Comprehensive Study Notes: Elementary Statistics – Data Collection, Representation, Central Tendency, and Variation

Foundations of Statistics and Data Classification

  • Definition of Data:

    • Data consist of information coming from observations, counts, measurements, or responses.
    • Real-world baseline examples illustrating data:
    • In modern digital connectivity, most U.S. adults report using the internet (95%95\%), owning a smartphone (90%90\%), or subscribing to home broadband (80%80\%).
    • Community cultural assessments indicate that 7 in 107\text{ in }10 Americans believe the arts unify their communities, and 2 in 52\text{ in }5 Americans have altered an opinion or perception based on an arts experience.
    • Demographic studies on youth technology adoption show that 21%21\% of children aged 8 to 118\text{ to }11 possess a personal social media profile.
  • Definition and Scope of Statistics:

    • Statistics is the science of collecting, organizing, analyzing, and interpreting data in order to make informed decisions.
    • Four sequential stages of a statistical workflow:
    • Collecting Data: Gathering observations through instruments such as surveys, experimental measurements, or archival searches.
    • Characterizing Data: Summarizing features using descriptive calculations such as the mean, median, or standard deviation.
    • Analyzing Data: Examining relationships, trends, distributions, and significant patterns.
    • Interpreting Data: Drawing valid conclusions and making decisions or predictions concerning a broader population.
  • Statistical Data Sets:

    • Population: The collection of all outcomes, responses, measurements, or counts that are of interest for a given inquiry.
    • Sample: A subset, or part, of the population selected for analysis.
    • Identifying Populations and Samples in Practice:
    • Electric Vehicle (EV) Interest Study:
      • Context: In a study of 11521152 adults in the United States, 207207 reported being likely to purchase an EV.
      • Target Population: The responses of all adults residing in the United States.
      • Sample: The responses of the 11521152 adults surveyed.
      • Sample Composition: Composed of 207207 respondents likely to buy an EV and 945945 respondents unlikely to buy an EV.
    • Workplace Stress Survey:
      • Context: In a survey of 834834 employees in the United States, 517517 stated that their jobs were highly stressful.
      • Target Population: The responses of all employees in the United States.
      • Sample: The responses of the 834834 employees surveyed.
      • Sample Composition: Composed of 517517 affirmative ("yes") responses and 317317 negative ("no") responses.
  • Parameters versus Statistics:

    • Parameter: A numerical description of a population characteristic.
    • Mnemonic aid: Both Population and Parameter begin with the letter P.
    • Example: The entire freshman class at a university has an average SAT math score of 514514. Because this measurement covers the entire freshman cohort, 514514 is a population parameter.
    • Statistic: A numerical description of a sample characteristic.
    • Mnemonic aid: Both Sample and Statistic begin with the letter S.
    • Example 1: In a survey of 27982798 U.S. households, the estimated average weekly grocery expenditure is "$165""\$165". Because this average reflects a subset of all U.S. households, it is a sample statistic.
    • Example 2: In a random inspection of several hundred retail grocery stores, 34%34\% failed to store fish at required food safety temperatures. Because this percentage was derived from a subset of stores, it represents a sample statistic.
  • Branches of Statistics:

    • Descriptive Statistics:
    • Focuses on the organization, summarization, and presentation of data.
    • Employs graphical displays, frequency tables, and summary numerical metrics.
    • Inferential Statistics:
    • Involves utilizing sample findings to draw conclusions, establish estimates, make predictions, and make generalizations about the larger population.
    • Practical Distinctions between Descriptive and Inferential Branches:
    • Relocation Cost Study:
      • Study Context: An investigation of 10001000 U.S. adults who moved within the prior five years determined that the average cost to relocate locally was "$1692""\$1692", while the average cost for a long-distance relocation was "$4401""\$4401".
      • Population: All U.S. adults who relocated within the past five years.
      • Sample: The 10001000 relocated adults who completed the study.
      • Descriptive Component: The reported average figures of "$1692""\$1692" and "$4401""\$4401" and their associated graphical bar charts.
      • Inferential Extension: Concluding that a long-distance move in the United States generally costs at least 2.52.5 times more than a local move.
    • Workplace Mental Health Disclosure Study:
      • Study Context: In a sample of 15411541 U.S. workers, 39%39\% reported worrying that disclosing a mental health condition to their employer would negatively affect their workplace standing.
      • Population: All employed workers in the United States.
      • Sample: The 15411541 U.S. workers who participated in the poll.
      • Descriptive Component: The calculation that 39%39\% of the sample expressed concern.
      • Inferential Extension: Inferring that nearly two in five U.S. workers fear negative career consequences if they disclose a mental health condition to their employers.
    • Income and Digital Access Study:
      • Study Context: In a sample of 15021502 U.S. adults, 18%18\% of individuals living in households earning under "$30,000""\$30,000" annually do not use the internet.
      • Population: All adults in the United States.
      • Sample: The 15021502 surveyed adults.
      • Descriptive Component: The statement that 18%18\% of respondents in the sub-"$30,000""\$30,000" income bracket do not access the internet.
      • Inferential Extension: Concluding that lower household income creates systemic barriers to digital accessibility.
    • Retirement Longevity Awareness Study:
      • Study Context: Among 10001000 surveyed U.S. 401(k) plan participants, 32%32\% did not know how many years their accumulated retirement savings would last.
      • Population: All active U.S. 401(k) retirement plan participants.
      • Sample: The 10001000 participants surveyed.
      • Descriptive Component: The measured statistic that 32%32\% lack knowledge regarding their savings' duration.
      • Inferential Extension: Inferring that calculating required retirement funds and personal financial longevity is fundamentally difficult for everyday investors.

