What Does a Standard Deviation Tell You?
In Standard Deviation Using a Calculator, you learned how to use 1-Var Stats and choose between \(Sx\) for a sample and \(\sigma x\) for a population. The next step is explaining what the reported value means. A standard deviation is not just a number to copy from the calculator: it describes how spread out observations are around their mean.
For sample data, the sample standard deviation \(s\) describes a typical distance of the sample observations from their sample mean \(\bar{x}\). For population data, the population standard deviation \(\sigma\) describes a typical distance of the population observations from the population mean \(\mu\). As in earlier tutorials, first identify which group the data represent and which standard deviation applies.
A strong interpretation names the group, the quantitative variable, its mean, and the standard deviation with units. It then states that observations are typically about the standard deviation’s value away from the mean. This model keeps the important reference point—the mean—explicit.
You can also write, “The standard deviation of [variable] for [group] is about [value] [units], so observations typically lie about [value] [units] from the mean.” The exact wording can vary, but the interpretation should retain the idea of a typical distance from the mean.
What “Typical Distance” Does—and Does Not—Mean
The phrase “typical distance” gives a useful sense of spread, but it is not a claim that every observation is exactly that far from the mean. Some observations may be closer, some farther away, and some may equal the mean. The standard deviation summarizes the distribution’s overall spread around its mean; it does not describe each individual observation.
It is also not the average of the ordinary absolute distances from the mean. In Calculating Standard Deviation by Hand, you saw that standard deviation is built using squared deviations, an appropriate denominator, and a square root. That calculation gives a useful measure of typical spread, but do not replace its interpretation with “the average absolute distance.”
The mean is the reference point. If a standard deviation is 5 minutes, that does not mean observations typically differ by 5 minutes from a goal, a median, or a minimum. It means their typical distance is about 5 minutes from the mean. If the question is about distance from a target, compare the mean with the target separately.
Standard deviation uses the same units as the data. A standard deviation of 2.4 seconds describes distances in seconds; it is not 2.4 seconds squared. Although the calculation squares deviations along the way, taking the square root returns the measure to the original units.
Finally, standard deviation alone does not tell you the shape of the distribution or the proportion of observations within a particular distance of the mean. Use a graph and the other appropriate summaries, as in the earlier tutorials on SOCS and choosing measures of spread. Do not apply a percentage rule unless the distribution and the relevant rule have been established.
Worked Examples: Putting Standard Deviation Into Words
Worked Example: Response Times at a Help Desk
A fictional sample of five help-desk chats had response times of 4, 6, 7, 8, and 10 minutes. Find the sample mean and standard deviation, then interpret the standard deviation in context.
Find the mean. The values sum to \(4+6+7+8+10=35\) minutes. Dividing by the sample size gives \(\bar{x}=35/5=7\) minutes.
Find the sample standard deviation. The deviations from 7 are \(-3,-1,0,1,3\) minutes. Their squares sum to \(9+1+0+1+9=20\) square minutes. Using the sample formula:
Check the calculation another way. The squared observations sum to \(4^2+6^2+7^2+8^2+10^2=265\). Subtract \(n\bar{x}^2=5(7^2)=245\) to recover the sum of squared deviations: \(265-245=20\). Then \(\sqrt{20/4}=\sqrt{5}\approx2.236\) minutes, matching the first calculation.
Interpret. The sample standard deviation of response times for these help-desk chats is about 2.236 minutes. In context, the response times typically differ from their sample mean of 7 minutes by about 2.236 minutes. This does not say that every chat took a time exactly 2.236 minutes from 7.
Worked Example: Errors in a Sample of Temperature Readings
A fictional technician checks five temperature readings by recording each reading’s error, defined as the reading minus the known reference temperature. The errors, in degrees Celsius, are \(2,-1,0,1,\) and \(3\). Find and interpret the sample standard deviation.
Find the mean error. The errors sum to \(2+(-1)+0+1+3=5\) degrees Celsius. Thus \(\bar{x}=5/5=1\) degree Celsius.
