Writing a CER that names its limitations
Build a claim-evidence-reasoning argument that uses statistical significance and replication, and honestly states its limitations.
- Claim, evidence, reasoning (CER): You must already know the three parts of a CER before adding statistical significance and limitations.
- Average and spread of data: Judging whether a difference is real depends on comparing group averages and how spread out the data is.
Prerequisites are inferred: pending teacher review.
Re-learn the skill with worked practice and clear examples.
A strong CER adds two things: it shows the difference is statistically significant (unlikely to be due to chance), and it relies on replication (the result repeating) rather than a single run.
Two studies test the same idea. Study X used 6 people and found Drug A beat Drug B. Study Y used 600 people and found the same difference, and it was statistically significant. Which gives stronger evidence, and why?
Reviewed- A.Study X, because smaller studies are easier to trust
- B.Study Y, because a larger sample makes a chance result less likely
- C.They are equally strong because both found the same direction
- D.Study X, because it was done first
Show the worked solution ▾
Answer: B. Study Y, because a larger sample makes a chance result less likely
- Step 1: Compare sample sizes: 6 people can show a difference by luck; 600 people make a lucky fluke far less likely.
- Step 2: Use significance: Study Y's difference was statistically significant, meaning it is unlikely to be chance, so its evidence is stronger.
Why it's right: A larger sample with a statistically significant result is much less likely to be a chance fluke, so Study Y is stronger.
- A: Smaller studies are usually easier to fool by chance, not easier to trust.
- C: Finding the same direction is not enough; sample size and significance change how convincing it is.
- D: Being done first does not make evidence stronger.
Aligned to BI 2.1: significance and sample size · reading level ~grade 9
A team gets an exciting result in one class of 20 students and wants to claim it is true for all teenagers. What is the best next step before making that big claim?
Reviewed- A.Publish the big claim right away
- B.Replicate the study with more and different participants
- C.Delete the data that disagrees
- D.Make the bar graph more colorful
Show the worked solution ▾
Answer: B. Replicate the study with more and different participants
- Step 1: Spot the overreach: One class of 20 is a narrow sample; a claim about all teenagers needs more support.
- Step 2: Choose the fix: Repeating the study with more and different participants (replication) tests whether the result holds.
Why it's right: Replicating with more and varied participants tests whether the one-class result is real before making a broad claim.
- A: Publishing a broad claim from one small class is exactly the overreach to avoid.
- C: Deleting disagreeing data is dishonest and is not replication.
- D: Graph color does not strengthen the evidence.
Aligned to BI 2.1: replication · reading level ~grade 9
- A drug is not approved on one small trial; agencies require larger, replicated, statistically significant results.
Fill these in as you work through the lesson.
- Claim (the answer to the question):
- Statistical significance (a result unlikely to be due to chance):
- Replication (repeating to see if the result holds):
- Limitation (something that weakens or narrows the conclusion):
A complete CER states a , supports it with and , shows the difference is unlikely to be , and names at least one of the study.
- Why does a study with only 4 people give weaker evidence than the same result in 400 people?
- What does it mean to say a difference is 'statistically significant'?
- Name one limitation a one-time, single-school study should admit.
A study finds Drug A lowered blood pressure more than Drug B, and the difference was statistically significant in a trial of 500 people, repeated at three hospitals. Write a one-paragraph CER (claim, evidence, reasoning) and add one honest limitation.