Introduction to Statistical Testing
Welcome to one of the most important parts of your Geographical and fieldwork skills! In your A Level Geography (9GE0) course, you aren't just expected to collect data; you need to prove what that data actually means. Statistical tests help us move from saying "this looks like a pattern" to saying "we are 95% certain this is a pattern."
These tests are essential for your Independent Investigation (NEA) and can appear in Paper 3, where you must interpret or evaluate data. Don't worry if you aren't a "maths person"—we will break these down step-by-step.
1. The Starting Point: Hypotheses
Before you run any test, you need to state what you are testing. In Geography, we use two types of hypotheses:
- The Null Hypothesis (\(H_0\)): This is the "no change" or "no relationship" statement. It assumes that any pattern you see is just down to chance. Example: "There is no significant relationship between distance from the CBD and building height."
- The Alternative Hypothesis (\(H_1\)): This is your prediction that there is a relationship or a difference. Example: "There is a significant positive correlation between distance from the CBD and building height."
Quick Tip: In Geography, we usually aim to reject the null hypothesis to prove our fieldwork discovered something meaningful!
2. Spearman's Rank Correlation Coefficient
What is it?
Spearman's Rank measures the association (relationship) between two sets of data. It tells you if, as one variable increases, the other increases (positive correlation) or decreases (negative correlation).
When to use it:
Use this when you have two sets of continuous data (e.g., pebble size and beach gradient) and you want to see if they are linked.
The Formula:
\(r_s = 1 - \frac{6 \sum d^2}{n(n^2 - 1)}\)
Step-by-Step Logic:
- Rank both sets of data from 1 (smallest) to \(n\) (largest).
- Find the difference (\(d\)) between the ranks of each pair.
- Square those differences (\(d^2\)) to get rid of negative numbers.
- Add them all up (\(\sum d^2\)).
- Plug the numbers into the formula!
Interpreting the Result:
- A result of \(+1\) is a perfect positive correlation.
- A result of \(-1\) is a perfect negative correlation.
- A result of \(0\) means there is no relationship at all.
Common Mistake: Forgetting to rank the data first! You cannot use the raw numbers (like the actual pebble size) in the Spearman's formula; you must use their rank position.
3. Student’s T-Test
What is it?
The t-test is a measure of difference. It tests whether the means (averages) of two groups are significantly different from each other, or if the difference happened by pure chance.
When to use it:
Use this when you have two different sites or groups and you want to compare them. For example: "Is the average pebble size on Beach A significantly different from Beach B?"
The Concept:
The test looks at the "spread" of the data (standard deviation) and the "average" (mean). If the two groups have very different means and the data isn't too messy or spread out, the t-test will give you a "significant" result.
Did you know? This test is great for comparing "before and after" scenarios, such as river velocity before and after a management scheme is installed.
4. Chi-Squared (\(\chi^2\))
What is it?
The Chi-squared test compares observed data (what you actually counted) with expected data (what you would expect to see if everything was equal or followed a specific theory).
When to use it:
Use this for categorical data (frequencies/counts). For example, if you are counting the number of different types of shops in two different towns, or the number of different plant species in two different sand dunes.
The Formula:
\(\chi^2 = \sum \frac{(O - E)^2}{E}\)
Where \(O\) = Observed frequency and \(E\) = Expected frequency.
The Logic:
- If your calculated \(\chi^2\) value is zero, your data perfectly matches what was expected.
- The larger the \(\chi^2\) value, the bigger the difference between what you saw and what was expected.
5. Making Sense of the Results (Significance)
Once you have your "calculated value" from any of the tests above, you aren't finished! You must compare it to a Critical Value found in a statistical table.
The 0.05 Significance Level
In Geography, we almost always use the 0.05 (5%) significance level. This means we are 95% confident that our results are not due to chance.
Degrees of Freedom (\(df\))
To use the tables, you need to know your \(df\). This is usually based on your sample size (\(n\)):
- Spearings Rank: \(df = n\)
- T-test: \(df = (n_1 + n_2) - 2\)
- Chi-squared: \(df = (\text{number of categories}) - 1\)
The Golden Rule:
If your calculated value is GREATER than the critical value from the table, you can reject your Null Hypothesis. This means your result is significant—you've found a real geographical pattern!
Don't worry if this seems tricky at first! Just remember: Calculated > Critical = Significant.
Quick Review: Which test should I use?
1. Are you looking for a relationship/link between two sets of numbers?
Use Spearman's Rank.
2. Are you comparing the averages (means) of two different places?
Use Student's T-Test.
3. Are you comparing frequencies/counts (e.g., number of people or shops)?
Use Chi-Squared.
Key Takeaways for the Exam
- AO3 Skill: You may be asked to "Analyse" or "Interpret" these tests. Always check if the result is higher than the critical value.
- Error Awareness: If a result isn't significant, it might be because your sample size (\(n\)) was too small or there were measurement errors in the field.
- Calculators: You are allowed to use them in Papers 1, 2, and 3. Make sure you practice the square root and squared functions!