Define a research question and its measurable outcome

“Better instructions” could mean several different things

• For the materials project, better might mean preferred, quicker to read or more likely to lead to correct sorting. Those are different outcomes. A question using only the word better leaves the team unable to decide what data to collect.
• This original GAC033-aligned lesson fixes one research question for a proposed study. The wider driving question remains how to improve return instructions. The narrower question will concern completely correct paper sorting, not every benefit a real classroom might value.

Name the changed feature precisely

• The independent variable is the feature deliberately changed: label layout A or B. Both use the same words and tray names. B separates each label more clearly and aligns it with its corresponding tray; A uses the existing closer layout.
• Do not change colour, wording, item difficulty and layout together if the claim is specifically about spacing and alignment. If several changes are combined, describe the comparison as a whole design package and avoid crediting one feature alone.

Define what counts as success

• The outcome is whether all four item placements on the paper sheet match a fixed key within two minutes. Each participant contributes one complete success or no complete success. A blank placement prevents a complete success; this rule is written before collection.
• The team may separately record error categories to explain difficulties, but those notes are not a substitute for the main measure. Four-item partial marks and whole-sheet successes have different denominators. Keep the primary count consistent.

{callout:example} “Among 24 willing peers in our class, does layout B yield more completely correct four-item paper sorting sheets than layout A, under the same instructions and a two-minute limit?” {/callout}

Specify people, setting and comparison

• The planned groups are 12 volunteers per condition, assigned by shuffled A/B slips. Every volunteer sees only one layout. This reduces practice effects that could occur if everyone tried A first and B second, but it does not ensure perfectly matched groups.
• State the actual setting and sample. The question concerns this small class task, not every student's behaviour at the real trays. If fewer volunteers participate, record the actual group sizes and compare proportions as well as counts rather than pretending the original plan occurred.

{callout:warn} A directional expectation such as “B may help” is not a result. Do not rewrite the outcome rule after seeing scores to make the preferred layout win. {/callout}

Check feasibility and alternative explanations

• Can the materials and key be prepared, approval obtained and task completed within the project schedule? If not, reduce scope openly. A feasible question is one the team can answer with its resources and permissions, not simply one that sounds impressive.
• Name plausible influences: reading familiarity, assistance from a friend, inconsistent timing or task order. The protocol controls assistance and timing; it records the remaining sample limit. Listing a limitation does not remove it, but it prevents an exaggerated interpretation.

Write the question and analysis together

• Before collecting, state how you will report success: complete successes out of participants in each group, with the same outcome definition. The proposed 12-person groups make the raw counts comparable; actual unequal groups require attention to their denominators.
• Write your own question with the changed feature, measured outcome, people, setting and conditions. Underline each component, then ask a partner what they would measure. If they name a different outcome, revise the wording until the task is clear.

{callout:key} A researchable question fixes the comparison, outcome rule, participants and conditions. The analysis must match that definition, even when the result is disappointing. {/callout}