Understanding Z-Scores in Descriptive Statistics

Z-Scores and Standardization

Introduction to Z-Scores

In descriptive statistics, most measures such as the mean, mode, and standard deviation describe characteristics of an entire data set. However, sometimes we are interested in how a single value compares to the rest of the data. Z-scores provide a standardized way to measure how far an individual value is from the mean, relative to the standard deviation.

• Z-score indicates how many standard deviations a value is from the mean.

• Positive z-scores indicate values above the mean; negative z-scores indicate values below the mean.

• Z-scores allow for comparison across different data sets or variables by standardizing values.

Deviations from the Mean

Each value in a data set has a deviation from the mean, calculated as the difference between the value and the mean. The sign of the deviation tells us whether the value is above or below the mean, and the magnitude tells us how far it is from the mean.

• Deviation for a value x: $x - \bar{x}$ (for a sample) or $x - \mu$ (for a population)

• Deviations alone do not indicate whether a value is unusually far from the mean without context.

Comparing Deviations Using Standard Deviation

The standard deviation is a measure of the average distance of values from the mean. To determine if a deviation is large or small, we compare it to the standard deviation by dividing the deviation by the standard deviation. This ratio is the z-score.

• If the deviation is greater than the standard deviation, the value is relatively far from the mean.

• If the deviation is less than the standard deviation, the value is relatively close to the mean.

Formula for Z-Score

The formula for the z-score depends on whether the data set is a sample or a population:

• Sample z-score:

$z = \frac{x - \bar{x}}{s}$

• Population z-score:

$z = \frac{x - \mu}{\sigma}$

• Where $x$ is the value, $\bar{x}$ is the sample mean, $s$ is the sample standard deviation, $\mu$ is the population mean, and $\sigma$ is the population standard deviation.

Worked Example: Calculating Z-Scores

Suppose the mean score on a nationwide test is 150, with a standard deviation of 10. Three friends score 170, 165, and 145. Their z-scores are calculated as follows:

• For 170: $z = \frac{170 - 150}{10} = 2$

• For 165: $z = \frac{165 - 150}{10} = 1.5$

• For 145: $z = \frac{145 - 150}{10} = -0.5$

Interpretation: 170 is 2 standard deviations above the mean, 165 is 1.5 above, and 145 is 0.5 below the mean.

Tabular Example: Z-Scores for a Sample Data Set

Consider the sample data set: 8, 9, 10, 13. The sample mean $\bar{x} = 10$, and the sample standard deviation $s = \sqrt{14/3} \approx 2.2$.

x	Deviation ($x - \bar{x}$)	Z-score ($\frac{x - \bar{x}}{s}$)
8	-2	-0.9
9	-1	-0.5
10	0	0
13	3	1.4

Visualizing Z-Scores: The Z-Score Diagram

A z-score diagram helps visualize the relationship between raw scores (x-values), the mean, and standard deviations. For the test score example:

• The mean (µ) is at 150.

• Each increment of 10 (the standard deviation) marks another point: 160 (µ+σ), 170 (µ+2σ), 180 (µ+3σ), and similarly to the left: 140 (µ–σ), 130 (µ–2σ), 120 (µ–3σ).

• The corresponding z-scores are: 0 at the mean, 1 at one standard deviation above, 2 at two above, 3 at three above, and negative values to the left.

Usual and Unusual Values: Interpreting Z-Scores

Most values in a data set have z-scores between -3 and +3. Values with z-scores beyond this range are considered unusual or outliers.

• Chebyshev's Theorem states that for any data set, at least 89% of values lie within 3 standard deviations of the mean (z-scores between -3 and +3).

• For bell-shaped (normal) distributions, about 98% of values fall within this range.

• Rule of thumb: Almost all values in a data set have z-scores between -3 and +3.

Summary Table: Z-Scores and Their Interpretation

Z-Score Range	Interpretation
z ≈ 0	Value is at the mean
0 < z < 1	Value is above the mean, but typical
1 < z < 2	Value is moderately above the mean
z > 2	Value is unusually high
-1 < z < 0	Value is below the mean, but typical
-2 < z < -1	Value is moderately below the mean
z < -2	Value is unusually low

Applications of Z-Scores

• Comparing scores from different distributions (e.g., test scores from different exams).

• Identifying outliers or unusual values in a data set.

• Standardizing data for further statistical analysis, such as in probability and inferential statistics.

Additional info: Z-scores are foundational for understanding the normal distribution, probability calculations, and hypothesis testing in statistics. Mastery of z-scores is essential for interpreting standardized test results and for comparing data across different scales.