Analysis of Frequency Distributions and Coefficient of Variation

1. The Need for a Relative Measure of Dispersion

Standard deviation is an excellent absolute measure of dispersion. It tells us the spread of data in the same units as the data itself (e.g., if heights are in cm, the standard deviation is also in cm). However, this becomes a problem when we want to compare the variability of two different datasets.

There are two main scenarios where standard deviation is not suitable for comparison:

• Datasets with Different Means: Imagine two cricket batsmen. 🏏

  • Batsman A: Mean Score $\bar{x}_A = 50$, Standard Deviation $\sigma_A = 10$.
  • Batsman B: Mean Score $\bar{x}_B = 100$, Standard Deviation $\sigma_B = 10$. Both have the same standard deviation, but a 10-run variation is much more significant for a batsman averaging 50 runs than for one averaging 100. Intuitively, Batsman B is more consistent.

• Datasets with Different Units: How would you compare the consistency of student heights (in cm) versus their weights (in kg)?

  • Heights: Mean = 160 cm, Standard Deviation = 8 cm.
  • Weights: Mean = 50 kg, Standard Deviation = 5 kg. Is a spread of 8 cm 'more' or 'less' than a spread of 5 kg? A direct comparison is meaningless.

To solve these issues, we need a relative measure of dispersion that is unitless and accounts for the mean. The most important one is the Coefficient of Variation.

---

2. Coefficient of Variation (C.V.)

The Coefficient of Variation (C.V.) is a standardized, relative measure of dispersion. It expresses the standard deviation as a percentage of the mean. This removes the units and allows for a fair comparison of variability between different datasets. 📊

$$\text{C.V.} = \frac{\sigma}{\bar{x}} \times 100 \quad (\text{where } \bar{x} \neq 0)$$

• $\sigma$: Standard Deviation of the data.
• $\bar{x}$: Mean of the data.

Example Calculation: Find the C.V. for the dataset {10, 15, 20, 25, 30}.

1. Calculate Mean ($\bar{x}$): $\bar{x} = \frac{10+15+20+25+30}{5} = \frac{100}{5} = 20$.
2. Calculate Variance ($\sigma^2$): $\sigma^2 = \frac{(10-20)^2 + (15-20)^2 + (20-20)^2 + (25-20)^2 + (30-20)^2}{5} = \frac{100+25+0+25+100}{5} = \frac{250}{5} = 50$.
3. Calculate Standard Deviation ($\sigma$): $\sigma = \sqrt{50} \approx 7.07$.
4. Calculate C.V.: C.V. = $\frac{7.07}{20} \times 100 = \mathbf{35.35\%}$.

---

3. Interpreting and Comparing C.V.

The C.V. is the perfect tool for comparing consistency. The rule is simple:

• Lower C.V. → Higher Consistency: A smaller C.V. means the data is tightly clustered around the mean, relative to the mean's value. This implies less variability.
• Higher C.V. → Lower Consistency: A larger C.V. means the data is more spread out relative to its mean. This implies more variability.

Decision Rule: For two datasets A and B:

• If C.V.(A) < C.V.(B), then dataset A is more consistent (less variable).
• If C.V.(A) > C.V.(B), then dataset B is more consistent (less variable).

Putting It All Together (Revisiting Earlier Examples):

• Cricket Batsmen:

  • C.V. (Batsman A) = $\frac{10}{50} \times 100 = 20\%$.
  • C.V. (Batsman B) = $\frac{10}{100} \times 100 = 10\%$. Since $10\% < 20\%$, our calculation confirms that Batsman B is the more consistent player.

• Heights vs. Weights:

  • C.V. (Heights) = $\frac{8 \text{ cm}}{160 \text{ cm}} \times 100 = 5\%$.
  • C.V. (Weights) = $\frac{5 \text{ kg}}{50 \text{ kg}} \times 100 = 10\%$. Since $5\% < 10\%$, we can conclude that the students' heights are relatively more consistent than their weights.