Find the sample standard deviation. The deviations from the mean error of 1 are \(1,-2,-1,0,\) and \(2\) degrees Celsius. Their squares sum to \(1+4+1+0+4=10\) square degrees Celsius. Therefore:
Check the calculation another way. The squared errors sum to \(2^2+(-1)^2+0^2+1^2+3^2=15\). Subtract \(n\bar{x}^2=5(1^2)=5\), giving a squared-deviation sum of \(15-5=10\). Then \(\sqrt{10/4}=\sqrt{2.5}\approx1.581\) degrees Celsius, as before.
Interpret carefully. In this sample, the temperature errors typically differ from their sample mean of 1 degree Celsius by about 1.581 degrees Celsius. The reference value of zero is not the mean error here. So this standard deviation describes spread around the mean error; by itself, it does not describe how far readings typically are from the reference temperature.
Worked Example: Equal Spread, Different Typical Times
Two fictional groups each record three delivery times, in minutes. Group A’s times are 10, 12, and 14; Group B’s are 20, 22, and 24. Calculate the sample mean and standard deviation for each group, then compare what the standard deviations say.
Group A. Its mean is \(\bar{x}_A=(10+12+14)/3=36/3=12\) minutes. The deviations from 12 are \(-2,0,2\), with squared deviations summing to 8. Therefore \(s_A=\sqrt{8/(3-1)}=\sqrt{4}=2\) minutes.
Check Group A. The squared times sum to \(100+144+196=440\), and \(n\bar{x}_A^2=3(144)=432\). The difference is \(440-432=8\); thus \(\sqrt{8/2}=2\) minutes.
Group B. Its mean is \(\bar{x}_B=(20+22+24)/3=66/3=22\) minutes. The deviations from 22 are also \(-2,0,2\), so the squared deviations again sum to 8. Thus \(s_B=\sqrt{8/(3-1)}=2\) minutes.
Check Group B. Its squared times sum to \(400+484+576=1460\), and \(n\bar{x}_B^2=3(484)=1452\). The difference is \(1460-1452=8\); \(\sqrt{8/2}=2\) minutes.
Compare in context. In each group, delivery times in the sample typically differ from that group’s own mean by about 2 minutes. The equal standard deviations indicate equal measured spread for these two lists, while their means differ: 12 minutes for Group A and 22 minutes for Group B. Equal standard deviations do not mean the groups have equal typical delivery times.
Common Mistakes and Full-Credit Wording
- Leaving out the mean. “The standard deviation is 2.2 minutes” reports a value but does not interpret it. A complete statement says what variable and group it describes and identifies the mean as the reference point.
- Using the wrong reference point. Standard deviation describes distance from the mean, not from zero, a target, or the median. Name the mean in the interpretation, especially when the context includes a meaningful target.
- Claiming every observation is that distance away. “Every response time is 2.2 minutes from the mean” is not what a standard deviation says. Use “typically about” and avoid describing every individual value.
- Calling it an average absolute distance. That wording describes a different calculation. For AP Statistics, use the standard interpretation: a typical distance from the mean.
- Using the wrong units. The standard deviation has the same units as the observations. For example, if delivery times are recorded in minutes, report minutes—not square minutes.
- Claiming a percentage or distribution shape from the value alone. A standard deviation does not, by itself, tell you the fraction of observations within a stated distance or whether the distribution is symmetric. Support those claims with appropriate information about the distribution.
For full-credit communication, adapt the model sentence to the actual setting. State the group and variable, report the mean and standard deviation with units, and describe the standard deviation as a typical distance from that mean. Keep the interpretation proportional to what the statistic tells you.
Check Your Understanding
For each question, focus on the mean as the reference point and use the variable’s units.
- A sample of practice runs has a mean time of 48 seconds and a standard deviation of 3 seconds. Write a context-based interpretation.
- A set of measurements has mean 15 centimeters and standard deviation 2 centimeters. Does the standard deviation say that every measurement is exactly 2 centimeters from 15? Explain.
- Five sample errors have mean \(0.4\) degrees and standard deviation \(1.1\) degrees. What value is the reference point for interpreting the standard deviation: zero or \(0.4\) degrees? Write an interpretation.
- Two groups have the same standard deviation but different means. What aspect of the groups is similar, and what aspect may differ?
- A student says, “The standard deviation is 4 minutes, so 4% of the observations are within 4 minutes of the mean.” Identify the error in this statement.