Data Collection and Experimental Design

  • Steps for Designing a Statistical Study:

    • 1. Identify the variable(s) of interest (the primary focus) and define the target population clearly.
    • 2. Develop a comprehensive, unbiased plan for collecting data; verify that any sample selected represents the population.
    • 3. Collect the data systematically.
    • 4. Summarize and display the data using descriptive statistics methods.
    • 5. Interpret the data and make inferences or decisions about the population using inferential techniques.
    • 6. Identify and document any possible sources of error, bias, or procedural limitations.
  • Methods of Data Collection:

    • Observational Study:
    • A researcher observes and measures characteristics of interest within a sample without modifying existing conditions or introducing interventions.
    • Data can be cross-sectional (measured at a single time point), historical (derived from past records), or longitudinal (tracked across time).
    • Example: An economic survey calling 10191019 U.S. adults to rate current national economic conditions and evaluate whether conditions are improving or deteriorating without attempting to influence their answers.
    • Experiment:
    • A specific intervention (treatment) is applied to an experimental group (treatment group), and the resulting responses are observed.
    • A separate control group is maintained without the active intervention to provide a baseline for comparison. Control subjects are frequently administered an inert treatment known as a placebo.
    • The individuals or entities undergoing the conditions are called experimental units (or human subjects).
    • Example: An investigation evaluating Vitamin D3D_3 supplementation in newly diagnosed viral infection patients, where 27002700 adults are administered either daily Vitamin D3D_3 or an identical placebo for four weeks.
    • Simulation:
    • Employs mathematical, computer, or physical models to recreate the operational conditions of a system or process.
    • Frequently utilized when real-world testing is dangerous, impractical, unethical, or cost-prohibitive.
    • Example: Automotive engineering firms using computerized models and crash test dummies to assess human vehicular impact forces.
    • Survey:
    • An investigation into one or more characteristics of a population conducted by asking questions of human subjects.
    • Channels include personal interviews, online questionnaires, telephone calls, or mailed forms.
    • Design consideration: Survey questions must be formulated neutrally to prevent leading wording that introduces bias.
    • Example: Questioning a sample of physicians from a standardized list of career motivators to establish whether financial stability served as their primary incentive.
  • Key Principles of Experimental Design:

    • The three fundamental pillars of experimental integrity are control, randomization, and replication.
    • Control of Confounding Variables:
    • A confounding variable occurs when a researcher cannot separate the individual effects of two or more distinct factors on the observed response variable.
    • Example: A local retail coffee shop undergoes an interior remodel at the exact time a major new shopping center opens next door; an ensuing spike in sales revenue cannot be clearly attributed to the remodel versus the new mall traffic.
    • Controlling Bias via the Placebo Effect and Blinding:
    • Placebo Effect: Occurs when an untreated subject demonstrates a favorable psychological or physiological response merely because they believe they are receiving active treatment.
    • Blinding: A technique where human subjects are kept unaware of whether they are assigned to the active treatment or the placebo group.
    • Double-Blind Experiment: An experimental protocol where neither the participating subjects nor the research personnel administering treatments know which group is receiving the active treatment versus the placebo.
    • Randomization Strategies:
    • Randomization is the process of allocating subjects into distinct treatment groups strictly through chance.
    • Completely Randomized Design: Subjects are assigned to experimental groups through pure random selection.
    • Randomized Block Design: Subjects possessing similar underlying traits (e.g., age, sex, risk status) are first partitioned into blocks. Within each block, subjects are then randomly assigned to the treatment or control groups.
    • Example: A clinical trial testing a dietary weight-loss beverage partitions participants into age blocks (30–3930\text{--}39, 40–4940\text{--}49, and over 5050 years old), followed by random allocation to either the active drink or the control within each age bracket.

Randomized Block Design diagram

  • Matched-Pairs Design:

    • Subjects are paired according to specific matching attributes (e.g., identical twins, similar baselines). One subject from the pair is randomly assigned to treatment while the other receives the alternative condition.
    • Alternatively, the same subject serves as their own matched control in repeated-measures pre-test and post-test designs.
  • Replication and Sample Size:

    • Sample size denotes the total count of experimental units evaluated in a study.
    • Replication is the repetition of an experiment under identical or closely matched conditions.
    • Maintaining a large sample size via replication ensures that experimental outcomes can be validated against random variation.
  • Critical Analysis of Flawed Experimental Designs:

    • Flaw Type 1 (Sample Size Deficiency):

      • Flawed Setup: Testing smoking cessation gum on 1010 heavy smokers, where 55 receive the gum and 55 receive a placebo; all 55 gum users quit smoking within two months.
      • Methodological Failure: A sample size of n=10n = 10 is far too small to confirm efficacy or rule out chance.
      • Necessary Correction: The experiment must be replicated across a significantly larger sample cohort.
    • Flaw Type 2 (Confounding Factors via Non-Random Blocking):

      • Flawed Setup: In a study of 10001000 smokers, all participants aged 18 to 3918\text{ to }39 receive active cessation gum, while all participants over age 3939 receive a placebo.
      • Methodological Failure: Treatment is confounded with subject age; the experiment cannot determine whether higher cessation in the younger group stems from the gum or younger age.
      • Necessary Correction: Subjects must be blocked by age, and then randomly allocated to active treatment or placebo within each block.
    • Sampling Techniques:

  • Census versus Sampling:

    • Census: An exhaustive enumeration and measurement of every single element in an entire population.
    • Sampling: The collection of data from a designated subset of the population.
    • Sampling Error: The unavoidable mathematical difference between the findings obtained from a sample statistic and the true population parameter.
  • Random Sample:

    • A sample where every individual member of the population possesses an equal probability of being chosen.
  • Simple Random Sample (SRS):

    • A sample where every possible sample of the specified size nn has the exact same probability of selection.
    • Mechanical Execution via Number Generators and Random Number Tables:
      • Each population member is assigned an identifying number from 1 to N1\text{ to }N.
      • Numbers are generated using calculators, computerized spreadsheets, or standardized random number tables.
      • Example: Selecting an SRS of n=8n = 8 from a roster of 731731 first-year college students.
      • Procedure: Digits from a random number table are read in three-digit groupings. Digits exceeding 731731 or duplicate numbers are bypassed until eight unique identifiers are found.
      • Selected student identifiers: 719,662,650,4,53,589,403,129719, 662, 650, 4, 53, 589, 403, 129.
  • Stratified Sample:

    • The population is divided into two or more homogeneous subsets called strata based on specific characteristics (e.g., socioeconomic bracket, department, major).
    • A random sample is drawn from every individual stratum in proportion to its size.

