Measures of Central Tendency & Statistical Dispersion Analysis
In descriptive statistics and exploratory data analysis (EDA), measures of central tendency identify the central or typical value around which an empirical dataset clusters. Complementary measures of dispersion quantify the extent of variation, scatter, or spread within the distribution. Together, these metrics summarize complex multidimensional datasets into intuitive analytical parameters.
1. The Arithmetic Mean ($\bar{x}$ and $\mu$)
The mean is the mathematical average of all observations in a dataset. For a sample of $n$ observations $\{x_1, x_2, \dots, x_n\}$, the sample mean $\bar{x}$ is defined as:
$$\bar{x} = \frac{1}{n} \sum_{i=1}^{n} x_i = \frac{x_1 + x_2 + \dots + x_n}{n}$$
For an entire population containing $N$ elements, the population mean $\mu$ shares the identical algebraic formulation with denominator $N$. The mean possesses the mathematical property that the sum of deviations of all observations from the mean is identically zero: $\sum_{i=1}^n (x_i - \bar{x}) = 0$. However, the arithmetic mean is highly sensitive to extreme outliers and asymmetric skewness.
2. The Median ($M$ or $\tilde{x}$)
The median is the positional middle value when the data elements are sorted in ascending numerical order ($x_{(1)} \le x_{(2)} \le \dots \le x_{(n)}$). It divides the probability distribution exactly into equal halves ($50\%$ below, $50\%$ above):
- Odd Sample Size ($n$ is odd): The median occupies the exact central position $k = \frac{n + 1}{2}$: $$M = x_{\left( \frac{n+1}{2} \right)}$$
- Even Sample Size ($n$ is even): The median is the arithmetic average of the two central order statistics: $$M = \frac{x_{\left(\frac{n}{2}\right)} + x_{\left(\frac{n}{2} + 1\right)}}{2}$$
Because the median depends solely on rank order rather than numerical magnitudes at the extremes, it is a robust statistic exhibiting high resistance against extreme outliers.
3. The Mode ($Mo$) & Multimodality
The mode is the data value or values that appear with the highest frequency within the dataset. A dataset may exhibit diverse modal characteristics:
- Unimodal: A unique value appears with the maximum frequency.
- Bimodal / Multimodal: Two or more distinct values share the identical maximum frequency.
- No Mode: Every element in the dataset occurs with identical frequency (e.g., all elements unique).
Distribution Skewness & Relationship Between Metrics
The relative geometric alignment of mean, median, and mode serves as a diagnostic indicator of distributional symmetry:
- Symmetric (Normal / Gaussian): $\text{Mean} \approx \text{Median} \approx \text{Mode}$. The distribution displays a bell curve with zero skewness.
- Positive Skew (Right-Skewed): $\text{Mode} < \text{Median} < \text{Mean}$. Long upper tail; extreme high values pull the mean upward.
- Negative Skew (Left-Skewed): $\text{Mean} < \text{Median} < \text{Mode}$. Long lower tail; extreme low values pull the mean downward.
Measures of Statistical Dispersion & Variability
To gauge the spread around central tendency, statistical dispersion metrics are calculated:
- Range: $R = x_{\max} - x_{\min}$, measuring the total span of values.
- Sample Variance ($s^2$): Incorporates Bessel's correction ($n-1$ degrees of freedom) to ensure an unbiased estimator of population variance: $$s^2 = \frac{1}{n - 1} \sum_{i=1}^n (x_i - \bar{x})^2$$
- Population Variance ($\sigma^2$): Uses total population count $N$: $$\sigma^2 = \frac{1}{N} \sum_{i=1}^N (x_i - \mu)^2$$
- Standard Deviation ($s$ and $\sigma$): The square root of variance, restoring variability into the original units of measurement: $$s = \sqrt{s^2}, \quad \sigma = \sqrt{\sigma^2}$$
- Interquartile Range ($IQR$): The distance between the 75th percentile ($Q_3$) and the 25th percentile ($Q_1$), $IQR = Q_3 - Q_1$, representing the spread of the central $50\%$ of observations.