Variance Vs Standard Deviation

Download as PDF

How do we describe the spread in a set of data? While the mean tells you the center, variance and standard deviation tell you how clustered or scattered the numbers are. Understanding this is key to making sense of data in science, sports, and everyday life.

What Are Variance and Standard Deviation?

Variance and standard deviation are measures of dispersion, which is a way of saying they tell us how spread out a set of data points is from its average (mean) value. A small variance or standard deviation means the data points are very close to the mean, indicating consistency. A large variance or standard deviation means the data points are far from the mean, indicating a wide range of values.

Think of it like shooting arrows at a target:

A low standard deviation is like having all your arrows land in a tight cluster near the bullseye. Your shots are consistent.
A high standard deviation is like having your arrows scattered all over the target. Your shots are inconsistent.

While they both measure the same thing—spread—they do so in slightly different ways, which we will explore. In short, the standard deviation is simply the square root of the variance.

Why Is Measuring Spread So Important?

Imagine two students, Alice and Bob, who both have an average test score of 85%. On the surface, they seem to be performing equally well. But let's look at their individual scores from five tests:

Alice's Scores: {84, 85, 85, 86, 85}
Bob's Scores: {100, 70, 90, 80, 85}

If you calculate the mean for both, you'll find it's 85 for each. But their performance is very different. Alice is incredibly consistent. Her scores are all tightly clustered around her average. Bob, on the other hand, is all over the place. He has a perfect score and a much lower score of 70. His scores are widely spread out.

This is where variance and standard deviation come in. They give us a number to describe this 'spread.' A teacher looking at these numbers would know that Alice is a very steady student, while Bob's performance is much less predictable. This context is lost if we only look at the average.

How Do You Calculate the Variance?

Calculating variance is a multi-step process, but each step is straightforward. We'll use the formula for population variance, usually denoted by the Greek letter sigma squared (" $σ^{2}$ "). A population includes every member of a group you're studying.

Here are the steps:

Find the mean (average) of the data set. The mean is denoted by $μ$ (the Greek letter mu).
Find the deviation for each data point. Subtract the mean from each individual data point ( $x_{i} - μ$ ). Some results will be positive, and some will be negative.
Square each deviation. Squaring each result from Step 2 makes all the values positive. This is important because it prevents the positive and negative deviations from canceling each other out.
Find the average of the squared deviations. Add up all the squared deviations from Step 3 and divide by the total number of data points ( $N$ ).

This process is summarized by the following formula:

σ^{2} = \frac{\sum_{i = 1}^{N} (x_{i} - μ)^{2}}{N}

Let's break down the formula's symbols:

$σ^{2}$ is the population variance.
$\sum$ (sigma) is a summation symbol, meaning "add everything up."
$x_{i}$ represents each individual data point.
$μ$ is the population mean.
$N$ is the total number of data points in the population.

Example 1

Let's find the variance of a small dataset representing the number of goals scored by a soccer player in 5 games: {2, 4, 5, 6, 8}.

Step 1: Find the mean ( $μ$ ).

$μ = \frac{2 + 4 + 5 + 6 + 8}{5} = \frac{25}{5} = 5$

The mean number of goals is $5$ .

Step 2 & 3: Find and square each deviation from the mean. We can organize this in a table:

Data Point ( $x_{i}$ )	Deviation ( $x_{i} - μ$ )	Squared Deviation ( $(x_{i} - μ)^{2}$ )
2	$2 - 5 = - 3$	$(- 3)^{2} = 9$
4	$4 - 5 = - 1$	$(- 1)^{2} = 1$
5	$5 - 5 = 0$	$0^{2} = 0$
6	$6 - 5 = 1$	$1^{2} = 1$
8	$8 - 5 = 3$	$3^{2} = 9$

Step 4: Find the average of the squared deviations.

First, sum the values in the last column: $9 + 1 + 0 + 1 + 9 = 20$ .

Now, divide by the number of data points ( $N = 5$ ):

$σ^{2} = \frac{20}{5} = 4$

The variance of the goals scored is $4$ .

How Do You Calculate the Standard Deviation?

