OpenStax Introductory Statistics: Measures of the Spread of the Data

Fundamental Concepts of Standard Deviation and Spread

  • Learning Objectives:

    • Recognize, describe, and calculate key measures of the spread of data: variance, standard deviation, and range.

  • Standard Deviation Definition and Core Characteristics:

    • The standard deviation is a numerical value measuring how far data points in a distribution are located relative to their mean.

    • It provides an overall quantitative measure of total variation present within a data set.

    • It serves as a metric to determine whether a given data value lies close to or far from the mean.

    • Standard deviation is always non-negative (s≥0s \ge 0 for samples and σ≥0\sigma \ge 0 for populations).

    • A small standard deviation indicates that data values are tightly clustered near the mean, demonstrating minimal variation or spread.

    • A large standard deviation indicates that data values are widely dispersed away from the mean, demonstrating higher variation or spread.

  • Supermarket Waiting Time Comparative Example:

    • Supermarket A: Mean wait time μ=5 minutes\mu = 5\,\text{minutes}, standard deviation σ=2 minutes\sigma = 2\,\text{minutes}.

    • Supermarket B: Mean wait time μ=5 minutes\mu = 5\,\text{minutes}, standard deviation σ=4 minutes\sigma = 4\,\text{minutes}.

    • Analytical Comparison: Although both supermarkets share an identical average customer wait time of 5 minutes5\,\text{minutes}, Supermarket B exhibits greater variability in wait times due to its larger standard deviation (4 minutes4\,\text{minutes} versus 2 minutes2\,\text{minutes}). Customers at Supermarket A experience wait times concentrated much closer to the average.

Standard Deviation as a Distance Measure from the Mean

  • Measuring Relative Distance:

    • Any data value xx can be expressed in terms of standard deviations relative to the mean using the linear relationship:     value=mean+(# of STDEVs)(standard deviation)\text{value} = \text{mean} + (\text{\# of STDEVs})(\text{standard deviation})

    • Sample expression:     x=xˉ+(# of STDEV)(s)x = \bar{x} + (\text{\# of STDEV})(s)

    • Population expression:     x=μ+(# of STDEV)(σ)x = \mu + (\text{\# of STDEV})(\sigma)

    • The multiplier # of STDEV\text{\# of STDEV} represents the exact number of standard deviations xx lies from the mean and does not need to be an integer.

    • A positive value for # of STDEV\text{\# of STDEV} places the data point to the right of (above) the mean on a number line, whereas a negative value places it to the left of (below) the mean.

  • Thresholds and Rules of Thumb for Extremity:

    • A data value located exactly 22 standard deviations away from the mean represents a general threshold for borderline extreme values.

    • Classifying a value as far from the mean when it is more than 22 standard deviations away is a practical rule of thumb rather than a rigid rule.

    • The precise proportion of a dataset falling beyond 22 standard deviations depends heavily on the overall shape of the data distribution.

  • Detailed Case Analysis at Supermarket A:

    • Baseline parameters: Mean xˉ=5 minutes\bar{x} = 5\,\text{minutes}, standard deviation s=2 minutess = 2\,\text{minutes}.

    • Rosa's Wait Time (7 minutes7\,\text{minutes}):

    • Difference from mean: 7−5=2 minutes7 - 5 = 2\,\text{minutes}.

    • Deviation in standard units: 2 minutes=1 standard deviation2\,\text{minutes} = 1\,\text{standard deviation}.

    • Equation: 7=5+(1)(2)7 = 5 + (1)(2).

    • Conclusion: Rosa's wait time is 1 standard deviation1\,\text{standard deviation} above (to the right of) the mean.

    • Ben's Wait Time (1 minute1\,\text{minute}):

    • Difference from mean: 1−5=−4 minutes1 - 5 = -4\,\text{minutes}.

    • Deviation in standard units: 4 minutes=2 standard deviations4\,\text{minutes} = 2\,\text{standard deviations}.

    • Equation: 1=5+(−2)(2)1 = 5 + (-2)(2).

