not separate every other change at those employers, so the affirmative should defend the mechanism and reach of its forecast rather than call AI’s economy-wide benefit established. This lowers confidence in the broader prediction; it does not erase the observed gains. Our displacement comparison remains an independent reason to challenge the overall balance.
The observational review does not isolate the tool’s contribution.
Say exactly what the limitation changes about their argument.
The negative can also develop a comparison of outcomes.
The challenge lowers confidence in the cause, and the negative pairs it with a real argument of its own.
Review the performance
- EvidenceI distinguish the observation from its cause.
- ArgumentsI identify what the comparison does not isolate.
- RefutationI explain the challenge’s significance without exaggerating it.
Choose one move to improve, revise your response, then try it aloud.
Common mistakes to avoid
- Treating a possible alternative explanation as proven.
- Mixing up random assignment with a representative sample.
Loading your progress…