- Foundation
- About 20 min
- 5 practice questions
Properties of r and the step deviation method
State the properties of the correlation coefficient, interpret r = +1, −1 and 0, and use the step deviation method (change of origin and scale) to calculate r.
If something looks wrong, please tell us.
By the end of this lesson you can
- State the properties of Karl Pearson's coefficient of correlation.
- Interpret values of r, including r = 0, and spot impossible values.
- Calculate r with the step deviation method.
Karl Pearson's r has a set of properties that make it easy to interpret and, in one case, easier to calculate.
Properties of r
- r has no unit; it is a pure number. r between height in feet and weight in kilograms is just a number such as 0.7.
- Negative r means an inverse relationship (a rise in price with a fall in demand; a rise in the interest rate with a fall in demand for funds). Positive r means movement in the same direction (price of coffee and demand for tea, a substitute; irrigation and yield).
- r lies between −1 and +1: −1 ≤ r ≤ 1. A value outside this range means an error in calculation.
- r is unaffected by a change of origin and scale. If U = (X − A) ÷ B and V = (Y − C) ÷ D, with B and D of the same sign, then r between U and V equals r between X and Y.
- r = 0 means the variables are uncorrelated: there is no linear relationship, though other (non-linear) relationships may exist.
- r = +1 or −1 means perfect correlation: an exact linear relationship.
- r close to ±1 means a strong linear relationship; r close to 0 means a weak linear relationship, though there may be a non-linear one.
Reading the value of r
- r = −1 Perfect negative linear relation
- r near −1 Strong negative
- r near 0 Weak or no linear relation
- r near +1 Strong positive
- r = +1 Perfect positive linear relation
Because r does not change when we subtract a constant from a variable (change of origin) or divide it by a positive constant (change of scale), we can calculate r from simplified values U = (X − A) ÷ B and V = (Y − C) ÷ D.
The step deviation method
When values are large, transform them: subtract an assumed mean and divide by a common factor. Then apply the actual-value formula to U and V:
r = [ΣUV − (ΣU)(ΣV) ÷ N] ÷ √{[ΣU² − (ΣU)² ÷ N] × [ΣV² − (ΣV)² ÷ N]}
Worked example 1: step deviation method
Illustrative: a price index X and money supply Y (₹ crore) in a made-up economy over five years: (110, 2,100), (130, 2,300), (150, 2,400), (170, 2,800), (190, 3,000). Take U = (X − 150) ÷ 20 and V = (Y − 2,400) ÷ 100.
| X | Y (₹ crore) | U = (X − 150) ÷ 20 | V = (Y − 2,400) ÷ 100 | U² | V² | UV |
|---|---|---|---|---|---|---|
| 110 | 2,100 | −2 | −3 | 4 | 9 | 6 |
| 130 | 2,300 | −1 | −1 | 1 | 1 | 1 |
| 150 | 2,400 | 0 | 0 | 0 | 0 | 0 |
| 170 | 2,800 | 1 | 4 | 1 | 16 | 4 |
| 190 | 3,000 | 2 | 6 | 4 | 36 | 12 |
| Total | 0 | 6 | 10 | 62 | 23 |
Step 1, totals: ΣU = 0, ΣV = 6, ΣU² = 10, ΣV² = 62, ΣUV = 23, N = 5. Step 2, numerator: 23 − (0 × 6) ÷ 5 = 23. Step 3, denominator: √[(10 − 0) × (62 − 36 ÷ 5)] = √(10 × 54.8) = √548 = 23.41. Step 4: r = 23 ÷ 23.41 = 0.98. A very high positive correlation between the price index and money supply. The arithmetic used numbers like 2 and 6 instead of 190 and 3,000, yet the answer is exactly what the original values would give.
Worked example 2: why change of origin and scale does not matter
Illustrative: take the fertiliser data of the previous lesson (X = 2, 4, 6, 8, 10; Y = 13, 15, 14, 18, 20), for which r = 0.92. Let U = (X − 6) ÷ 2 = −2, −1, 0, 1, 2 and V = Y − 16 = −3, −1, −2, 2, 4. Then ΣUV = 6 + 1 + 0 + 2 + 8 = 17, ΣU² = 10, ΣV² = 34, and both ΣU and ΣV are 0. r = 17 ÷ √(10 × 34) = 17 ÷ √340 = 17 ÷ 18.44 = 0.92. Same r. Halving X halved every product and every x², and the halving cancels in the formula.
Worked example 3: spotting errors and interpreting
| Reported value (illustrative) | Verdict | Meaning |
|---|---|---|
| r = 1.25 | Impossible | r cannot exceed 1: there is a calculation error |
| r = −0.95 | Valid | Strong negative linear relationship |
| r = 0 | Valid | No linear relationship; a curved one may still exist |
| r = 0.15 | Valid | Weak positive linear relationship |
| r = −1 | Valid | Perfect negative linear relationship |
Remember the U-shaped example (Y = X squared for X = −3 to 3): r = 0 there, yet Y depends exactly on X. So zero correlation does not mean independence; it means no linear relation.
Key takeaways: r is a pure number between −1 and +1; its sign shows direction and its size shows the strength of the linear relationship. r = ±1 is perfect, r = 0 means no linear relation (not necessarily independence). r is unaffected by change of origin and scale, which gives the step deviation method: work with U = (X − A) ÷ B and V = (Y − C) ÷ D.
Check your understanding
-
1. The unit of the correlation coefficient between income in rupees and expenditure in rupees is:
- rupees
- rupees squared
- per cent
- none: r is a pure number
Show answer
Answer: D. r has no unit; it is a pure number whatever units X and Y are measured in.
-
2. A student reports r = −1.4. This shows that:
- the correlation is very strongly negative
- there is an error in calculation
- the relation is non-linear
- the variables are independent
Show answer
Answer: B. r always lies between −1 and +1; a value outside this range indicates a calculation error.
-
3. If r = 0 between X and Y, then:
- there is no linear relationship, though a non-linear one may exist
- X and Y must be independent
- X and Y are perfectly related
- the scatter points lie on a horizontal line only
Show answer
Answer: A. r = 0 means the variables are uncorrelated in the linear sense; a non-linear relation (such as a U-shape) may still exist, so independence does not follow.
-
4. Illustrative: with U and V from step deviations, ΣUV = 12, ΣU = 0, ΣV = 5, ΣU² = 8, ΣV² = 30 and N = 5. The value of r is:
- 0.6
- 0.95
- 0.85
- 1.2
Show answer
Answer: C. Numerator = 12 − (0 × 5) ÷ 5 = 12. Denominator = √[(8 − 0) × (30 − 25 ÷ 5)] = √(8 × 25) = √200 = 14.14. r = 12 ÷ 14.14 = 0.85. A value such as 1.2 is impossible, since r cannot exceed 1.
-
5. Assertion (A): Karl Pearson's r can be calculated from U = (X − A) ÷ B and V = (Y − C) ÷ D instead of X and Y. Reason (R): r is independent of change of origin and scale (with B and D of the same sign).
- Both A and R are true, and R is the correct explanation of A
- Both A and R are true, but R is not the correct explanation of A
- A is true, but R is false
- A is false, but R is true
Show answer
Answer: A. Both are true, and R explains A: since subtracting constants and dividing by positive constants leave r unchanged, the simplified U and V give the same r as X and Y.
Sign in to keep track of the lessons you finish.