    • Conclusion: Ben's wait time is 2 standard deviations2\,\text{standard deviations} below (to the left of) the mean, placing it on the borderline for being considered far from the average.

Population vs. Sample Statistics and Notation

  • Concept of Deviations:

    • A deviation is defined as the signed difference between an individual data point xx and the set mean: x−meanx - \text{mean}.

    • Every single data point in a dataset yields its own corresponding deviation value; thus, there are as many deviations as there are data items.

    • Population deviation notation: x−μx - \mu

    • Sample deviation notation: x−xˉx - \bar{x}

  • Standard Deviation and Variance Notation:

    • Variance represents the average of the squared deviations from the mean.

    • Standard deviation is the square root of variance.

Notation summary table for population vs sample statistics
  • Distinction Between Population and Sample Calculations:

    • Population Calculations: Use total population count NN in the denominator when calculating average squared deviation.

    • Sample Calculations: Use degrees of freedom n−1n - 1 (sample size minus 1) in the denominator to correct for sample estimation bias.

Mathematical Formulas for Variance and Standard Deviation

  • Sample Standard Deviation Formulas:

    • Standard formula for raw individual data:     s=∑(x−xˉ)2n−1s = \sqrt{\frac{\sum (x - \bar{x})^2}{n - 1}}

    • Formula for frequency-grouped data:     s=∑f(x−xˉ)2n−1s = \sqrt{\frac{\sum f(x - \bar{x})^2}{n - 1}}

    • Sample variance formula:     s2=∑f(x−xˉ)2n−1s^2 = \frac{\sum f(x - \bar{x})^2}{n - 1}

    • Key note: The denominator for sample standard deviation and variance is explicitly n−1n - 1

  • Population Standard Deviation Formulas:

    • Standard formula for raw individual data:     σ=∑(x−μ)2N\sigma = \sqrt{\frac{\sum (x - \mu)^2}{N}}

    • Formula for frequency-grouped data:     σ=∑f(x−μ)2N\sigma = \sqrt{\frac{\sum f(x - \mu)^2}{N}}

    • Population variance formula:     σ2=∑f(x−μ)2N\sigma^2 = \frac{\sum f(x - \mu)^2}{N}

    • Key note: The denominator for population standard deviation and variance is explicitly NN

  • Role of Frequency (ff):

    • The term ff denotes the absolute frequency with which a distinct data value xx appears in the data set.

    • If a value appears once, f=1f = 1. If a value appears three times, f=3f = 3

Worked Examples and Detailed Step-by-Step Calculations

  • Example 1: Small Population Dataset:

    • Dataset representing a complete population (N=6N = 6): 10,60,50,30,40,2010, 60, 50, 30, 40, 20

    • Step 1: Compute Population Mean (μ\mu):     μ=10+60+50+30+40+206=2106=35\mu = \frac{10 + 60 + 50 + 30 + 40 + 20}{6} = \frac{210}{6} = 35

    • Step 2: Compute Squared Deviations and Sum:

    • (10−35)2=(−25)2=625(10 - 35)^2 = (-25)^2 = 625

    • (60−35)2=(25)2=625(60 - 35)^2 = (25)^2 = 625

    • (50−35)2=(15)2=225(50 - 35)^2 = (15)^2 = 225

    • (30−35)2=(−5)2=25(30 - 35)^2 = (-5)^2 = 25

    • (40−35)2=(5)2=25(40 - 35)^2 = (5)^2 = 25

    • (20−35)2=(−15)2=225(20 - 35)^2 = (-15)^2 = 225

    • Total sum of squared deviations: 625+625+225+25+25+225=1750625 + 625 + 225 + 25 + 25 + 225 = 1750

    • Step 3: Compute Population Variance (σ2\sigma^2):     σ2=17506≈291.67\sigma^2 = \frac{1750}{6} \approx 291.67

