How to Write Claim, Evidence, Reasoning (CER)

CER stands for claim, evidence, reasoning. The table shows what each part answers and the most common way a draft goes wrong, so a teacher or a student can check a response quickly. The sections after it cover each part in detail, with sentence starters, worked examples, and a before-and-after revision.

The claim answers the question or states the conclusion. Evidence is the relevant information used to support that conclusion, such as observations, measurements, patterns, results, or source-based facts allowed by the task. Reasoning is the justification that explains why the evidence supports the claim through scientific ideas, principles, mechanisms, or relationships.

These categories are connected but not interchangeable. 'The plant near the lamp grew taller' may be evidence if it reports an observation. 'The lamp caused more growth' is a claim about the relationship. Reasoning would have to explain why the measured pattern supports that conclusion and address whether the investigation actually controlled other important conditions. Labeling the jobs separately makes those distinctions easier to inspect.

A claim should be clear enough to evaluate. If the question asks which material is the best insulator in an investigation, the claim identifies the material supported by the results. It does not need to repeat every measurement or explain the mechanism yet. When the question asks for a relationship, the claim should state that relationship at the level the evidence can support.

Treat the claim as revisable. Scientific explanations are not stronger because a student refuses to change an early answer. If the data do not support the first conclusion, update the claim. This makes CER consistent with evidence-based reasoning rather than turning the claim box into a guess that must later be defended at all costs.

Evidence should bear directly on the claim. In an investigation, that may mean measurements, repeated observations, comparison groups, or a pattern across trials. In a source-based science task, it may come from a graph, model, text, or data table. Preserve units and conditions when they change the meaning of the result. A number without context is not automatically useful evidence.

Do not list every observation because it was collected. Ask which results discriminate among possible claims. NGSS materials repeatedly emphasize valid and reliable evidence and, at higher grades, multiple sources or lines of evidence when appropriate. Sufficiency depends on the task: a tightly scoped classroom question may sometimes be answered with one strong observation, but repeated observations are stronger when variability matters, and broader explanations may require several independent supports.

Reasoning is usually the hardest part because it cannot be copied directly from the data table. It identifies the scientific idea that makes the evidence meaningful. If an investigation shows ice melting faster on one surface, the reasoning needs the relevant thermal-energy relationship; repeating the melting times is still evidence, not reasoning. NSTA's CER guidance similarly distinguishes data from the justification that links data to a claim.

A useful prompt is: 'What science idea explains why this result supports my conclusion?' Another is: 'If a reader accepted the evidence but did not yet accept the claim, what relationship would I need to explain?' Those questions move students away from empty phrases such as 'This proves my claim' and toward an actual explanatory bridge.

Many students know what a claim is and still freeze at the first word. These starters give them a way in. Use them for the first few responses, then ask students to set the starter aside and write in their own words. Otherwise every paper in the class begins with the same phrase.

Suppose a class tests how much water equal-sized pieces of two unlabeled paper-towel samples absorb, using the same water amount, contact time, draining method, and measurement procedure for both samples. Across three matched trials, Sample B absorbs more water each time. Claim: under the tested conditions, Sample B absorbed more water than Sample A. Evidence: the recorded amount for B is higher in all three matched trials, and the average for B is also higher. Those are observations from the investigation rather than opinions about which brand seems better.

Reasoning: using equal-sized samples while holding the water amount, contact time, draining method, and measurement procedure consistent makes the absorption results more comparable, and the repeated pattern across matched trials supports the conclusion that B performed better on this measured property under those conditions. Notice the limits. The CER does not justify a broader claim that B is 'the best paper towel' overall, because the investigation did not test cost, strength, different liquids, or every competing product. Good reasoning also keeps the conclusion proportional to what was tested.

Use a classroom investigation with the same mass of sugar and water, but two water temperatures. In this illustrative data set, the sugar dissolves in 24 seconds at 50°C and 78 seconds at 20°C, with both cups left unstirred. Claim: the sugar dissolved faster in the warmer water under these test conditions. Evidence: the recorded dissolving time was 24 seconds in the 50°C water and 78 seconds in the 20°C water. Reasoning: warmer water has greater average molecular kinetic energy, so water molecules move faster and interact with the sugar more rapidly; that scientific idea explains why the measured time was shorter. The evidence supports this comparison, but one trial does not establish the exact result for every sugar grain, water volume, or classroom setup.

Compare this with the paper-towel example: the claim answers the investigation question, the evidence reports measured results with their conditions, and the reasoning supplies the scientific relationship that connects those results to the claim. The topic changes; the three jobs do not.

