Revision notes

Chapter 6 Correlation: Foundation

The key points, rules to remember, tables and diagrams from all 5 lessons, on one page for quick revision. To keep a copy, use your browser's Print and choose "Save as PDF".

Lesson 1

Meaning of correlation: relationships, causation and direction

Why two variables may move together

  • Cause and effect One variable influences the other
    • Low rainfall and low farm productivity
    • Price and quantity demanded
  • Pure coincidence No sensible causal link
    • Arrival of migratory birds and the local birth rate
    • Shoe size and money in your pocket
  • A third variable Both are driven by something else
    • Ice-cream sales and drowning deaths both rise with temperature

Correlation tells us which way and how strongly two variables move together. It describes moving together (covariation), not one causing the other, so it is never by itself proof of cause and effect.

Direction of correlation

  • Positive correlation
    • Variables move in the same direction
    • Income rises, consumption rises; income falls, consumption falls
    • Temperature and ice-cream sales
    • Price of coffee and demand for tea (a substitute)
  • Negative correlation
    • Variables move in opposite directions
    • Price of apples falls, demand for apples rises
    • More hours of study, fewer chances of failing
    • Rate of interest rises, demand for funds falls
Illustrative data
Week (illustrative)Tomato arrivals at a mandi (tonnes)Price (₹ per kg)
12038
23530
35022
47014
5908
Pair of variables (illustrative)DirectionBest explanation
Family income and spending on clothesPositiveCause and effect: more income allows more spending
Sales of umbrellas and sales of raincoats in a townPositiveThird variable: both rise with rainfall
Number of films released in a year and number of mangoes harvestedCould be anythingCoincidence: no sensible link
Price of a bus ticket and number of bus passengersNegativeCause and effect: a higher fare discourages travel

Key takeaways: correlation studies whether, in which direction and how strongly two variables move together. Relationships may be causal, coincidental or due to a third variable; correlation measures covariation, not causation. Positive correlation means movement in the same direction; negative, in opposite directions.

Read the full lesson

Lesson 2

Scatter diagrams

Reading a scatter diagram
Pattern of pointsWhat it shows
Scattered around an upward-rising linePositive correlation: X up, Y up
Scattered around a downward-sloping lineNegative correlation: X up, Y down
No upward or downward line; points spread everywhereNo correlation
All points exactly on an upward-rising linePerfect positive correlation (r = +1)
All points exactly on a downward-sloping linePerfect negative correlation (r = −1)
Points close to a lineHigh (strong) correlation
Points widely dispersed around the lineLow (weak) correlation
Points along a curve (rising then flattening, or U-shaped)Non-linear relationship

A scatter diagram shows the nature of the relationship visually and is not confined to linear relations; it gives no numerical value. If all points lie on a line, the correlation is perfect.

Patterns seen in scatter diagrams

  • Linear
    • Positive (upward line)
    • Negative (downward line)
    • Perfect positive or perfect negative (on the line)
  • No correlation
    • Points with no upward or downward line
  • Non-linear
    • Positive non-linear (rising curve)
    • Negative non-linear (falling curve)
    • U-shaped or other curves

Scatter diagram as a tool

  • Strengths
    • Quick visual idea of direction and closeness
    • Shows non-linear patterns too
    • Helps decide whether r is appropriate
  • Limitations
    • Gives no numerical value
    • Judgement of 'high' or 'low' is rough
    • Hard to compare two scatters precisely

Key takeaways: the three tools are the scatter diagram, Karl Pearson's r and Spearman's rank correlation. A scatter diagram plots pairs as points; upward drift means positive, downward means negative, no drift means none, points on a line mean perfect correlation, and a curve means a non-linear relation. Check the scatter before computing r.

Read the full lesson

Lesson 3

Karl Pearson's coefficient of correlation

r = Cov(X, Y) ÷ (σx × σy) = Σxy ÷ (N σx σy) = Σxy ÷ √(Σx² × Σy²), where x = X − X̄ and y = Y − Ȳ.

Illustrative data
XYx = X − 6y = Y − 16x²y²xy
213−4−316912
415−2−1412
6140−2040
81822444
102044161616
ΣX = 30ΣY = 8000403434

Steps to calculate Karl Pearson's r

  1. Find the means X̄ = ΣX ÷ N, Ȳ = ΣY ÷ N
  2. Deviations x = X − X̄, y = Y − Ȳ (each sums to zero)
  3. Squares and products x², y², xy
  4. Totals Σx², Σy², Σxy
  5. Apply the formula r = Σxy ÷ √(Σx² Σy²)

Key takeaways: Karl Pearson's r measures the degree of linear relationship. With deviations from means, r = Σxy ÷ √(Σx² Σy²); from actual values, r = [NΣXY − ΣXΣY] ÷ √{[NΣX² − (ΣX)²][NΣY² − (ΣY)²]}. The sign of the covariance gives the sign of r; values near ±1 mean strong, near 0 weak, linear relationships.

Read the full lesson

Lesson 4

Properties of r and the step deviation method

Reading the value of r

  1. r = −1 Perfect negative linear relation
  2. r near −1 Strong negative
  3. r near 0 Weak or no linear relation
  4. r near +1 Strong positive
  5. 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.

r = [ΣUV − (ΣU)(ΣV) ÷ N] ÷ √{[ΣU² − (ΣU)² ÷ N] × [ΣV² − (ΣV)² ÷ N]}

Illustrative data
XY (₹ crore)U = (X − 150) ÷ 20V = (Y − 2,400) ÷ 100U²V²UV
1102,100−2−3496
1302,300−1−1111
1502,40000000
1702,800141164
1903,0002643612
Total06106223
Reported value (illustrative)VerdictMeaning
r = 1.25Impossibler cannot exceed 1: there is a calculation error
r = −0.95ValidStrong negative linear relationship
r = 0ValidNo linear relationship; a curved one may still exist
r = 0.15ValidWeak positive linear relationship
r = −1ValidPerfect negative linear relationship

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.

Read the full lesson

Lesson 5

Spearman's rank correlation

Spearman's rank correlation coefficient: rs = 1 − 6ΣD² ÷ (n³ − n), where n is the number of pairs and D is the difference between the two ranks of each item. Like r, it lies between −1 and +1 and is interpreted the same way.

ContestantJudge PJudge QD = P − QD²
112−11
22111
334−11
44311
556−11
66511
Total06
StudentEconomicsAccountancyRank in EconomicsRank in AccountancyDD²
A72682200
B58624311
C655534−11
D80751100
E50485500
Total02

With ties: rs = 1 − 6[ΣD² + (m1³ − m1) ÷ 12 + (m2³ − m2) ÷ 12 + ...] ÷ (n³ − n), where m1, m2, ... are the numbers of items in each tied group.

FirmRank of XRank of YDD²
112−11
22.511.52.25
32.53.5−11
443.50.50.25
55500
66600
Total04.5

Karl Pearson's r and Spearman's rs

  • Karl Pearson's r
    • Uses actual values
    • Needs measurable variables
    • Affected by extreme values
    • More accurate for precisely measured data
  • Spearman's rs
    • Uses ranks
    • Works for attributes and unmeasurable variables
    • Not affected by extreme values
    • Loses some information; generally rs ≤ r

Key takeaways: Spearman's rank correlation, rs = 1 − 6ΣD² ÷ (n³ − n), is used for attributes, for variables that can be ranked but not measured, for some non-linear relations and when extreme values are present. Tied items share the average rank, and (m³ − m) ÷ 12 is added to ΣD² for each tied group. It lies between −1 and +1 and does not imply causation.

Read the full lesson

Privacy · Terms · Refunds · Delivery