Can the reading pilot support expansion?
Check whether missing outcomes change the case for expansion.
The question
You review a reading pilot in India. The study randomly assigned 400 pupils. One group received an offer of extra reading sessions. The other kept usual support. Compare the groups as assigned.
| Offered sessions | Usual support | |
|---|---|---|
| Pupils assigned | 200 | 200 |
| Reached the reading threshold | 110 | 100 |
Both groups use the same reading threshold. The study owner can verify whether all outcomes are present.
“What does the draft show, and what would you check before recommending expansion?”
Not saved by this site. Your browser may restore text when you navigate back.
A strong answer, step by step
Open each step in turn. This is one strong answer, not the only one.
Step 1: Calculate the draft gap
What would you do here? Open the move you’d make.
Report an apparent five-point gap and check whether the outcome file is complete.
The strong move. Each initial rate uses 200 assigned pupils. The draft gap depends on complete outcomes.
Say the programme raised reading results by 10 percentage points.
Tempting, but it misses something. The gap between 55% and 50% is five percentage points. The outcome file still needs checking.
See how a strong candidate says it
“The draft gives 55% for the offer group and 50% for usual support. That is a gap of five percentage points. I’d check the outcome records behind both counts.”
Why it works: Each initial rate uses 200 assigned pupils. The draft gap depends on complete outcomes.
Step 2: Check comparability
What would you do here? Open the move you’d make.
Ask for the observed and missing counts by assigned group.
The strong move. The assignment records tell us who belongs in each group. The outcome file tells us whose results were measured.
Use the reported rates as final because the study used random assignment.
Tempting, but it misses something. Random assignment still leaves room for missing outcomes at the end of the study.
See how a strong candidate says it
“I’d ask the study owner for the number of outcomes observed and any unmatched records in each assigned group. I’d keep the decision on expansion open until they check.”
Why it works: The assignment records tell us who belongs in each group. The outcome file tells us whose results were measured.
Step 3: Bound the missing results
The study owner confirms all offer-group outcomes. There are 20 missing outcomes in usual support. The draft counted these 20 as below the threshold. There are 100 known successes among 180 observed outcomes in that group.
What would you do here? Open the move you’d make.
Show the range from −5 to +5 percentage points and ask for the missing results.
The strong move. Use all 200 assigned pupils. The missing results can move the comparison in either direction.
Keep the five-point gain because the missing pupils have no recorded pass.
Tempting, but it misses something. The missing scores are unknown. If all 20 met the threshold, the comparison rate would be 60%.
See how a strong candidate says it
“The 100 known successes put usual support at 50% if none of the 20 missing pupils met the threshold, or 60% if all did. With 55% in the offer group, the gap could be −5 to +5 percentage points.”
Why it works: Use all 200 assigned pupils. The missing results can move the comparison in either direction.
Step 4: Advise on expansion
What would you do here? Open the move you’d make.
Keep current support while the study owner checks the outcomes and uncertainty.
The strong move. The missing-data range allows a lower or higher rate in the offer group. Sampling uncertainty needs its own check.
End the pilot because the result could be negative.
Tempting, but it misses something. The same missing outcomes also allow a positive gap. The sponsor needs the verified evidence before changing course.
See how a strong candidate says it
“I’d keep current support and leave expansion open. I’d ask the study owner to recover or explain the missing outcomes and check sampling uncertainty. Then I’d ask the sponsor to review the evidence and delivery capacity.”
Why it works: The missing-data range allows a lower or higher rate in the offer group. Sampling uncertainty needs its own check.
The answer, in short
- Decision
- Keep current support and ask the study owner to verify the outcome file before expansion.
- Because
- The missing comparison outcomes give an offer-minus-usual-support range of −5 to +5 percentage points.
- What would change it
- Verified outcomes and a review of uncertainty and delivery capacity would reopen the decision.
Common traps
- Treating missing scores as failed tests.
- Treating a draft gap as a firm result.
Check your answer
- I used the groups as assigned.
- I checked which outcomes were present.
- I tested both ends of the range.
- I gave the sponsor a next evidence check.
Last reviewed Nov 11, 2026