Correlation: Understanding Relationships Between Variables (Class 11 Statistics)

What is Correlation and Its Types?

Correlation refers to the statistical relationship between two or more variables. When two variables move in a similar direction or in opposite directions, they are said to be correlated. For instance, if an increase in one variable tends to be associated with an increase in another, they are positively correlated. If an increase in one is associated with a decrease in the other, they are negatively correlated. If there's no consistent relationship, they are said to have zero correlation.

Understanding correlation begins with visualizing data using a Scatter Diagram. This is a graphical representation of the relationship between two variables, where each pair of values (X, Y) is plotted as a point. By observing the pattern of these points, we can get an initial idea of the type and strength of correlation:

1. Positive Correlation: The points on the scatter diagram tend to rise from the lower left to the upper right. This indicates that as one variable increases, the other also tends to increase. Examples include height and weight, or hours studied and exam scores.
2. Negative Correlation: The points tend to fall from the upper left to the lower right. This suggests that as one variable increases, the other tends to decrease. Examples include the price of a commodity and its demand, or car speed and travel time for a fixed distance.
3. Zero (or No) Correlation: The points are scattered randomly across the graph, showing no discernible pattern or direction. This implies that there is no linear relationship between the two variables. For example, the height of a person and their intelligence.

Karl Pearson's Coefficient of Correlation (Product-Moment Correlation)

Karl Pearson's coefficient of correlation, denoted by 'r', is the most widely used measure of linear correlation. It quantifies both the strength and direction of the linear relationship between two quantitative variables. The value of 'r' always lies between -1 and +1, inclusive. A value of +1 indicates perfect positive correlation, -1 indicates perfect negative correlation, and 0 indicates no linear correlation.

The formula for Karl Pearson's coefficient of correlation (using the direct method for raw data) is:

$$r = \frac{N\sum XY - (\sum X)(\sum Y)}{\sqrt{[N\sum X^2 - (\sum X)^2][N\sum Y^2 - (\sum Y)^2]}}$$

Where:

• N = Number of pairs of observations
• ΣXY = Sum of the products of the paired values of X and Y
• ΣX = Sum of the values of X
• ΣY = Sum of the values of Y
• ΣX² = Sum of the squares of the values of X
• ΣY² = Sum of the squares of the values of Y

Interpretation of 'r' values:

• r = +1: Perfect positive correlation
• r between +0.7 and +1: Strong positive correlation
• r between +0.3 and +0.7: Moderate positive correlation
• r between 0 and +0.3: Weak positive correlation
• r = 0: No linear correlation
• r between -0.3 and 0: Weak negative correlation
• r between -0.7 and -0.3: Moderate negative correlation
• r between -1 and -0.7: Strong negative correlation
• r = -1: Perfect negative correlation

It's important to remember that correlation does not imply causation. A strong correlation only suggests that two variables tend to move together, but it doesn't necessarily mean that one causes the other. There might be a third, unobserved variable influencing both, or the relationship might be purely coincidental.

Worked Examples: Calculating Pearson's Coefficient

• Example 1: Calculating Pearson's 'r' for Study Hours and Marks

  Let's consider the following data for 5 students showing their weekly study hours (X) and their marks out of 100 (Y) in a test:

  | Student | Study Hours (X) | Marks (Y) |
  |---|---|---|
  | 1 | 3 | 50 |
  | 2 | 5 | 70 |
  | 3 | 2 | 40 |
  | 4 | 6 | 80 |
  | 5 | 4 | 60 |

  Calculate Pearson's coefficient of correlation.

  Step 1: Create a table for necessary calculations.
  We need X, Y, X², Y², and XY.

  | X | Y | X² | Y² | XY |
  |---|---|---|---|---|
  | 3 | 50 | 9 | 2500 | 150 |
  | 5 | 70 | 25 | 4900 | 350 |
  | 2 | 40 | 4 | 1600 | 80 |
  | 6 | 80 | 36 | 6400 | 480 |
  | 4 | 60 | 16 | 3600 | 240 |
  | ΣX=20 | ΣY=300 | ΣX²=90 | ΣY²=19000 | ΣXY=1300 |

  Step 2: Identify N.
  N = 5 (number of students/pairs of observations).