Question: which melts an ice cube faster, a metal pan or a wooden board? In this illustrative test, two ice cubes of the same size start on surfaces that have been sitting in the same room at about 22°C (72°F). The cube on the metal pan melts completely in 6 minutes. The cube on the wooden board takes 14 minutes.

Before reading the revision, look for three problems in the first draft. The claim describes the setup without answering the question. The evidence box holds an explanation instead of a measurement. The reasoning repeats the result and adds the word “proves.” Then compare each cell with the revised column.

CER works in any class where an answer needs support. In English, the evidence is a detail or quotation from the text, and the reasoning explains what that detail shows about the idea in the claim. In social studies, the evidence is a fact from a source, and the reasoning links it to the historical idea. The organizer's last box is labeled for a science idea, but students can use it for the text-based or historical idea instead.

The table answers two questions. English: is Charlotte a loyal friend to Wilbur in Charlotte's Web? Social studies: why did colonists dump tea into Boston Harbor in 1773?

Some English and social studies classes teach claim, evidence, commentary (CEC) instead of CER. The structure is the same: the claim answers the question, the evidence is the quotation or fact, and the commentary does the job the reasoning does in science by explaining how and why that evidence supports the claim. Commentary is where the analysis lives, so a sentence that only repeats the quotation in other words is not commentary yet.

CERER is the same routine with two supports: claim, evidence, reasoning, evidence, reasoning. It matches the two evidence-reasoning rows on the organizer, and you can write it as one paragraph. Whichever label your teacher uses, check each part with the same question: does this sentence do its own job, or is it repeating another part? In the organizer's last box, write the science idea, or in English or social studies write the idea the commentary is building toward.

Score each part 0, 1, or 2: 0 means missing or unsupported, 1 means present but incomplete, and 2 means clear and supported. Claim: does it answer the actual question at a scope the investigation can support? Evidence: does it include relevant observations or data, with units or conditions when those matter? Reasoning: does it explain the scientific idea that links the evidence to the claim instead of repeating the result? A response with a low score in one part shows what to revise; it does not need extra evidence merely to reach a fixed sentence count.

For a first elementary practice, a teacher may accept one clear observation and an oral or sentence-stem explanation. In later grades, the same check can require repeated measurements, attention to variability, a mechanism, or a limit on the conclusion. Match the expected evidence and reasoning to the question and available data rather than applying one rubric as a universal standard.

The printable provides room for two pairs because many explanations benefit from more than one support. Pair one might describe the main quantitative pattern; pair two might identify a repeated observation, comparison, or second data source. Each reasoning box should explain the significance of that particular evidence instead of copying one generic sentence twice.

If the second piece adds nothing, do not force it. A requirement such as 'always use two pieces of evidence' can lead students to duplicate weak information. Conversely, a complex claim may need more than the worksheet provides. The organizer is a planning surface, not a rule about how much evidence science permits.

Real data are not always perfectly tidy. If one trial differs from the others, note it instead of silently discarding it. Check for measurement error, procedure differences, variability, or another explanation that the task allows you to investigate. The presence of an exception does not automatically destroy a claim, but the reasoning should not pretend contradictory evidence does not exist.

A stronger revision may narrow the claim: 'Sample B absorbed more water in two of three trials' is more accurate than a sweeping conclusion when the pattern is inconsistent. At advanced levels, students can discuss reliability, uncertainty, and competing explanations. The core habit stays the same: the claim and reasoning should respond to the evidence actually available.

Mistake one is writing a claim that simply repeats the question. Make a definite answer. Mistake two is putting an explanation in the evidence box. Ask whether the sentence reports what was observed or measured; if not, it may belong in reasoning. Mistake three is using every datum. Select information that directly bears on the claim. Mistake four is treating correlation as automatic causation when the investigation does not justify it.

Mistake five is repeating evidence in the reasoning section. Add the scientific principle or mechanism that explains the connection. Mistake six is keeping a claim that the results do not support. Revise it. Mistake seven is writing a conclusion broader than the study. Match the wording to the variables, samples, conditions, and evidence that were actually examined.

After planning, combine claim, evidence, and reasoning into connected prose. A short response may use one paragraph; a longer explanation may revisit the reasoning after several pieces of evidence. The reader should still be able to identify the three functions, but the final writing does not need to announce 'Claim:' and 'Evidence:' unless the teacher or assessment specifically requests those labels.

For younger learners, oral CER or color coding can make the distinctions visible before extended writing. For older students, add source evaluation, uncertainty, alternative explanations, and more demanding reasoning. The NGSS progression increases expectations for evidence and scientific reasoning across grade bands, so the scaffold should become more sophisticated as students gain knowledge rather than remain a fixed fill-in-the-blank routine.