Write an evidence brief

Write an evidence brief: guided practice

An evidence brief makes it easy to see what is known, what is inferred, and what is still missing.

Use paper or a free local text editor. This is an original fictional practice case; do not submit its records to a real service. Start this lesson with these facts, even if an earlier variation changed them. Original fictional paper P1, “Signs and Sorting”: research question—do picture labels change sorting accuracy at one campus recycling station? Methods—40 volunteers randomly assigned to picture labels or text labels, 20 per group, one supervised session, identical items. Results—picture group 16/20 correct; text group 12/20 correct. Discussion—picture labels may help in this setting; volunteers and one session limit broader conclusions. No uncertainty interval or replication is supplied. This is a teaching example, not a published study. For the final comparison use P2 from Lesson 5: museum, different weeks, 50 per group, picture 40 correct, text 38 correct, different items/staff, no random assignment. A club asks whether picture labels deserve a small local trial.

1. 1. Lead with the decision

   Write a modest recommendation that fits the two fictional reports.

   You should see: A small local trial is worth considering; the reports do not guarantee improvement.

   The requested decision is a local test, not universal adoption.

2. 2. Show the evidence

   Report P1 and P2 with denominators and difference units.

   You should see: P1 16/20 versus 12/20, 20 points; P2 40/50 versus 38/50, 4 points.

   A reader can check calculations instead of relying on adjectives.

3. 3. State uncertainty

   Name different settings and designs and propose a useful next question.

   You should see: P1 uses random assignment at one campus station; P2 compares different museum weeks with different items and staff. Neither report supplies uncertainty intervals or replication. Would the club’s own identical-item comparison produce a repeatable difference?

   The next evidence request should target the main remaining uncertainty.

4. 4. Assemble the brief

   Combine question, design map, results, claim correction, comparison, and a 150–250-word brief; label P1/P2 fictional.

   You should see: Six-part folder assembled: question, design map, results, claim correction, comparison, and this completed brief. A small local trial of picture labels is worth considering, but these two fictional practice reports do not guarantee improvement. In P1, 40 campus volunteers were randomly assigned to picture or text labels for one supervised session using identical items. Correct sorting was 16/20 (80%) with pictures and 12/20 (60%) with text, a 20-percentage-point difference. In P2, museum visitors used the label types in different weeks without random assignment. Correct sorting was 40/50 (80%) with pictures and 38/50 (76%) with text, a 4-percentage-point difference. Both samples favored picture labels, but their effect sizes and designs differ. P2 also changed items and staff, so those differences may help explain its result. P1 concerns volunteers at one station during one session. Neither supplied report includes an uncertainty interval or replication. Sorting accuracy is not the total amount recycled, and these findings do not establish a permanent effect. The club could test whether picture labels improve its own sorting accuracy by randomly assigning participants to label formats with identical items and otherwise consistent conditions, then repeating the comparison. Record correct counts and denominators and reassess before broader adoption.

   Complete work includes an honest uncertainty statement, even when a firm conclusion is unavailable.