  Step 3: Apply Pearson's formula.
  $$r = \frac{N\sum XY - (\sum X)(\sum Y)}{\sqrt{[N\sum X^2 - (\sum X)^2][N\sum Y^2 - (\sum Y)^2]}}$$

  $$r = \frac{(5)(1300) - (20)(300)}{\sqrt{[(5)(90) - (20)^2][(5)(19000) - (300)^2]}}$$

  $$r = \frac{6500 - 6000}{\sqrt{[450 - 400][95000 - 90000]}}$$

  $$r = \frac{500}{\sqrt{[50][5000]}}$$

  $$r = \frac{500}{\sqrt{250000}}$$

  $$r = \frac{500}{500}$$

  $$r = 1$$

  Interpretation: There is a perfect positive correlation (r=1) between study hours and marks. This means as study hours increase, marks consistently increase proportionally.

Spearman's Rank Correlation Coefficient

While Pearson's coefficient is excellent for quantitative variables with a linear relationship, sometimes we deal with qualitative data (like beauty, honesty, or rankings in a competition) or situations where extreme values might distort the linear relationship. In such cases, Spearman's Rank Correlation Coefficient, denoted by Rs or ρ (rho), is more appropriate.

Spearman's coefficient measures the association between the ranks of two variables rather than their actual values. It is particularly useful when:

• The data is ordinal (ranked).
• The relationship between variables is non-linear but monotonic (always increasing or always decreasing).
• Extreme values are present that could unduly influence Pearson's 'r'.

The formula for Spearman's Rank Correlation Coefficient (when there are no tied ranks) is:

$$R_s = 1 - \frac{6\sum d^2}{N(N^2 - 1)}$$

Where:

• N = Number of pairs of observations
• d = Difference between the ranks of corresponding values of X and Y (Rank_X - Rank_Y)

Steps to calculate Spearman's 'Rs':

1. Assign ranks to the values of the first variable (X). If values are identical (tied ranks), assign the average of the ranks they would have occupied.
2. Assign ranks to the values of the second variable (Y) in the same manner.
3. Calculate the difference (d) between the ranks for each pair of observations.
4. Square each difference (d²).
5. Sum all the d² values (Σd²).
6. Substitute the sum into the formula.

Handling Tied Ranks: When two or more observations have the same value, they are assigned the average of the ranks they would have received. For example, if two values are tied for the 3rd and 4th positions, both are assigned a rank of (3+4)/2 = 3.5. If ties are present, a correction factor is added to the Σd² term in the numerator, but for Class 11 CBSE, the basic formula is generally sufficient unless specified.

Exam Tip: Mastering Correlation

When tackling correlation problems in your CBSE exams, keep these crucial points in mind:

1. Correlation vs. Causation: Always remember that correlation does not imply causation. Just because two variables move together doesn't mean one causes the other. For example, ice cream sales and drowning incidents might be positively correlated, but neither causes the other; a third variable (hot weather) causes both.
2. Formula Accuracy: Be precise with your formulas for both Pearson's and Spearman's coefficients. A small error in squaring or summing can lead to a completely wrong answer. Write down the formula first, then substitute values carefully.
3. Step-by-Step Calculation: Especially for Pearson's 'r', create a clear table for X, Y, X², Y², and XY. This systematic approach reduces calculation errors and helps you organize your work, making it easier for the examiner to follow.
4. Interpretation: Don't just calculate the coefficient; interpret its value. State whether the correlation is positive, negative, or zero, and comment on its strength (weak, moderate, strong, perfect). This shows a deeper understanding.
5. Scatter Diagrams: Practice drawing and interpreting scatter diagrams. They provide a visual check for your calculated correlation coefficient – a positive slope in the diagram should correspond to a positive 'r' or 'Rs'.

Practice Questions with Solutions

• Q: The following data shows the number of hours spent by 5 students on social media (X) and their marks (Y) in a math test:

  | Student | Social Media Hours (X) | Math Marks (Y) |
  |---|---|---|
  | 1 | 8 | 30 |
  | 2 | 6 | 45 |
  | 3 | 10 | 20 |
  | 4 | 7 | 40 |
  | 5 | 9 | 25 |

  Calculate Pearson's coefficient of correlation (r) and interpret the result.
  A: Step 1: Create a calculation table.

  | X | Y | X² | Y² | XY |
  |---|---|---|---|---|
  | 8 | 30 | 64 | 900 | 240 |
  | 6 | 45 | 36 | 2025 | 270 |
  | 10 | 20 | 100 | 400 | 200 |
  | 7 | 40 | 49 | 1600 | 280 |
  | 9 | 25 | 81 | 625 | 225 |
  | ΣX=40 | ΣY=160 | ΣX²=330 | ΣY²=5550 | ΣXY=1215 |

  Step 2: Identify N.
  N = 5.

