Your starting point
Prerequisites: Use the ideas and vocabulary from lesson 25; revisit its worked examples if needed.
Definition: A hypothesis is a proposed explanation or expected relationship that can be examined with evidence. An assumption is a condition taken as given in the analysis. A descriptive comparison is not automatically a formal hypothesis test.
State what could challenge your expectation
“Practice helps” is difficult to evaluate without defining practice, outcome, timing and comparison. A useful expectation names an observable pattern and a result that would make you reconsider. Writing it before analysis helps distinguish an initial hypothesis from a story invented after the outcome is known. Unexpected findings can generate new hypotheses, but label them as exploratory.
Make hidden assumptions visible
A comparison may assume consistent measurement, comparable groups, independent records or stable definitions. These conditions are not guaranteed by a correct formula. If one group’s outcomes were measured differently, the difference may reflect the instrument rather than the underlying capability. Record assumptions alongside the calculation and identify which can be checked from the available data.
Use sensitivity rather than certainty
When an assumption is uncertain, examine how the conclusion changes under a plausible alternative. For example, compare a result including and excluding records with an unresolved status, clearly labeling both. This does not eliminate uncertainty, but it reveals whether the decision depends heavily on one choice. AI can help generate alternative explanations; it cannot certify that unobserved assumptions hold.
Two worked examples
A learning expectation
A teacher predicts that learners who complete two retrieval attempts will perform better on a later independent task. Prior knowledge and voluntary participation may differ between groups. The observed comparison can examine the pattern, but it cannot isolate the effect of retrieval unless the design addresses those alternatives.
A reporting assumption
Two teams report completed tasks, but one counts a task when drafted and the other when reviewed. A comparison assumes a shared definition that is false. Aligning the completion rule is more important than calculating the difference to additional decimal places.
Where this becomes useful
A manager wants to test whether a new support script reduces response time. State the expected direction, define the time measurement and document whether ticket complexity changed before attributing an observed difference to the script.
Make an attempt
Use AI Analysis Verification Challenge. Before revealing the notes, write one hypothesis suggested by the evidence and two assumptions needed for a stronger conclusion. Identify a possible observation that would contradict your preferred explanation.
Misconception check
Claim: A hypothesis is useful only if the data confirms it.
Inspect the correction
A clear expectation is also useful when evidence challenges it and helps you revise the explanation.
Independent application
Write a hypothesis for a fictional service change, then list three assumptions and a plausible alternative explanation. Specify a comparison and an observation that would make you revise the hypothesis. Keep exploratory follow-up questions separate from the original plan.
Completion records practice and self-review. It is not certification of statistical competence.
