5 Evaluating Evidence and Improving Investigations

Assessing the strength of evidence

Evidence is data or observations relevant to a scientific question and used to support or challenge a claim. Its strength depends on how carefully it was collected, how consistently it was measured, and how directly it addresses the question.

When assessing evidence, consider whether the measurements are relevant and reliable, whether there are enough trials or observations to distinguish a pattern from chance variation, and whether important variables were controlled. Also check whether the tools, units, and procedures were appropriate and clearly recorded. Examine the full set of results, including variation and observations that do not fit the expected pattern.

For example, one test of each paper towel brand provides weak evidence about which absorbs the most water: a spill or an unusual sheet could affect the result. Testing several sheets of each brand with the same amount of water and the same timing makes the comparison more dependable. Repeated measurements do not automatically remove every source of bias, so the method and its limitations still matter.

Recognizing limits and uncertainty

Evidence has Limits: it supports conclusions only within the conditions and range tested. A classroom comparison of a few paper towel brands cannot establish which brand is best in every setting or whether every sheet from a brand performs identically. Scientific evidence can support or challenge explanations, but it rarely proves a broad claim beyond all doubt.

State conclusions no more broadly than the evidence allows. Identify what was not tested and any uncertainty that could affect how confidently the results support the claim.

Considering alternative explanations

A result may have more than one possible cause. Before concluding that a changed outcome was caused by the independent variable, consider Alternative explanations, including uncontrolled variables, measurement errors, differences between samples, and chance variation.

For example, if seedlings given more light grow taller, additional light may have increased growth. But if those seedlings also received more water or were planted in different soil, those differences are alternative explanations. A stronger investigation keeps water, soil, seed type, and other relevant conditions the same while changing the amount of light.

Ask whether each possible explanation fits the data and what new test could distinguish among the explanations. A strong explanation accounts for important evidence, not only results that support it. Investigate unexpected or inconsistent results rather than discarding them without a reason.

Communicating results and conclusions

A useful Scientific report lets other people judge how well a conclusion follows from the evidence and, when possible, repeat the investigation. It should distinguish measured results from interpretations and should not hide data that do not fit the expected pattern.

A clear report includes:

• Question and claim: What was investigated, and what conclusion is supported?

• Method: What was changed, measured, and kept the same? How were observations or measurements collected?

• Results: What data were obtained? Tables and graphs can help reveal patterns and variation.

• Reasoning: How do the results support, challenge, or qualify the claim?

• Limitations and uncertainty: What could have affected the results, and what was not tested?

• Next steps: What further test could help answer an unresolved question?

For example: “In our test, Brand A absorbed more water on average than Brand B across five trials. This supports the claim that Brand A absorbed more under our test conditions. The small number of trials and the single measurement method limit the conclusion.” This is more accurate than saying that Brand A is always the most absorbent.

Refining investigations

Results can show that a method needs improvement or that a new question is worth testing. If measurements vary widely, investigators might standardize the procedure, choose a more suitable measuring tool, or collect more trials. If an alternative explanation remains plausible, they can control the relevant variable or design a comparison that separates possible explanations. If a pattern appears only over part of the tested range, they can collect data at additional values.

For instance, if a paper towel absorbs more water in one trial but less in another, the investigator might find that the pieces were different sizes. A refined method could use equal-sized pieces, the same water volume and soaking time, and repeated trials for each brand. This is more than merely repeating the investigation: the method uses what was learned to improve the evidence.

Scientific investigation is an Iterative process. Evidence informs explanations, critique identifies uncertainties, and further investigations can strengthen, revise, or limit a conclusion. Analyzing data, arguing from evidence, and communicating information work together rather than forming a fixed recipe of steps.