Choosing a statistical test

Laura Armstrong & Joe Wolfensohn

Teachers

Laura Armstrong Joe Wolfensohn

Recall Questions

This topic requires some prior knowledge of statistics in Biology and quantitative investigations of variation from year 12. You can test your knowledge on these below.

Why do scientists use statistical tests in biology?

To determine whether the differences/ correlations observed are significant or due to chance.

What statistical tests are on your specification?

Student T-test, Chi Squared and Correlation Co-efficient (Spearman’s rank)

Will you ever be asked to recall the equations for these tests on AQA?

No!

Topic Explainer Video

Check out this @LauraDoesBiology video that explains choosing a statistical test or read the full notes below. Once you've gone through the whole note, try out the practice questions!

Why do We Use Statistical Tests?

In Biology, results are often influenced by natural variation. Statistical tests help scientists:

  • Determine if differences between groups or relationships between variables are significant.

  • Test hypotheses objectively.

  • Decide whether to accept or reject the null hypothesis.

Null Hypothesis = Assumes no difference, no effect, or no correlation.

Which Statistical Test Should You Use?

Choosing the right test depends on:

  • The type of data (categoric data or continuous data)

  • What you're testing (difference or correlation)

1. Student’s t-test (Difference between means)

Use When...

Details

Comparing means of two groups to determine if there is a significant difference

e.g., Mean height of plants in two different soils

Data must be continuous 

What it shows: Is there a significant difference between the means?

2. Spearman’s Rank Correlation Coefficient (Relationship between variables)

Use When...

Details

Testing for a correlation between two variables to determine if there is a significant correlation

e.g., Does light intensity correlate with photosynthesis rate?

Data must be collected from the same sample and ranked.

What it shows: Strength and direction of a correlation (positive or negative).

3. Chi-squared    Test

Use When...

Details

Testing if observed frequencies are significantly different from an expected ratio 

e.g., Mendelian genetics ratios like 

Data must be categoric (counts, not measurements)

What it shows: Is there a significant difference between observed frequencies and expected frequencies?

Standard Deviation

Standard deviation is not a statistical test.

Explanation:

  • Standard deviation (SD) is a descriptive statistic.
  • It summarises how spread out the data are around the mean.
  • It helps you describe and visualise variation in your data, but it does not tell you whether differences or correlations are statistically significant.

 

Re-cap on standard deviation

  • Standard deviation (SD) measures the spread of values around the mean.
  • A small SD = data points are close to the mean → high precision and low variability.
  • large SD = data points are spread out → low precision and high variability.

Use standard deviation bars on graphs to judge whether differences between groups are meaningful: if the SD bars overlap, differences are unlikely to be significant.

Exam Tip

  • If a question asks, "which statistical test would you use?" standard deviation is NOT the answer.
  • If a question asks, "how could you assess variability?", then standard deviation is the correct answer.
No answer provided.

Summary

Key Terms

  • Null Hypothesis (H₀): Assumes no effect or no difference between groups
  • Student’s t-test: Tests for a difference between the means of two groups
  • Spearman’s Rank Correlation: Tests for a correlation between two variables
  • Chi-squared Test: Tests for a significant difference between observed and expected frequencies
  • Standard Deviation: Measures the amount of variation or spread around the mean
No answer provided.

Exam Tip

  • Choose your statistical test based on the data type and question aim (difference or correlation.
  • Remember: Means = t-test, Correlation = Spearman’s Rank, Categories = Chi-squared.

You will not have to calculate these tests from scratch, so don't try to learn the formulae!

You will be asked to state which type of test you would use for the experimental data described, and give your reasons why.

You may also be provided with the formula and asked to finish off calculating the test statistic.

No answer provided.

The doctors carried out a statistical test to determine whether antibody concentrations were significantly different in girls given two doses of a vaccine, compared with girls given three doses. They determined the mean concentrations of antibody 12 months after the first dose of vaccine.

What statistical test should the doctors have used? Give the reason for your choice.

T-test

Comparing two mean values (to see if there is a significant difference)

A team of scientists investigated whether the number of copies of the amylase gene, AMY1, was associated with the concentration of amylase in the saliva of 58 human volunteers. Give the name of the statistical test they should use to analyse their data, give a reason for your choice.

Correlation coefficient/ Spearman’s rank

Looking for a correlation/ association between two variables (to see if there is a significant correlation)

Name the statistical test you should use to determine if the observed frequencies of the four phenotypes differed significantly from the frequencies expected according to the Hardy–Weinberg equation.

Chi squared / X2

(As you have categoric data and are looking for a significant difference between observed and expected frequencies).

Practice Question

Try to answer the practice question from the TikTok on your own, then watch the video to see how well you did!