Descriptive Statistics Calculations

Why Descriptive Statistics Matter for Safety Professionals

Safety professionals are surrounded by data: injury counts, exposure sample results, near-miss reports, audit scores, noise dosimetry readings, and training completion rates. Raw numbers by themselves tell you very little. Descriptive statistics turn a pile of data into a handful of meaningful values that answer two essential questions: Where is the center of this data? and How spread out is it?

The BCSP includes descriptive statistics in the ASP blueprint under Mathematical Calculations because a competent safety professional must be able to summarize data accurately, interpret exposure monitoring results against occupational exposure limits, recognize when a data set is skewed, and communicate trends to management in a credible way. On the exam, these questions are highly predictable and highly learnable. They are among the most reliable points you can earn if you know the formulas and work carefully.

What Descriptive Statistics Are

Descriptive statistics summarize and describe the features of a data set. They do not make predictions or draw conclusions about a larger population (that is inferential statistics, which includes hypothesis testing and confidence intervals). Descriptive statistics fall into two main families:

1. Measures of Central Tendency (where the data clusters)
• Mean – the arithmetic average
• Median – the middle value when data are ordered
• Mode – the most frequently occurring value
• Geometric Mean – used for lognormally distributed data such as industrial hygiene exposure samples
• Weighted Mean – an average where some values count more than others (the 8-hour TWA is a weighted mean)

2. Measures of Dispersion or Variability (how spread out the data is)
• Range – highest value minus lowest value
• Variance – the average of the squared deviations from the mean
• Standard Deviation – the square root of the variance, expressed in the same units as the data
• Coefficient of Variation (CV) – standard deviation divided by the mean, expressed as a percentage; allows comparison of variability between data sets with different units or magnitudes
• Percentiles and Quartiles – values below which a given percentage of the data falls

How the Calculations Work

Mean (x̄)
x̄ = Σx / n
Add all values and divide by the number of values. The symbol Σ (sigma) means sum of. For a population the mean is written as μ (mu), but the calculation is identical.

Median
Arrange the data in ascending order. If n is odd, the median is the middle value. If n is even, the median is the average of the two middle values. The median is resistant to outliers, which is why it is preferred when a data set contains extreme values.

Mode
The value that appears most often. A data set may have no mode, one mode (unimodal), two modes (bimodal), or more. The mode is the only measure of central tendency that can be used with categorical data (for example, the most common body part injured).

Range
Range = Maximum − Minimum
Simple, but heavily influenced by a single outlier.

Variance
Sample variance: s² = Σ(x − x̄)² / (n − 1)
Population variance: σ² = Σ(x − μ)² / N
Note the critical difference: a sample divides by n − 1 (called degrees of freedom), while a population divides by N. Exam questions almost always involve a sample unless they explicitly say the data represent the entire population.

Standard Deviation
Sample: s = √[ Σ(x − x̄)² / (n − 1) ]
Population: σ = √[ Σ(x − μ)² / N ]
Standard deviation is simply the square root of the variance. It is the single most important dispersion statistic on the exam.

Coefficient of Variation
CV = (s / x̄) × 100%
A CV of 10% or less is often considered acceptable precision for sampling and analytical methods in industrial hygiene.

Geometric Mean
GM = (x₁ × x₂ × ... × xₙ)^(1/n), or equivalently the antilog of the mean of the logarithms of the values. Exposure data are typically lognormal (a few very high results and many lower results), so the geometric mean and geometric standard deviation (GSD) describe them more accurately than the arithmetic mean.

Fully Worked Example

A safety manager records the number of recordable injuries per month over a 6-month period: 2, 4, 4, 5, 7, 8

Step 1 – Mean: Σx = 2 + 4 + 4 + 5 + 7 + 8 = 30. x̄ = 30 / 6 = 5.0 injuries per month

Step 2 – Median: Data are already in order. n = 6 (even), so average the 3rd and 4th values: (4 + 5) / 2 = 4.5

Step 3 – Mode: The value 4 appears twice; all others appear once. Mode = 4

Step 4 – Range: 8 − 2 = 6

Step 5 – Deviations from the mean and their squares:
2 − 5 = −3 → 9
4 − 5 = −1 → 1
4 − 5 = −1 → 1
5 − 5 = 0 → 0
7 − 5 = 2 → 4
8 − 5 = 3 → 9
Σ(x − x̄)² = 9 + 1 + 1 + 0 + 4 + 9 = 24
(Quick check: the deviations themselves must sum to zero: −3 −1 −1 + 0 + 2 + 3 = 0. ✓)

Step 6 – Sample variance: s² = 24 / (6 − 1) = 24 / 5 = 4.8

Step 7 – Sample standard deviation: s = √4.8 = 2.19 injuries per month

Step 8 – Coefficient of variation: CV = (2.19 / 5.0) × 100% = 43.8%

If the question had stated these six months were the entire population of interest, the variance would be 24 / 6 = 4.0 and the standard deviation would be √4.0 = 2.0. Notice how different the answer is; this n versus n − 1 distinction is a favorite exam distractor.