Stratified sampling diagram

  • Cluster Sample:
    • The population is naturally partitioned into geographic or administrative subsets called clusters (e.g., city blocks, school districts, zip codes).
    • Several clusters are randomly selected, and every individual member within those selected clusters is sampled, while unselected clusters remain completely unmeasured.

Cluster sampling zip code zones diagram

  • Systematic Sample:
    • Population items are ordered in a sequential list.
    • A random starting entry is selected, and every kthk\text{th} element is systematically included.
    • Example: In West Ridge County, households are assigned numbers, a random starting value of 8686 is drawn, and every 100th100\text{th} household is sampled (86,186,286,386,48686, 186, 286, 386, 486).

Systematic sampling household selection diagram

  • Convenience Sample:
    • Selects available subjects who are easily accessible.
    • High susceptibility to selection bias; not recommended for rigorous statistical inference.
  • Real-World Sampling Evaluation Examples:
    • Academic Major Stratification: Partitioning a college student body by declared academic majors and randomly sampling students from each major constitutes a stratified sample.
    • Random Number Identification: Assigning all enrolled students a numerical ID and using a random number generator to pick participants constitutes a simple random sample.
    • Classroom Sampling: Surveying students enrolled in one's personal biology section regarding opinions on stem cell research constitutes a convenience sample; it introduces bias because biology students likely possess disproportionate background knowledge and stronger viewpoints than the general student body.

Graphical Displays for Data Sets

  • Stem-and-Leaf Plots:
    • A quantitative graphing method where each data point is separated into a stem (the leading digits) and a leaf (the final rightmost digit).
    • Functions similarly to a histogram turned on its side, but preserves the original raw data values.
    • Standard Construction Procedure:
    • Identify the minimum and maximum data values to define the range of stems.
    • Align stems vertically to the left of a continuous vertical boundary line.
    • For every data observation, record its trailing digit (leaf) in the corresponding stem row to the right of the line.
    • Provide an explicit key (e.g., Key: 10∣4=10410 | 4 = 104).
    • Worked Example (Daily Text Messages Sent by 5050 Adults):
    • Raw Data Array (n=50n = 50):       754910459881237510968816680786955761149873184284465225252633252032244317492732292940233330413538365430148\begin{matrix} 75 & 49 & 104 & 59 & 88 \\ 123 & 75 & 109 & 68 & 81 \\ 66 & 80 & 78 & 69 & 55 \\ 76 & 114 & 98 & 73 & 18 \\ 42 & 84 & 46 & 52 & 25 \\ 25 & 26 & 33 & 25 & 20 \\ 32 & 24 & 43 & 17 & 49 \\ 27 & 32 & 29 & 29 & 40 \\ 23 & 33 & 30 & 41 & 35 \\ 38 & 36 & 54 & 30 & 148 \end{matrix}
    • Values span from a low of 1717 to a high of 148148, requiring stems running from 11 to 1414.
    • Standard Stem-and-Leaf Representation:

Stem-and-leaf plot showing text messages sent

* Analytic Insight: More than 50%50\% of surveyed cell phone users sent between 2020 and 5050 text messages per day.
  • Variations: Split-Stem Plots (Two Rows per Stem):

    • Used to provide a more detailed visual picture when data cluster densely.

    • Each stem value is listed twice:

      • First stem row contains leaves 0,1,2,3,40, 1, 2, 3, 4.
      • Second stem row contains leaves 5,6,7,8,95, 6, 7, 8, 9.
    • Demonstrates that the vast majority of U.S. adults sent between 2020 and 8080 text messages per day.

    • Dot Plots:

  • A quantitative display where each numerical observation is positioned as a point (dot) above its value on a horizontal axis.

  • Highlights clusters, gaps, spread, and isolated values while retaining all individual values.

  • Application to Text Messaging Data:

    • Constructed across a horizontal continuum spanning from 1515 to 150150.

Dot plot showing distribution of text messages sent

* Visual Insights: Shows dense grouping between 2020 and 8080, reveals that only 55 respondents sent over 100100 text messages, and clearly identifies 148148 as an extreme outlier.
  • Pie Charts:
    • Circular displays designed for qualitative data, showing categories as proportional slices (sectors) of a whole (100%100\%).
    • Formula for Central Angle of a Sector:     Central Angle=360∘×(Relative Frequency)\text{Central Angle} = 360^\circ \times (\text{Relative Frequency})
    • Worked Example (Earned Degrees Conferred in the U.S. in 2022):
    • Degree distribution in thousands:
      • Associate's: f=1008f = 1008
      • Bachelor's: f=2015f = 2015
      • Master's: f=880f = 880
      • Doctoral: f=204f = 204
    • Total degrees conferred (nn):       n=1008+2015+880+204=4107 thousandn = 1008 + 2015 + 880 + 204 = 4107\text{ thousand}
    • Relative frequency and angle calculations:
      • Associate's:         Relative Frequency=10084107≈0.2454  ⟹  24.5%\text{Relative Frequency} = \frac{1008}{4107} \approx 0.2454 \implies 24.5\%\text{Central Angle} = 360^\circ \times 0.2454 \approx 88.2^\circ\n * Bachelor's:\n        \text{Relative Frequency} = \frac{2015}{4107} \approx 0.4906 \implies 49.1\%Central Angle=360∘×0.4906≈176.8∘\text{Central Angle} = 360^\circ \times 0.4906 \approx 176.8^\circ
      • Master's:         Relative Frequency=8804107≈0.2143  ⟹  21.4%\text{Relative Frequency} = \frac{880}{4107} \approx 0.2143 \implies 21.4\%\text{Central Angle} = 360^\circ \times 0.2143 \approx 77.0^\circ\n * Doctoral:\n        \text{Relative Frequency} = \frac{204}{4107} \approx 0.0497 \implies 5.0\%Central Angle=360∘×0.0497≈18.0∘\text{Central Angle} = 360^\circ \times 0.0497 \approx 18.0^\circ