This is the easy part! The standard deviation is simply the square root of the variance. It is denoted by $σ$ (sigma, without the square).

Once you have the variance ( $σ^{2}$ ), you just perform one more calculation.

σ = \sqrt{σ^{2}} = \sqrt{\frac{\sum_{i = 1}^{N} (x_{i} - μ)^{2}}{N}}

Let's continue with our previous example.

Example 1 (Continued)

We calculated the variance of the soccer player's goals to be $σ^{2} = 4$ .

Step 5: Take the square root of the variance.

$σ = \sqrt{4} = 2$

So, the standard deviation of the goals scored is $2$ . This means that, on average, the player's goals per game are about $2$ goals away from their mean of $5$ .

What's the Real Difference Between Variance and Standard Deviation?

If the standard deviation is just the square root of the variance, why do we need both? The key difference, and the reason standard deviation is often more useful for interpretation, is the units.

Variance is measured in squared units. In our soccer example, the variance is $4$ "goals squared." What is a squared goal? It's not intuitive and has no real-world meaning.
Standard Deviation is measured in the same units as the original data. The standard deviation is $2$ goals. This is much easier to understand. It represents a typical distance of a data point from the mean.

Let's say we measured the heights of five dogs in centimeters: {60, 47, 17, 43, 30}.

The mean height is $39.4$ cm.
The variance would be calculated as $217.04$ cm². (centimeters squared!)
The standard deviation would be $\sqrt{217.04} \approx 14.73$ cm.

It's much more useful to say "The typical deviation from the average height is about $14.73$ cm" than to talk about $217.04$ squared centimeters. Standard deviation brings the measure of spread back into a real-world, understandable context.

Here's a table summarizing the key differences:

Feature	Variance ( $σ^{2}$ )	Standard Deviation ( $σ$ )
Purpose	Measures the average squared distance from the mean.	Measures the typical distance from the mean.
Units	Squared units of the original data (e.g., cm², points²).	Same units as the original data (e.g., cm, points).
Interpretation	Difficult to interpret directly due to squared units.	Easy to interpret as a typical amount of variation.
Calculation	A key step in the process.	The final, more interpretable result ( $\sqrt{variance}$ ).

Population vs. Sample: A Quick Note

So far, we've been calculating the population variance and standard deviation. A population is the entire group you're interested in (e.g., all students in a school, every car produced by a factory).

However, it's often impossible to collect data from an entire population. Instead, we take a sample—a smaller, manageable subset of the population (e.g., 50 students from the school, 100 cars from the factory).

When you calculate variance and standard deviation for a sample, the formula changes slightly. For sample variance (denoted $s^{2}$ ), you divide by $n - 1$ instead of $N$ , where $n$ is the sample size.

Sample Variance:

s^{2} = \frac{\sum_{i = 1}^{n} (x_{i} - \bar{x})^{2}}{n - 1}

Sample Standard Deviation:

s = \sqrt{\frac{\sum_{i = 1}^{n} (x_{i} - \bar{x})^{2}}{n - 1}}

Note the use of $s$ instead of $σ$ and $\bar{x}$ (the sample mean) instead of $μ$ . Why divide by $n - 1$ ? In short, using $n - 1$ gives a better, more accurate estimate of the true population variance. It's a concept called "Bessel's correction." For now, the most important thing is to know that two formulas exist and to use the one your teacher specifies. Unless told otherwise, you are likely working with a population.

Worked Example: Comparing Two Datasets

Example 3

Let's compare the daily high temperatures (°C) for two cities, Sunville and Mistburg, over a 7-day period.

Sunville Temperatures: {25, 26, 25, 24, 25, 27, 26}
Mistburg Temperatures: {20, 30, 22, 28, 19, 31, 25}

Which city has a more consistent climate?