Interpreting the Results

Because the mean (5.0) is greater than the median (4.5), the distribution is slightly right-skewed (positively skewed), pulled upward by the larger values. In a perfectly symmetrical (normal) distribution, mean = median = mode. When mean < median, the data are left-skewed (negatively skewed). Exam questions frequently test this relationship conceptually without requiring any arithmetic.

The Normal Distribution and the Empirical Rule

Although strictly a bridge toward inferential statistics, the ASP exam expects you to know the 68-95-99.7 rule for normally distributed data:
• About 68% of values fall within ±1 standard deviation of the mean
• About 95% fall within ±2 standard deviations
• About 99.7% fall within ±3 standard deviations

Example: If noise exposures average 87 dBA with a standard deviation of 2 dBA and are normally distributed, roughly 95% of workers fall between 83 and 91 dBA. You should also be able to compute a z-score: z = (x − x̄) / s, which expresses how many standard deviations a single value lies from the mean.

How to Answer Descriptive Statistics Questions on the Exam

Step 1 – Identify exactly what is being asked. Mean, median, mode, range, variance, standard deviation, or CV? Read the final sentence of the question first.

Step 2 – Determine sample versus population. Look for words like sample, randomly selected, or representative (use n − 1) versus all, every, or entire (use N). When in doubt, assume sample.

Step 3 – Order the data. Even if the question asks for the mean, sorting first makes the median, mode, and range immediately visible and reduces errors.

Step 4 – Write out the formula from the provided reference sheet before plugging in numbers. The BCSP exam supplies formulas; your job is to select the right one and execute it correctly.

Step 5 – Compute methodically. Build a small table of x, (x − x̄), and (x − x̄)² on your scratch paper. Verify the deviations sum to zero.

Step 6 – Finish the calculation. The most common error is stopping at the variance when the question asks for standard deviation, or forgetting to take the square root. Another is dividing by n instead of n − 1.

Step 7 – Sanity-check the answer. The standard deviation can never be negative, can never be larger than the range, and is usually roughly one quarter to one third of the range for small data sets. The mean must lie between the minimum and maximum values.

Exam Tips: Answering Questions on Descriptive Statistics Calculations

• Memorize the vocabulary, not just the math. Many questions are purely conceptual: Which measure of central tendency is least affected by outliers? (Median.) Which measure can be used with nominal data? (Mode.) Which statistic has the same units as the original data? (Standard deviation, not variance.)

• Know the n − 1 rule cold. Expect at least one answer choice computed with N instead of n − 1 as a deliberate distractor. Compute it the correct way and you will see the trap.

• Do not confuse variance with standard deviation. If both appear among the answer choices, the exam is testing whether you took the square root.

• Use your calculator statistics mode wisely. The exam calculator may allow statistical entry, but make sure you know which key gives sample standard deviation (often labeled s, sₓ, or σₙ₋₁) versus population standard deviation (σ or σₙ). If unsure, compute by hand; the data sets are deliberately small.

• Watch for even versus odd data counts when finding the median. With an even count, you must average the two middle values.

• Recognize lognormal data. If a question mentions industrial hygiene exposure results or says the data are lognormally distributed, the geometric mean and geometric standard deviation are the appropriate descriptors, not the arithmetic mean.

• Connect skewness to the mean-median relationship. Mean > median means right-skewed; mean < median means left-skewed; mean = median = mode means symmetrical. Injury and exposure data are usually right-skewed.

• Remember the empirical rule percentages. 68, 95, and 99.7 are tested repeatedly in both calculation and interpretation questions.

• Carry extra decimal places until the final step. Rounding the mean early can shift the standard deviation enough to land on a wrong answer choice.

• Include units in your answer. If the data are in ppm, the mean and standard deviation are in ppm; the variance is in ppm²; the CV is a unitless percentage.

• Budget your time. A full standard deviation calculation for six to eight data points should take two to three minutes. If a question seems to require far more arithmetic, re-read it; you may be misreading what is asked.

Quick Reference Summary

Mean: x̄ = Σx / n
Median: middle value of ordered data (average of two middle values if n is even)
Mode: most frequent value
Range: max − min
Sample variance: s² = Σ(x − x̄)² / (n − 1)
Sample standard deviation: s = √s²
Population standard deviation: σ = √[ Σ(x − μ)² / N ]
Coefficient of variation: CV = (s / x̄) × 100%
z-score: z = (x − x̄) / s
Empirical rule: 68% / 95% / 99.7% within 1 / 2 / 3 standard deviations

Master these fundamentals and descriptive statistics questions become some of the most dependable points you will earn on the ASP exam. More importantly, you will be equipped to turn raw safety data into clear, defensible information that drives better decisions in the workplace.