Pie chart of earned degrees conferred in 2022

* Key Finding: Almost half (49.1%49.1\%) of all higher education degrees conferred in 2022 were bachelor's degrees.
  • Pareto Charts:
    • A vertical bar graph for qualitative data where bar height represents frequency or relative frequency.
    • Ordering rule: Bars are arranged in descending order from left to right, placing the most frequent category at the far left.
    • Worked Example (Top Five Causes of Death in the U.S., 2022):
    • Data breakdown:
      • Heart disease: 702,880702,880
      • Cancer: 608,371608,371
      • Accidents: 227,039227,039
      • COVID-19: 186,552186,552
      • Stroke (cerebrovascular diseases): 165,393165,393

Pareto chart of top five causes of death in the United States

* Analytical Interpretation: Heart disease was the leading cause of death in the United States in 2022. Heart disease and cancer combined caused more deaths than accidents, COVID-19, and stroke combined.
  • Displays for Paired Data Sets:
    • Scatter Plots:
    • Graph paired quantitative data points (x,y)(x, y) on a Cartesian coordinate grid.
    • Used to examine the direction, form, and strength of the relationship between two variables.
    • Example: Ronald Fisher's 1936 Iris Data Set:
      • Examines petal length (xx, in millimeters) versus petal width (yy, in millimeters) across three species of iris.

Scatter plot of Fisher's Iris data set showing petal length versus petal width

* Finding: As petal length increases, petal width tends to increase, showing a positive association.
  • Time Series Charts:
    • Displays quantitative data measured at regular intervals over time, plotting time on the horizontal axis and connecting points with line segments.
    • Example: U.S. Electric Vehicle (EV) Sales and Energy Demand (2013–2023):
    • 2013: 97 thousand97\text{ thousand} EVs sold; 560 GWh560\text{ GWh} electricity demand
    • 2014: 118 thousand118\text{ thousand} EVs sold; 840 GWh840\text{ GWh} electricity demand
    • 2015: 114 thousand114\text{ thousand} EVs sold; 1900 GWh1900\text{ GWh} electricity demand
    • 2016: 161 thousand161\text{ thousand} EVs sold; 3500 GWh3500\text{ GWh} electricity demand
    • 2017: 196 thousand196\text{ thousand} EVs sold; 3700 GWh3700\text{ GWh} electricity demand
    • 2018: 363 thousand363\text{ thousand} EVs sold; 5000 GWh5000\text{ GWh} electricity demand
    • 2019: 327 thousand327\text{ thousand} EVs sold; 6300 GWh6300\text{ GWh} electricity demand
    • 2020: 295 thousand295\text{ thousand} EVs sold; 5400 GWh5400\text{ GWh} electricity demand
    • 2021: 633 thousand633\text{ thousand} EVs sold; 7000 GWh7000\text{ GWh} electricity demand
    • 2022: 993 thousand993\text{ thousand} EVs sold; 7200 GWh7200\text{ GWh} electricity demand
    • 2023: 1393 thousand1393\text{ thousand} EVs sold; 22000 GWh22000\text{ GWh} electricity demand

Time series chart of electric vehicles sold from 2013 to 2023

  • Trend Analysis: Annual EV sales increased from 2013 through 2018, experienced a slight decline in 2019–2020, and expanded rapidly from 2021 through 2023.

Measures of Central Tendency

  • Core Measures of Central Tendency:

    • A measure of central tendency is a single numerical value that represents the center or typical value of a data distribution.
    • The three primary measures are the mean, the median, and the mode.
  • The Mean:

    • The sum of all data values divided by the total count of values.
    • Sigma Notation: ∑x\sum x denotes the sum of all individual data values xx.
    • Population Mean Formula:     μ=∑xN\mu = \frac{\sum x}{N}     where NN is the total number of entries in the population.
    • Sample Mean Formula:     x‾=∑xn\overline{x} = \frac{\sum x}{n}     where nn is the total number of entries in the sample.
    • Worked Example (Adult Pre-Study Weights in Pounds):
    • Raw data (n=7n = 7): 274,235,223,268,290,285,235274, 235, 223, 268, 290, 285, 235.
    • Sum of values:       ∑x=274+235+223+268+290+285+235=1810\sum x = 274 + 235 + 223 + 268 + 290 + 285 + 235 = 1810
    • Sample mean:       x‾=18107≈258.6 pounds\overline{x} = \frac{1810}{7} \approx 258.6\text{ pounds}
  • The Median:

    • The value that lies in the middle of a data set when the values are arranged in ascending or descending order.
    • Divides an ordered data set into two equal halves.
    • Odd Number of Entries: The median is the single middle data entry.
    • Even Number of Entries: The median is the mean of the two middle data entries.
    • Worked Examples:
    • Example 1 (Odd Sample Size, n=7n = 7):
      • Ordered data: 223,235,235,268,274,285,290223, 235, 235, 268, 274, 285, 290.
      • The median is the fourth entry: 268 pounds268\text{ pounds}.
    • Example 2 (Even Sample Size, n=6n = 6):
      • If the participant weighing 285 pounds285\text{ pounds} withdraws: 223,235,235,268,274,290223, 235, 235, 268, 274, 290.
      • The two middle entries are 235235 and 268268.
      • Calculation:         Median=235+2682=251.5 pounds\text{Median} = \frac{235 + 268}{2} = 251.5\text{ pounds}
  • The Mode:

    • The data entry that occurs with the greatest frequency.
    • Special Conditions: If no entry repeats, the data set has no mode. If two entries share the same highest frequency, the data set is bimodal.
    • Worked Examples:
    • Numerical Data: In the weights set (223,235,235,268,274,285,290223, 235, 235, 268, 274, 285, 290), 235235 occurs twice while all other entries occur once. Mode = 235 pounds235\text{ pounds}.
    • Categorical Data: At a political debate, audience members reported their party affiliations:
      • Democrat: f=46f = 46
      • Republican: f=34f = 34
      • Independent: f=39f = 39
      • Other / don't know: f=5f = 5
      • Mode = Democrat (the only measure of central tendency that can be used with nominal qualitative data).
  • Comparing the Mean, Median, and Mode:

    • Mean Advantages & Disadvantages:
    • Advantage: Reliable because it accounts for every entry in the data set.
    • Disadvantage: Sensitive to outliers (values far removed from the rest of the data).
    • Worked Example (Classroom Student Ages):
    • Data set (n=20n = 20): 20,20,20,20,20,20,21,21,21,21,22,22,22,23,23,23,23,24,24,6520, 20, 20, 20, 20, 20, 21, 21, 21, 21, 22, 22, 22, 23, 23, 23, 23, 24, 24, 65.
    • Mean:       x‾=47520=23.75 years\overline{x} = \frac{475}{20} = 23.75\text{ years}
    • Median (mean of 10th and 11th entries, 2121 and 2222):       Median=21+222=21.5 years\text{Median} = \frac{21 + 22}{2} = 21.5\text{ years}
    • Mode: 20 years20\text{ years} (occurs 66 times).