Part A: Sunville Analysis

1. Mean ( $μ_{S}$ ): $\frac{25 + 26 + 25 + 24 + 25 + 27 + 26}{7} = \frac{178}{7} \approx 25.43$ °C

2. Squared Deviations: We find $(x_{i} - 25.43)^{2}$ for each temperature. The sum is approx. $6.86$ .

3. Variance ( $σ_{S}^{2}$ ): $\frac{6.86}{7} \approx 0.98$ °C²

4. Standard Deviation ( $σ_{S}$ ): $\sqrt{0.98} \approx 0.99$ °C

Part B: Mistburg Analysis

1. Mean ( $μ_{M}$ ): $\frac{20 + 30 + 22 + 28 + 19 + 31 + 25}{7} = \frac{175}{7} = 25$ °C

2. Squared Deviations: We find $(x_{i} - 25)^{2}$ for each temperature: $25, 25, 9, 9, 36, 36, 0$ . The sum is $140$ .

3. Variance ( $σ_{M}^{2}$ ): $\frac{140}{7} = 20$ °C²

4. Standard Deviation ( $σ_{M}$ ): $\sqrt{20} \approx 4.47$ °C

Conclusion: Even though the cities have very similar average temperatures ( $25.43$ °C vs $25$ °C), Mistburg's standard deviation ( $4.47$ °C) is much larger than Sunville's ( $0.99$ °C). This tells us that Sunville has a very consistent, stable climate, while Mistburg's weather is far more varied and unpredictable.

Common Mistakes to Avoid

Forgetting to square the deviations. If you just add the deviations ( $x_{i} - μ$ ), they will always sum to zero! Squaring makes them all positive.
Forgetting to take the square root. A common error is to calculate the variance perfectly and then forget the final step of taking the square root to find the standard deviation.
Mistaking the range for standard deviation. The range is simply the highest value minus the lowest value. Standard deviation is a more complex and informative measure of spread that considers every data point.
Using the wrong formula. Be careful to use the population formula (dividing by $N$ ) or the sample formula (dividing by $n - 1$ ) as required by the problem.
Calculation errors. There are many steps, which means many opportunities for small arithmetic mistakes. Double-check your work, especially when summing the squared deviations.

Quick Summary: Variance vs. Standard Deviation

Here's a quick reference to keep the concepts straight.

Goal: Both measure how spread out data is from the mean.
Variance ( $σ^{2}$ ): The average of the squared differences from the mean. Its units are squared (e.g., meters²), making it hard to interpret.
Standard Deviation ( $σ$ ): The square root of the variance. Its units are the same as the original data (e.g., meters), making it the preferred measure for describing spread in an understandable way.
Relationship: $σ = \sqrt{σ^{2}}$ .
Key Takeaway: Low SD means data is clustered and consistent. High SD means data is scattered and variable.

Frequently Asked Questions

Can variance or standard deviation be negative?

No, they can never be negative. The variance is calculated from the sum of squared numbers, which are always non-negative. Since the standard deviation is the square root of the variance, it will also always be non-negative.

What does a standard deviation of 0 mean?

A standard deviation of 0 means there is no spread in the data at all. This only happens when every single data point in the set is exactly the same, so they are all equal to the mean.

Is a smaller or larger standard deviation better?

It completely depends on the context. In manufacturing, a small standard deviation for product size is good because it means consistency. In investing, a high standard deviation means high risk but also potentially high reward.

Why do we square the differences instead of just using their absolute value?

Using the absolute value gives a valid measure of spread called the Mean Absolute Deviation (MAD). However, squaring the differences gives more weight to points that are far from the mean (outliers). This property, along with others, makes standard deviation more useful for advanced statistical theories.

How is standard deviation used in the real world?

It's used everywhere! Meteorologists use it to describe the variability in temperature forecasts. In finance, it represents the risk of an investment. In quality control, it ensures that products like screws or phone screens are manufactured to consistent specifications.

Why do we divide by n-1 when calculating sample variance?

Dividing by n-1, known as Bessel's correction, provides an 'unbiased estimate' of the true population variance. Because a sample is smaller than the full population, its spread is likely to be a bit smaller. Using n-1 adjusts the calculation to better predict the spread of the larger population.

Are variance and standard deviation sensitive to outliers?

Yes, they are very sensitive to outliers. Because the distance from the mean is squared in the calculation, a single data point that is very far from the mean will have a very large impact on the final result, increasing both the variance and standard deviation significantly.