10.x. Variance and Standard Deviation

Before defining variance and standard deviation, we first need the mean (μ) of a dataset.

For a discrete dataset with values x₁, x₂,..., xₙ, the mean is:

μ = ∑xᵢ / n (for i = 1 to n)

For a probability distribution, the expected value (mean) is: E[X] = ∑xᵢ ⋅ P(xᵢ) (Discrete) E[X] = ∫ x⋅f(x) dx (Continuous)

The mean is a fundamental concept in statistics, representing the average of a set of numbers. For a simple dataset of discrete values, we find the mean by adding all the values together and then dividing by the number of values (n). In statistics, the mean can also refer to the expected value of a random variable, calculated differently for discrete (by summing up the values multiplied by their probabilities) and continuous distributions (through integration). This average value serves as the central point around which the data points tend to cluster.

Imagine a classroom where five students scored the following points in a test: 4, 8, 6, 5, and 3. To find the average score (mean), we add all their scores (4 + 8 + 6 + 5 + 3 = 26) and divide by the number of students (5). Thus, the average score is 5.2. This number gives us a quick idea of how well the students performed overall.

The variance measures the average of the squared differences from the mean.

Formula (sample data): Variance (s²) = ∑(xᵢ − μ)² / n (for i = 1 to n)

Formula (population): Variance (σ²) = E[(X − μ)²]

Interpretation: • A higher variance indicates more spread-out data. • A lower variance indicates that the data points are close to the mean.

Variance quantifies the degree to which data points differ from the mean. To calculate variance, we first find the differences of each data point from the mean and then square those differences (to eliminate any negatives). We then average these squared differences. A high variance means the data points are widely spread out from the mean, while a low variance suggests that most data points are close to the mean, indicating less variability.

Continuing with the classroom example, if we calculate the variance of the students' scores, we'll find that scores are varied: students scored differently. If most students score around the average (5.2), variance will be low, indicating consistency. However, if one student scores 10 while others score around 3-6, variance will be higher, indicating a wider spread in scores.

The standard deviation (SD) is simply the square root of the variance, providing a measure of dispersion in the same unit as the original data.

s = √Variance = √(∑(xᵢ − μ)² / n (for i = 1 to n)

Why standard deviation? • It is more interpretable than variance due to being in the same unit as the data. • Often used in engineering to understand fluctuations, measurement errors, and signal noise.

Standard deviation is another measure of spread in a dataset, calculated as the square root of variance. Because it is expressed in the same units as the original data, it is often more understandable and relatable than variance. Engineers and scientists prefer using standard deviation as it helps them readily interpret the level of variability or uncertainty in measurements or data.

Think of standard deviation as measuring how far the students' test scores are from the average test score. For example, if we found that the standard deviation of the scores was 1.72, we could say most students scored within 1.72 points of the average score of 5.2, giving a clearer picture of how consistent the students are in their performances.

1. Non-Negative: Both variance and SD are always ≥ 0.
2. Affected by Outliers: Both increase if data has extreme values.
3. Additive Property (Independent variables): o Var(X+Y) = Var(X) + Var(Y)
4. Scaling Rule: o If Y = aX + b, then Var(Y) = a² ⋅ Var(X) o SD(Y) = |a| ⋅ SD(X)

Variance and standard deviation have specific properties that are crucial for data analysis. Firstly, they cannot be negative as they measure the spread of data. When extreme values are present (outliers), they can inflate both variance and standard deviation, masking the true variability of a data set. The additive property means that if you combine independent random variables, you can calculate the variance of their sum simply. The scaling rule indicates how variance and standard deviation change when data is scaled or shifted.

Imagine pouring two different types of liquids with different consistencies into the same container (independent variables). Even if one liquid is thicker (higher variance) than the other, the overall thickness in the container can be calculated by adding their respective thicknesses. If you pour more of one liquid compared to the other, the overall texture changes predictably based on how much you poured — that's the additive property and scaling rule in action!

Example 1: Find the variance and standard deviation of the data: 4,8,6,5,3

Step 1: Mean (4 + 8 + 6 + 5 + 3) = 26 μ = 26 / 5 = 5.2

Step 2: Variance s² = [(4−5.2)² + (8−5.2)² + (6−5.2)² + (5−5.2)² + (3−5.2)²] / 5 = (1.44 + 7.84 + 0.64 + 0.04 + 4.84) / 5 = 14.8 / 5 = 2.96

Step 3: Standard Deviation s = √(2.96) ≈ 1.72

In this example, we calculate the variance and standard deviation for a given set of data. First, we found the mean by averaging the test scores. Then we calculated variance by finding the average of the squared differences from the mean, which quantified how spread out the scores were. Finally, taking the square root of variance gave us the standard deviation, which informs us about the typical distance of each score from the average, helping to understand the consistency in the students' performance.

Let's say we are analyzing the scores of students in a class test. By following the calculated steps, we find the average score and the extent of variation in students' performances. If most students scored around 5, feeling secure, we could conclude that they understood the material well, whereas significant deviation might indicate topics needing more attention.

• In solving PDEs numerically (e.g., finite difference/finite element methods), variance in measurements can influence error estimation. • Standard deviation is often used in stochastic PDEs or uncertainty quantification, where input data or boundary conditions are probabilistic. • In signal processing (often modeled using PDEs), SD quantifies noise level. • Mechanical or electrical systems modeled by PDEs may use SD to assess stability or fluctuation tolerance.

Variance and standard deviation have significant applications in engineering, particularly in the context of partial differential equations (PDEs). For example, numerical methods for solving PDEs require an understanding of variance to refine error estimates related to measurements. The presence of stochastic elements in PDEs makes standard deviation critical in modeling scenarios where variables are not deterministic but probabilistic. This quantification provides engineers with insights into noise levels or stability in dynamic systems.

When engineers model the behavior of a bridge using PDEs, they must account for variations in material properties and loading conditions. By analyzing the variance and standard deviation of these factors, they can predict how the bridge behaves under different conditions, helping ensure safety and durability. If the variance of the applied forces is understood, engineers can design the structure to withstand expected fluctuations without failure.

--