Histogram of student ages displaying outlier, gap, mean, median, and mode

* Comparison: The outlier (6565) pulls the mean up to 23.7523.75, which is higher than 1919 of the 2020 student ages. The mode (2020) is lower than most entries. The median (21.521.5) best describes the typical student age in this class.
  • Weighted Mean:

    • The mean of a data set whose entries have varying weights.
    • Formula:     x‾=∑(x⋅w)∑w\overline{x} = \frac{\sum (x \cdot w)}{\sum w}     where ww is the weight of each entry xx.
    • Worked Example (Grade Point Average Calculation):
    • Grading scale: A=4A = 4, B=3B = 3, C=2C = 2, D=1D = 1, F=0F = 0.
    • Course grades and credit hours:
      • Grade C (x=2x = 2), Credit hours w=3w = 3: x⋅w=6x \cdot w = 6
      • Grade C (x=2x = 2), Credit hours w=4w = 4: x⋅w=8x \cdot w = 8
      • Grade D (x=1x = 1), Credit hours w=1w = 1: x⋅w=1x \cdot w = 1
      • Grade A (x=4x = 4), Credit hours w=3w = 3: x⋅w=12x \cdot w = 12
      • Grade C (x=2x = 2), Credit hours w=2w = 2: x⋅w=4x \cdot w = 4
      • Grade B (x=3x = 3), Credit hours w=3w = 3: x⋅w=9x \cdot w = 9
    • Sum of weights:       ∑w=3+4+1+3+2+3=16\sum w = 3 + 4 + 1 + 3 + 2 + 3 = 16
    • Sum of weighted values:       ∑(x⋅w)=6+8+1+12+4+9=40\sum (x \cdot w) = 6 + 8 + 1 + 12 + 4 + 9 = 40
    • Weighted Mean (GPA):       x‾=4016=2.5\overline{x} = \frac{40}{16} = 2.5
  • Mean of Grouped Data (Frequency Distributions):

    • Approximates the sample mean for grouped data when original values are not accessible.
    • Formula:     x‾=∑(x⋅f)n\overline{x} = \frac{\sum (x \cdot f)}{n}     where xx is the class midpoint, ff is the class frequency, and n=∑fn = \sum f.
    • Class Midpoint Formula:     x=Lower Class Limit+Upper Class Limit2x = \frac{\text{Lower Class Limit} + \text{Upper Class Limit}}{2}
    • Step-by-Step Procedure:
    • 1. Calculate the midpoint xx for each class.
    • 2. Multiply each midpoint by its class frequency: x⋅fx \cdot f.
    • 3. Sum the products: ∑(x⋅f)\sum (x \cdot f).
    • 4. Sum the frequencies to find the sample size: n=∑fn = \sum f.
    • 5. Divide the sum of the products by the sample size: x‾=∑(x⋅f)n\overline{x} = \frac{\sum (x \cdot f)}{n}.
    • Worked Example (Daily Screen Time in Minutes for 3030 Adults):
    • Class Midpoint x=172.5x = 172.5, Frequency f=3f = 3: x⋅f=517.5x \cdot f = 517.5
    • Class Midpoint x=208.5x = 208.5, Frequency f=2f = 2: x⋅f=417.0x \cdot f = 417.0
    • Class Midpoint x=244.5x = 244.5, Frequency f=5f = 5: x⋅f=1222.5x \cdot f = 1222.5
    • Class Midpoint x=280.5x = 280.5, Frequency f=6f = 6: x⋅f=1683.0x \cdot f = 1683.0
    • Class Midpoint x=316.5x = 316.5, Frequency f=7f = 7: x⋅f=2215.5x \cdot f = 2215.5
    • Class Midpoint x=352.5x = 352.5, Frequency f=4f = 4: x⋅f=1410.0x \cdot f = 1410.0
    • Class Midpoint x=388.5x = 388.5, Frequency f=3f = 3: x⋅f=1165.5x \cdot f = 1165.5
    • Sum of frequencies: n=∑f=30n = \sum f = 30.
    • Sum of products: ∑(x⋅f)=8631.0\sum (x \cdot f) = 8631.0.
    • Estimated sample mean:       x‾=8631.030=287.7 minutes\overline{x} = \frac{8631.0}{30} = 287.7\text{ minutes}
    • Comparison: The mean calculated from the original raw data is 285.5 minutes285.5\text{ minutes}, demonstrating that grouped estimation provides a close approximation (2.2 minutes2.2\text{ minutes} difference).
  • Distribution Shapes:

    • Symmetric Distribution:
    • A vertical line drawn down the center divides the graph into approximate mirror images.
    • The mean, median, and mode are equal and located at the center.

Symmetric distribution histogram

  • Uniform Distribution (Rectangular):
    • All entries or classes have equal or approximately equal frequencies.
    • It is symmetric, with the mean and median located at the center.