  Step 3: Apply Pearson's formula.
  $$r = \frac{N\sum XY - (\sum X)(\sum Y)}{\sqrt{[N\sum X^2 - (\sum X)^2][N\sum Y^2 - (\sum Y)^2]}}$$

  $$r = \frac{(5)(1215) - (40)(160)}{\sqrt{[(5)(330) - (40)^2][(5)(5550) - (160)^2]}}$$

  $$r = \frac{6075 - 6400}{\sqrt{[1650 - 1600][27750 - 25600]}}$$

  $$r = \frac{-325}{\sqrt{[50][2150]}}$$

  $$r = \frac{-325}{\sqrt{107500}}$$

  $$r = \frac{-325}{327.87}$$

  Final answer: r ≈ -0.991. This indicates a very strong negative correlation between social media hours and math marks. As social media hours increase, math marks tend to decrease significantly.

• Q: Two judges (A and B) ranked 6 participants in a dance competition as follows:

  | Participant | Judge A's Rank | Judge B's Rank |
  |---|---|---|
  | 1 | 2 | 3 |
  | 2 | 1 | 2 |
  | 3 | 4 | 4 |
  | 4 | 5 | 6 |
  | 5 | 3 | 1 |
  | 6 | 6 | 5 |

  Calculate Spearman's Rank Correlation Coefficient (Rs) and interpret the result.
  A: Step 1: Create a table including ranks, differences (d), and squared differences (d²).

  | Judge A Rank (R₁) | Judge B Rank (R₂) | d = R₁ - R₂ | d² |
  |---|---|---|---|
  | 2 | 3 | -1 | 1 |
  | 1 | 2 | -1 | 1 |
  | 4 | 4 | 0 | 0 |
  | 5 | 6 | -1 | 1 |
  | 3 | 1 | 2 | 4 |
  | 6 | 5 | 1 | 1 |
  | | | | Σd²=8 |

  Step 2: Identify N.
  N = 6 (number of participants).

  Step 3: Apply Spearman's formula.
  $$R_s = 1 - \frac{6\sum d^2}{N(N^2 - 1)}$$

  $$R_s = 1 - \frac{6 \times 8}{6(6^2 - 1)}$$

  $$R_s = 1 - \frac{48}{6(36 - 1)}$$

  $$R_s = 1 - \frac{48}{6 \times 35}$$

  $$R_s = 1 - \frac{48}{210}$$

  $$R_s = 1 - 0.22857$$

  Final answer: Rs ≈ 0.771. This indicates a strong positive rank correlation between the rankings given by Judge A and Judge B, meaning they largely agree in their assessment of the participants.

• Q: What does a correlation coefficient of r = -0.15 signify about the relationship between two variables?
  A: Step 1: Analyze the sign of the coefficient.
  The coefficient is negative (-0.15), which indicates a negative relationship.

  Step 2: Analyze the magnitude of the coefficient.
  The magnitude of 0.15 is close to 0, which implies a very weak relationship.

  Final answer: A correlation coefficient of r = -0.15 signifies a very weak negative linear relationship between the two variables. This means that as one variable increases, the other tends to slightly decrease, but the connection is not strong or consistent.

• Q: Discuss one situation where Spearman's Rank Correlation Coefficient would be more appropriate than Pearson's Coefficient of Correlation.
  A: Step 1: Consider the nature of data.
  Pearson's coefficient works best for quantitative, interval/ratio data that have a linear relationship and are normally distributed. Spearman's is better for ordinal data or when the relationship is monotonic but not necessarily linear.

  Step 2: Provide an example.
  Consider ranking contestants in a beauty pageant or evaluating the quality of different art pieces. These involve subjective judgments that are best expressed as ranks (ordinal data) rather than precise numerical scores on an interval scale. Using Pearson's 'r' on such subjective 'scores' might be misleading or inappropriate, whereas Spearman's 'Rs' directly addresses the consistency of rankings.

  Final answer: Spearman's Rank Correlation Coefficient would be more appropriate when dealing with qualitative attributes that can only be ranked, such as assessing the attractiveness of models, the honesty of employees, or the quality of essays. It's also suitable when the relationship between variables is clearly monotonic (consistently increasing or decreasing) but not necessarily linear, or when there are outliers that could heavily skew Pearson's 'r'.

Frequently Asked Questions