Evidence is the foundation of research claims. A critical researcher does not merely note what evidence is presented — they evaluate whether it is strong enough to support the conclusions drawn. This is one of the most demanding and highest-value skills in Extended Investigation.
For every piece of evidence offered in support of a finding, ask: “Is this evidence sufficient, reliable and valid to warrant the conclusion drawn from it?”
If the answer is no, the conclusion may still be true — but it is not adequately supported by this particular evidence.
KEY TAKEAWAY: A conclusion is only as strong as the evidence and reasoning that support it. Your job is to assess the warrant — the degree to which the evidence justifies the conclusion. A weak link here undermines the whole argument.
A useful acronym for evaluating evidence offered by sources:
| Letter | Criterion | Question to Ask |
|---|---|---|
| R | Reputation | Does this source/researcher have credibility in this field? |
| A | Ability to see | Did they have direct access to what they claim to have observed? |
| V | Vested interest | Do they have anything to gain from a particular conclusion? |
| E | Expertise | Are they qualified to make this judgement? |
| N | Neutrality | Is there a political, commercial or ideological motive? |
Observational studies can establish correlation — they generally cannot establish causation. Look for:
- Random assignment (experiments) as the gold standard for causal claims
- Confounding variables that might explain the association
- Language that overstates causal claims from correlational data
EXAM TIP: A very common exam question presents a study with a methodological flaw and asks: “How does this weakness affect the conclusion?” Structure your answer: (1) name the flaw, (2) explain its specific effect on validity, (3) state what the researcher should have done.
Even perfectly valid evidence may not support a particular conclusion if it addresses a different question. Ask:
- Is this evidence directly relevant to the claim being made?
- Is there a logical gap between what was measured and what was concluded?
- Does the evidence address the specific population, context or timeframe of the claim?
| Weakness | Description |
|---|---|
| Small sample | Findings may not be generalisable |
| Non-representative sample | Findings may reflect the sample rather than the population |
| Self-report bias | Participants may respond inaccurately due to social desirability |
| Correlation ≠ causation | Association does not establish cause |
| Insufficient controls | Other explanations not ruled out |
| Publication bias | Only positive results published; effect may be overstated |
| Outdated data | Findings may not apply to current context |
APPLICATION: When summarising a source in your Journal, add a “strengths of evidence” and “limitations of evidence” row to your analysis table. This habit ensures you are always thinking evaluatively, not just descriptively.
COMMON MISTAKE: Citing the statistical significance of a finding as proof of its importance. Statistical significance (p < 0.05) means only that the result is unlikely to be due to chance in that sample — it says nothing about effect size, practical importance, or generalisability. Always look for effect sizes alongside significance values.