Uniform distribution histogram

  • Skewed Left Distribution (Negatively Skewed):
    • The tail of the graph extends more to the left.
    • The mean is pulled to the left of the median:     Mean<Median≤Mode\text{Mean} < \text{Median} \le \text{Mode}

Skewed left distribution histogram

  • Skewed Right Distribution (Positively Skewed):
    • The tail of the graph extends more to the right.
    • The mean is pulled to the right of the median:     Mode≤Median<Mean\text{Mode} \le \text{Median} < \text{Mean}

Skewed right distribution histogram

Measures of Variation

  • The Range:
    • The difference between the maximum and minimum entries in a quantitative data set.
    • Formula:     Range=Maximum Entry−Minimum Entry\text{Range} = \text{Maximum Entry} - \text{Minimum Entry}
    • Worked Example (Graduate Starting Salaries in Thousands of Dollars):
    • Corporation A: 51,48,49,55,57,51,54,51,47,5251, 48, 49, 55, 57, 51, 54, 51, 47, 52
      • Ordered: 47,48,49,51,51,51,52,54,55,5747, 48, 49, 51, 51, 51, 52, 54, 55, 57
      • Range calculation:         Range=57−47=10 (or $10,000)\text{Range} = 57 - 47 = 10\text{ (or \$10,000)}
    • Corporation B: 50,33,51,60,59,42,51,39,62,6850, 33, 51, 60, 59, 42, 51, 39, 62, 68
      • Ordered: 33,39,42,50,51,51,59,60,62,6833, 39, 42, 50, 51, 51, 59, 60, 62, 68
      • Range calculation:         Range=68−33=35 (or $35,000)\text{Range} = 68 - 33 = 35\text{ (or \$35,000)}
    • Comparison: Both corporations share identical central tendencies (Mean = 51.551.5, Median = 5151, Mode = 5151), but Corporation B exhibits much greater variation.

Histogram of starting salaries for Corporation A

Histogram of starting salaries for Corporation B

  • Deviation, Variance, and Standard Deviation:

    • Deviation: The difference between a data entry xx and the mean of the data set:     Population Deviation=x−μ\text{Population Deviation} = x - \muSample Deviation=x−x‾\text{Sample Deviation} = x - \overline{x}
    • The sum of all individual deviations in a data set always equals zero:     ∑(x−μ)=0\sum (x - \mu) = 0
    • Population Variance and Standard Deviation:
    • Population Variance (σ2\sigma^2):       σ2=∑(x−μ)2N\sigma^2 = \frac{\sum (x - \mu)^2}{N}
    • Population Standard Deviation (σ\sigma):       σ=∑(x−μ)2N\sigma = \sqrt{\frac{\sum (x - \mu)^2}{N}}
    • Properties of Standard Deviation:
      • Measures the spread of data about the mean in the same units as the data set.
      • Always non-negative (σ≥0\sigma \ge 0); σ=0\sigma = 0 only when all data entries are identical.
      • Increases as entries become more spread out from the mean.
    • Sample Variance and Standard Deviation:
    • Sample Variance (s2s^2):       s2=∑(x−x‾)2n−1s^2 = \frac{\sum (x - \overline{x})^2}{n - 1}
      • Dividing by n−1n - 1 corrects for bias, ensuring s2s^2 is an unbiased estimator of σ2\sigma^2.
    • Sample Standard Deviation (ss):       s=∑(x−x‾)2n−1s = \sqrt{\frac{\sum (x - \overline{x})^2}{n - 1}}
  • Step-by-Step Calculation of Population Variance and Standard Deviation:

    • 1. Calculate the population mean: μ=∑xN\mu = \frac{\sum x}{N}.
    • 2. Find the deviation for each entry: x−μx - \mu.
    • 3. Square each deviation: (x−μ)2(x - \mu)^2.
    • 4. Add the squared deviations to find the sum of squares: SSx=∑(x−μ)2SS_x = \sum (x - \mu)^2.
    • 5. Divide the sum of squares by the population size NN to find the population variance: σ2=SSxN\sigma^2 = \frac{SS_x}{N}.
    • 6. Take the square root of the population variance to find the population standard deviation: σ=σ2\sigma = \sqrt{\sigma^2}.
    • Worked Example (Corporation A Starting Salaries, N=10N = 10, μ=51.5\mu = 51.5):
    • 51−51.5=−0.5  ⟹  (−0.5)2=0.2551 - 51.5 = -0.5 \implies (-0.5)^2 = 0.25
    • 48−51.5=−3.5  ⟹  (−3.5)2=12.2548 - 51.5 = -3.5 \implies (-3.5)^2 = 12.25
    • 49−51.5=−2.5  ⟹  (−2.5)2=6.2549 - 51.5 = -2.5 \implies (-2.5)^2 = 6.25
    • 55−51.5=3.5  ⟹  (3.5)2=12.2555 - 51.5 = 3.5 \implies (3.5)^2 = 12.25
    • 57−51.5=5.5  ⟹  (5.5)2=30.2557 - 51.5 = 5.5 \implies (5.5)^2 = 30.25
    • 51−51.5=−0.5  ⟹  (−0.5)2=0.2551 - 51.5 = -0.5 \implies (-0.5)^2 = 0.25
    • 54−51.5=2.5  ⟹  (2.5)2=6.2554 - 51.5 = 2.5 \implies (2.5)^2 = 6.25
    • 51−51.5=−0.5  ⟹  (−0.5)2=0.2551 - 51.5 = -0.5 \implies (-0.5)^2 = 0.25
    • 47−51.5=−4.5  ⟹  (−4.5)2=20.2547 - 51.5 = -4.5 \implies (-4.5)^2 = 20.25
    • 52−51.5=0.5  ⟹  (0.5)2=0.2552 - 51.5 = 0.5 \implies (0.5)^2 = 0.25
    • Sum of Squares (SSxSS_x):       SSx=0.25+12.25+6.25+12.25+30.25+0.25+6.25+0.25+20.25+0.25=88.5SS_x = 0.25 + 12.25 + 6.25 + 12.25 + 30.25 + 0.25 + 6.25 + 0.25 + 20.25 + 0.25 = 88.5
    • Population Variance:       σ2=88.510=8.85≈8.9\sigma^2 = \frac{88.5}{10} = 8.85 \approx 8.9
    • Population Standard Deviation:       σ=8.85≈2.97≈3.0 (or $3000)\sigma = \sqrt{8.85} \approx 2.97 \approx 3.0\text{ (or \$3000)}
  • Step-by-Step Calculation of Sample Variance and Standard Deviation:

    • 1. Calculate the sample mean: x‾=∑xn\overline{x} = \frac{\sum x}{n}.
    • 2. Find the deviation for each entry: x−x‾x - \overline{x}.
    • 3. Square each deviation: (x−x‾)2(x - \overline{x})^2.
    • 4. Sum the squared deviations to find the sum of squares: SSx=∑(x−x‾)2SS_x = \sum (x - \overline{x})^2.
    • 5. Divide by n−1n - 1 to find the sample variance: s2=SSxn−1s^2 = \frac{SS_x}{n - 1}.
    • 6. Take the square root to find the sample standard deviation: s=s2s = \sqrt{s^2}.
    • Worked Example (High School Football Concussion Recovery Times in Days):
    • Data (n=12n = 12): 8,10,4,6,7,7,9,10,7,6,5,118, 10, 4, 6, 7, 7, 9, 10, 7, 6, 5, 11.
    • Mean calculation:       ∑x=90  ⟹  x‾=9012=7.5 days\sum x = 90 \implies \overline{x} = \frac{90}{12} = 7.5\text{ days}
    • Squared deviations from the mean (7.57.5):
      • 8:(8−7.5)2=(0.5)2=0.258: (8 - 7.5)^2 = (0.5)^2 = 0.25
      • 10:(10−7.5)2=(2.5)2=6.2510: (10 - 7.5)^2 = (2.5)^2 = 6.25
      • 4:(4−7.5)2=(−3.5)2=12.254: (4 - 7.5)^2 = (-3.5)^2 = 12.25
      • 6:(6−7.5)2=(−1.5)2=2.256: (6 - 7.5)^2 = (-1.5)^2 = 2.25
      • 7:(7−7.5)2=(−0.5)2=0.257: (7 - 7.5)^2 = (-0.5)^2 = 0.25
      • 7:(7−7.5)2=(−0.5)2=0.257: (7 - 7.5)^2 = (-0.5)^2 = 0.25
      • 9:(9−7.5)2=(1.5)2=2.259: (9 - 7.5)^2 = (1.5)^2 = 2.25
      • 10:(10−7.5)2=(2.5)2=6.2510: (10 - 7.5)^2 = (2.5)^2 = 6.25
      • 7:(7−7.5)2=(−0.5)2=0.257: (7 - 7.5)^2 = (-0.5)^2 = 0.25
      • 6:(6−7.5)2=(−1.5)2=2.256: (6 - 7.5)^2 = (-1.5)^2 = 2.25
      • 5:(5−7.5)2=(−2.5)2=6.255: (5 - 7.5)^2 = (-2.5)^2 = 6.25
      • 11:(11−7.5)2=(3.5)2=12.2511: (11 - 7.5)^2 = (3.5)^2 = 12.25
    • Sum of Squares (SSxSS_x):       SSx=51.0SS_x = 51.0
    • Sample Variance:       s2=51.012−1=51.011≈4.636≈4.6s^2 = \frac{51.0}{12 - 1} = \frac{51.0}{11} \approx 4.636 \approx 4.6
    • Sample Standard Deviation:       s=51.011≈2.15≈2.2 dayss = \sqrt{\frac{51.0}{11}} \approx 2.15 \approx 2.2\text{ days}
  • Estimating and Interpreting Standard Deviation:

    • Standard deviation measures the typical amount by which entries deviate from the mean.
    • Visual Comparison of Three Populations (N=8N = 8, μ=4\mu = 4):
    • Distribution 1: Eight entries equal to 44 (4,4,4,4,4,4,4,44, 4, 4, 4, 4, 4, 4, 4).
      • Each deviation is 00, so σ=0\sigma = 0.
    • Distribution 2: Four entries at 33, four entries at 55 (3,3,3,3,5,5,5,53, 3, 3, 3, 5, 5, 5, 5).
      • Each deviation is ±1\pm 1, so σ=1.0\sigma = 1.0.
    • Distribution 3: Two entries at 11, two at 33, two at 55, two at 77 (1,1,3,3,5,5,7,71, 1, 3, 3, 5, 5, 7, 7).
      • Deviations are ±3\pm 3 and ±1\pm 1, yielding σ=2(9)+2(1)+2(1)+2(9)8=5≈2.2\sigma = \sqrt{\frac{2(9) + 2(1) + 2(1) + 2(9)}{8}} = \sqrt{5} \approx 2.2.
  • The Empirical Rule (68−95−99.768 - 95 - 99.7 Rule):

    • Applies exclusively to symmetric, bell-shaped distributions.
    • Distribution Characteristics:
    • Approximately 68%68\% of data lie within one standard deviation of the mean: [μ−σ,μ+σ][\mu - \sigma, \mu + \sigma].
    • Approximately 95%95\% of data lie within two standard deviations of the mean: [μ−2σ,μ+2σ][\mu - 2\sigma, \mu + 2\sigma].
    • Approximately 99.7%99.7\% of data lie within three standard deviations of the mean: [μ−3σ,μ+3σ][\mu - 3\sigma, \mu + 3\sigma].
    • Detailed Sub-Intervals:
    • 34.13%34.13\% between μ−σ\mu - \sigma and μ\mu, and 34.13%34.13\% between μ\mu and μ+σ\mu + \sigma.
    • 13.59%13.59\% between μ−2σ\mu - 2\sigma and μ−σ\mu - \sigma, and 13.59%13.59\% between μ+σ\mu + \sigma and μ+2σ\mu + 2\sigma.
    • 2.14%2.14\% between μ−3σ\mu - 3\sigma and μ−2σ\mu - 2\sigma, and 2.14%2.14\% between μ+2σ\mu + 2\sigma and μ+3σ\mu + 3\sigma.
    • 0.135%0.135\% in each outer tail beyond 3σ3\sigma.
    • Worked Example (Female Heights in the U.S., Ages 20–29):
    • Survey parameters: Mean height x‾=64.1 inches\overline{x} = 64.1\text{ inches}, Standard deviation s=2.6 inchess = 2.6\text{ inches}.