    • Step 4: Compute Population Standard Deviation (σ\sigma):     σ=σ2=291.67≈17.1\sigma = \sqrt{\sigma^2} = \sqrt{291.67} \approx 17.1

  • Example 2.32: Sample of Fifth Grade Student Ages:

    • Dataset representing a sample of n=20n = 20 fifth grade students (ages rounded to nearest half year):     9,9.5,9.5,10,10,10,10,10.5,10.5,10.5,10.5,11,11,11,11,11,11,11.5,11.5,11.59, 9.5, 9.5, 10, 10, 10, 10, 10.5, 10.5, 10.5, 10.5, 11, 11, 11, 11, 11, 11, 11.5, 11.5, 11.5

    • Step 1: Compute Sample Mean (xˉ\bar{x}):     xˉ=∑(x⋅f)20=10.525≈10.53 years\bar{x} = \frac{\sum (x \cdot f)}{20} = 10.525 \approx 10.53\,\text{years}

    • Step 2: Tabulate Deviations and Squared Deviations:

    • Data x=9x = 9, frequency f=1f = 1:

      • Deviation x−xˉ=9−10.525=−1.525x - \bar{x} = 9 - 10.525 = -1.525

      • Squared deviation (x−xˉ)2=(−1.525)2=2.325625(x - \bar{x})^2 = (-1.525)^2 = 2.325625

      • Weighted product f(x−xˉ)2=1×2.325625=2.325625f(x - \bar{x})^2 = 1 \times 2.325625 = 2.325625

    • Data x=9.5x = 9.5, frequency f=2f = 2:

      • Deviation x−xˉ=9.5−10.525=−1.025x - \bar{x} = 9.5 - 10.525 = -1.025

      • Squared deviation (x−xˉ)2=(−1.025)2=1.050625(x - \bar{x})^2 = (-1.025)^2 = 1.050625

      • Weighted product f(x−xˉ)2=2×1.050625=2.101250f(x - \bar{x})^2 = 2 \times 1.050625 = 2.101250

    • Data x=10x = 10, frequency f=4f = 4:

      • Deviation x−xˉ=10−10.525=−0.525x - \bar{x} = 10 - 10.525 = -0.525

      • Squared deviation (x−xˉ)2=(−0.525)2=0.275625(x - \bar{x})^2 = (-0.525)^2 = 0.275625

      • Weighted product f(x−xˉ)2=4×0.275625=1.102500f(x - \bar{x})^2 = 4 \times 0.275625 = 1.102500

    • Data x=10.5x = 10.5, frequency f=4f = 4:

      • Deviation x−xˉ=10.5−10.525=−0.025x - \bar{x} = 10.5 - 10.525 = -0.025

      • Squared deviation (x−xˉ)2=(−0.025)2=0.000625(x - \bar{x})^2 = (-0.025)^2 = 0.000625

      • Weighted product f(x−xˉ)2=4×0.000625=0.002500f(x - \bar{x})^2 = 4 \times 0.000625 = 0.002500

    • Data x=11x = 11, frequency f=6f = 6:

      • Deviation x−xˉ=11−10.525=0.475x - \bar{x} = 11 - 10.525 = 0.475

      • Squared deviation (x−xˉ)2=(0.475)2=0.225625(x - \bar{x})^2 = (0.475)^2 = 0.225625

      • Weighted product f(x−xˉ)2=6×0.225625=1.353750f(x - \bar{x})^2 = 6 \times 0.225625 = 1.353750

    • Data x=11.5x = 11.5, frequency f=3f = 3:

      • Deviation x−xˉ=11.5−10.525=0.975x - \bar{x} = 11.5 - 10.525 = 0.975

      • Squared deviation (x−xˉ)2=(0.975)2=0.950625(x - \bar{x})^2 = (0.975)^2 = 0.950625

      • Weighted product f(x−xˉ)2=3×0.950625=2.851875f(x - \bar{x})^2 = 3 \times 0.950625 = 2.851875

    • Sum of weighted squared deviations: ∑f(x−xˉ)2=9.7375\sum f(x - \bar{x})^2 = 9.7375

