Welcome to the World of Precision Spread!
In your Statistics journey so far, you’ve likely met the Range (the whole spread) and the Interquartile Range (IQR) (the middle 50%). While these are great, sometimes statisticians want to be even more specific about how "bunched up" or "spread out" data is. That is where Interpercentile and Interdecile ranges come in!
These measures are exclusive to the Higher Tier. They allow us to ignore extreme values (outliers) while looking at a much larger chunk of the data than the IQR does. Don't worry if it sounds complicated—it’s just subtraction with a fancy name!
1. What is an Interpercentile Range?
A percentile is a value that divides a data set into 100 equal parts. For example, the 10th percentile (\(P_{10}\)) is the value that has 10% of the data below it.
An Interpercentile Range (IPR) is simply the difference between two specific percentiles. It tells us the spread of the middle "chunk" of data between those two points.
The Formula:
\( \text{Interpercentile Range} = P_{\text{upper}} - P_{\text{lower}} \)
Example: If you are asked for the 20th to 80th interpercentile range, you would calculate: \( P_{80} - P_{20} \).
Quick Review: Think of this like the Interquartile Range (\(UQ - LQ\)). In fact, the IQR is actually just a specific interpercentile range—the 25th to 75th range!
2. The Interdecile Range (IDR)
A decile is a specific type of percentile that splits data into 10 parts (10%, 20%, 30%, etc.).
The Interdecile Range usually refers to the 10th to 90th interpercentile range. This is the most common version you will see in your Edexcel exam.
How to calculate the 10th to 90th Interdecile Range:
1. Find the value of the 90th percentile (\(P_{90}\)).
2. Find the value of the 10th percentile (\(P_{10}\)).
3. Subtract the smaller from the larger: \( P_{90} - P_{10} \).
Why 10th to 90th?
By looking at the spread between the 10th and 90th percentiles, we are looking at the middle 80% of the data. This is great because it ignores the bottom 10% and top 10%, which are often where weird "outliers" hide, but it still gives us a much broader picture than the IQR (which only looks at the middle 50%).
3. Step-by-Step Example
Imagine you have a set of test scores for 100 students. You have already calculated the following percentiles:
\( P_{90} = 85 \text{ marks} \)
\( P_{10} = 30 \text{ marks} \)
Find the 10th to 90th interdecile range:
\( \text{IDR} = P_{90} - P_{10} \)
\( \text{IDR} = 85 - 30 = 55 \text{ marks} \)
Interpretation: The middle 80% of students have scores that span across a range of 55 marks.
4. Comparing Data Sets
In your exam, you might be asked to compare two different distributions using these ranges. This is a key skill for AO2 (Interpretation) and AO3 (Evaluation) marks.
- A smaller range (Interpercentile or Interdecile) means the data is more consistent or less spread out.
- A larger range means the data is more varied or spread out.
Example Analogy: Imagine two bus routes. Route A has a 10th-90th interdecile range of 5 minutes for wait times. Route B has an IDR of 20 minutes. Route A is much more reliable because the middle 80% of wait times are clustered closer together.
Key Comparison Tip:
When comparing, always use the context provided in the question (e.g., "The heights of plants in Greenhouse A are more consistent because their interdecile range is lower than Greenhouse B").
5. Common Mistakes to Avoid
- Mixing up Position and Value: Don't confuse the position (e.g., the 10th person) with the value (e.g., the 10th person's score is 15). The range is the difference between the values.
- Forgetting to Subtract: Some students find \(P_{90}\) and \(P_{10}\) and stop there. Remember, a "range" is a single number found by subtracting!
- Using the wrong percentiles: Read the question carefully. If it asks for the "5th to 95th interpercentile range," don't use the 10th and 90th by habit.
Quick Summary Checklist
- Interpercentile Range: \( P_{upper} - P_{lower} \).
- Interdecile Range: Usually \( P_{90} - P_{10} \) (the middle 80%).
- Purpose: Measures spread while ignoring extreme values at the very ends.
- Comparison: A lower range = more consistent data; a higher range = more spread out data.
Note: For more ways to measure spread, see the chapters on Standard Deviation or Range and IQR.