Intermediateexam strategy

Statistics for Biology Exams: Means, SD, and Significance

Learning objectives

  • Compute and interpret means, medians, and standard deviation.
  • Read error bars and significance markers on exam figures.
  • Understand what t-tests and p-values tell you — and what they don't.
  • Write statistically literate conclusions for data questions.

Descriptive Statistics: Mean, Median, Spread

Exams expect fluency with the basics. The mean is the sum divided by n — sensitive to outliers, which is exactly why questions sometimes ask for the median instead (the middle value, robust against extremes). Range is max minus min; standard deviation quantifies typical spread around the mean. You rarely compute SD by hand in exams, but you must interpret it: a large SD means variable data, overlapping SD bars between groups suggest no meaningful difference. When asked which average to use, the reasoning is the mark: 'median, because the outlier skews the mean.' Practice spotting skewed datasets and justifying the choice — the justification, not the number, is usually what's scored. For every practice dataset, compute mean and median and decide which fits; the decision practice matters more than arithmetic.

Error Bars: SD, SE, and Confidence Intervals

Error bars come in flavors, and questions test whether you know what the bars represent. Standard deviation bars show data spread; standard error bars show precision of the mean estimate; 95% confidence intervals show the range likely containing the true mean. The practical exam rule: non-overlapping error bars suggest a real difference between groups, while substantially overlapping bars mean you cannot claim one. A subtle trap: SE bars are narrower than SD bars for the same data — conclusions must respect which is plotted. When a figure legend specifies the bar type, read it; when it doesn't, state your assumption. Either way, the interpretation sentence should reference the bars explicitly: 'the error bars overlap substantially, so no significant difference can be claimed.'

The overlap rule

Non-overlapping error bars → likely real difference. Substantially overlapping → cannot claim a difference. Check the legend for whether bars are SD, SE, or CI.

Tests and P-Values: What They Actually Say

School biology mentions t-tests (comparing two means) and chi-square (counts vs expectation); exams test interpretation more than computation. The p-value is the probability of seeing results this extreme by chance if the null hypothesis were true — it is not the probability the hypothesis is true, a misreading examiners love to probe. The convention: p < 0.05 counts as statistically significant. A significant result means the difference is unlikely to be chance — it does not measure how large or important the difference is. 'Statistical significance' and 'biological importance' are different claims; strong answers distinguish them. When a question provides a p-value, your conclusion must cite it: 'p = 0.03, below 0.05, so the difference is significant.'

Writing Statistically Literate Conclusions

Data conclusions follow a template that examiners reward: (1) describe the observed difference with numbers and units; (2) state the statistical evidence (error-bar overlap or p-value vs 0.05); (3) give the verdict (significant or not); (4) connect to the biological hypothesis cautiously. Example: 'Group A grew 12 mm vs 8 mm in group B; error bars do not overlap and p = 0.02 (< 0.05), so the difference is significant, consistent with the hypothesis that light increases growth — though causation requires the controlled design to rule out confounders.' Note the caution in step 4: significance supports the pattern, while the experimental design determines what the pattern proves. Conclusions that skip step 2 are opinions; conclusions with all four steps are science.

Conclusion template

Numbers → statistical evidence (bars/p-value) → significance verdict → cautious biological interpretation. All four steps, every data conclusion.

Key points

  • Choose mean vs median deliberately; justify with outlier reasoning when asked.
  • Read the error-bar legend (SD, SE, or CI) and apply the overlap rule before claiming differences.
  • p < 0.05 means significant; the p-value is not the probability the hypothesis is true.
  • Significance measures unlikeliness-by-chance, not size or biological importance.
  • Write conclusions as numbers → statistical evidence → verdict → cautious interpretation.

Common misconceptions

✗ Myth: A significant result means the effect is large or important.

✓ Truth: Significance only addresses whether chance explains the result. A tiny, trivial difference can be significant with large samples.

✗ Myth: p = 0.06 means there's no effect.

✓ Truth: It means the evidence didn't cross the conventional threshold. 'Not significant' is not proof of no difference — it's inconclusive.

Frequently asked questions

Do I need to calculate standard deviation by hand?

Rarely — exams focus on interpreting given SD values and error bars. Know the formula's logic (average squared deviation) well enough to explain what a large or small SD means.

When is a t-test used vs chi-square?

t-test compares means of continuous measurements between two groups; chi-square compares observed counts against expected frequencies. The data type — measurements vs counts — decides.

What if error bars just barely touch?

Treat borderline overlap as inconclusive and say so: 'the bars overlap slightly, so no firm conclusion about a difference can be drawn.' Examiners reward this honesty over forced claims.