    • Step 3: Compute Sample Variance (s2s^2):     s2=∑f(x−xˉ)2n−1=9.737520−1=9.737519=0.5125s^2 = \frac{\sum f(x - \bar{x})^2}{n - 1} = \frac{9.7375}{20 - 1} = \frac{9.7375}{19} = 0.5125

    • Step 4: Compute Sample Standard Deviation (ss):     s=0.5125≈0.715891≈0.72 yearss = \sqrt{0.5125} \approx 0.715891 \approx 0.72\,\text{years}

    • Step 5: Calculation of Standard Deviation Intervals:

    • One standard deviation above the mean:       xˉ+1s=10.53+(1)(0.72)=11.25\bar{x} + 1s = 10.53 + (1)(0.72) = 11.25

    • Two standard deviations below the mean:       xˉ−2s=10.53−(2)(0.72)=9.09\bar{x} - 2s = 10.53 - (2)(0.72) = 9.09

    • 1.51.5 standard deviations below the mean:       xˉ−1.5s=10.53−(1.5)(0.72)=9.45\bar{x} - 1.5s = 10.53 - (1.5)(0.72) = 9.45

    • 1.51.5 standard deviations above the mean:       xˉ+1.5s=10.53+(1.5)(0.72)=11.61\bar{x} + 1.5s = 10.53 + (1.5)(0.72) = 11.61

  • WebAssign Homework Example: Student Movie Watching Frequency:

    • Sample size n=25n = 25 randomly selected students asked about movies watched in the previous week.

    • Frequency Distribution:

    • 0 movies:f=50\,\text{movies}: f = 5

    • 1 movie:f=91\,\text{movie}: f = 9

    • 2 movies:f=72\,\text{movies}: f = 7

    • 3 movies:f=33\,\text{movies}: f = 3

    • 4 movies:f=14\,\text{movies}: f = 1

    • Step 1: Calculate Sample Mean (xˉ\bar{x}):     xˉ=(0⋅5)+(1⋅9)+(2⋅7)+(3⋅3)+(4⋅1)25=0+9+14+9+425=3625=1.44 movies\bar{x} = \frac{(0 \cdot 5) + (1 \cdot 9) + (2 \cdot 7) + (3 \cdot 3) + (4 \cdot 1)}{25} = \frac{0 + 9 + 14 + 9 + 4}{25} = \frac{36}{25} = 1.44\,\text{movies}

    • Step 2: Calculate Weighted Sum of Squared Deviations:     ∑f(x−xˉ)2=5(0−1.44)2+9(1−1.44)2+7(2−1.44)2+3(3−1.44)2+1(4−1.44)2\sum f(x - \bar{x})^2 = 5(0 - 1.44)^2 + 9(1 - 1.44)^2 + 7(2 - 1.44)^2 + 3(3 - 1.44)^2 + 1(4 - 1.44)^2     =5(−1.44)2+9(−0.44)2+7(0.56)2+3(1.56)2+1(2.56)2= 5(-1.44)^2 + 9(-0.44)^2 + 7(0.56)^2 + 3(1.56)^2 + 1(2.56)^2     =5(2.0736)+9(0.1936)+7(0.3136)+3(2.4336)+1(6.5536)= 5(2.0736) + 9(0.1936) + 7(0.3136) + 3(2.4336) + 1(6.5536)     =10.368+1.7424+2.1952+7.3008+6.5536=28.16= 10.368 + 1.7424 + 2.1952 + 7.3008 + 6.5536 = 28.16

    • Step 3: Calculate Sample Variance (s2s^2):     s2=28.1625−1=28.1624=1.173333333s^2 = \frac{28.16}{25 - 1} = \frac{28.16}{24} = 1.173333333

    • Step 4: Calculate Sample Standard Deviation (ss):     s=1.173333333≈1.08320512≈1.08 moviess = \sqrt{1.173333333} \approx 1.08320512 \approx 1.08\,\text{movies}