Normal distribution of female heights illustrating the Empirical Rule

* Find the percentage of women with heights between 58.958.9 and 64.1 inches64.1\text{ inches}:
  * Calculate the distance in standard deviations:

        x‾−2s=64.1−2(2.6)=58.9 inches\overline{x} - 2s = 64.1 - 2(2.6) = 58.9\text{ inches} * Height 58.9 inches58.9\text{ inches} lies exactly two standard deviations below the mean. * Percentage sum:         13.59%+34.13%=47.72%13.59\% + 34.13\% = 47.72\% * Approximately 47.72%47.72\% of females in this age group are between 58.958.9 and 64.1 inches64.1\text{ inches} tall.

  • Chebyshev's Theorem:
    • Applies to any distribution regardless of its shape.
    • Mathematical Rule: The proportion of any data set lying within kk standard deviations of the mean (k>1k > 1) is at least:     1−1k21 - \frac{1}{k^2}
    • Benchmark Values:
    • For k=2k = 2: At least 1−122=1−14=34=75%1 - \frac{1}{2^2} = 1 - \frac{1}{4} = \frac{3}{4} = 75\% of all data lie within two standard deviations of the mean: [μ−2σ,μ+2σ][\mu - 2\sigma, \mu + 2\sigma].
    • For k=3k = 3: At least 1−132=1−19=89≈88.9%1 - \frac{1}{3^2} = 1 - \frac{1}{9} = \frac{8}{9} \approx 88.9\% of all data lie within three standard deviations of the mean: [μ−3σ,μ+3σ][\mu - 3\sigma, \mu + 3\sigma].
    • Worked Example (Age Distributions in Georgia and Iowa):
    • Parameters for Georgia: Mean μ≈38.2 years\mu \approx 38.2\text{ years}, Standard deviation σ≈22.6 years\sigma \approx 22.6\text{ years}.
    • Parameters for Iowa: Mean μ≈39.6 years\mu \approx 39.6\text{ years}, Standard deviation σ≈23.6 years\sigma \approx 23.6\text{ years}.

Population age distribution histogram for Georgia

Population age distribution histogram for Iowa

* Apply Chebyshev's Theorem to Georgia for k=2k = 2:
  * Interval limits:

        μ−2σ=38.2−2(22.6)=38.2−45.2=−7.0 (use 0 since age cannot be negative)\mu - 2\sigma = 38.2 - 2(22.6) = 38.2 - 45.2 = -7.0\text{ (use }0\text{ since age cannot be negative)}\mu + 2\sigma = 38.2 + 2(22.6) = 38.2 + 45.2 = 83.4\text{ years}\n * Conclusion: At least 75\%ofthepopulationofGeorgiaisbetweenbirth(of the population of Georgia is between birth (0)and) and83.4\text{ years} old.\n * Evaluation of an Age of 90:Because: Because90 > 83.4,anageof, an age of90 lies more than two standard deviations above the mean, making it an unusual age for a Georgia resident.\n\n* **Standard Deviation for Grouped Data:**\n * Formula for estimating the sample standard deviation of a frequency distribution:\n    s = \sqrt{\frac{\sum (x - \overline{x})^2 f}{n - 1}}\n    where xistheclassmidpoint,is the class midpoint,fistheclassfrequency,andis the class frequency, andn = \sum f.\n * Worked Example 1 (Number of Children in 50 Households):\n * Frequency table values:\n * x = 0,,f = 10::x \cdot f = 0\n * x = 1,,f = 19::x \cdot f = 19\n * x = 2,,f = 7::x \cdot f = 14\n * x = 3,,f = 7::x \cdot f = 21\n * x = 4,,f = 2::x \cdot f = 8\n * x = 5,,f = 1::x \cdot f = 5\n * x = 6,,f = 4::x \cdot f = 24\n * Sample size: n = \sum f = 50.\n * Mean calculation:\n      \sum xf = 91 \implies \overline{x} = \frac{91}{50} = 1.82 \approx 1.8\text{ children}\n * Squared deviations multiplied by frequencies:\n * x = 0::(0 - 1.82)^2 \times 10 = (-1.82)^2 \times 10 = 3.3124 \times 10 = 33.1240\n * x = 1::(1 - 1.82)^2 \times 19 = (-0.82)^2 \times 19 = 0.6724 \times 19 = 12.7756\n * x = 2::(2 - 1.82)^2 \times 7 = (0.18)^2 \times 7 = 0.0324 \times 7 = 0.2268\n * x = 3::(3 - 1.82)^2 \times 7 = (1.18)^2 \times 7 = 1.3924 \times 7 = 9.7468\n * x = 4::(4 - 1.82)^2 \times 2 = (2.18)^2 \times 2 = 4.7524 \times 2 = 9.5048\n * x = 5::(5 - 1.82)^2 \times 1 = (3.18)^2 \times 1 = 10.1124 \times 1 = 10.1124\n * x = 6::(6 - 1.82)^2 \times 4 = (4.18)^2 \times 4 = 17.4724 \times 4 = 69.8896\n * Sum of squared terms:\n      \sum (x - \overline{x})^2 f = 145.38\n * Sample Standard Deviation:\n      s = \sqrt{\frac{145.38}{50 - 1}} = \sqrt{\frac{145.38}{49}} \approx 1.72 \approx 1.7\text{ children}\n * Worked Example 2 (Home Sales Prices in Thousands of Dollars):\n * Classes and midpoints (for open-ended class 500+,midpoint, midpoint599.5 is used):\n * 0\text{--}99:midpoint: midpointx = 49.5,,f = 59,,xf = 2920.5\n * 100\text{--}199:midpoint: midpointx = 149.5,,f = 136,,xf = 20332.0\n * 200\text{--}299:midpoint: midpointx = 249.5,,f = 79,,xf = 19710.5\n * 300\text{--}399:midpoint: midpointx = 349.5,,f = 41,,xf = 14329.5\n * 400\text{--}499:midpoint: midpointx = 449.5,,f = 18,,xf = 8091.0\n * 500+:midpoint: midpointx = 599.5,,f = 42,,xf = 25179.0\n * Totals: n = \sum f = 375,,\sum xf = 90562.50.\n * Estimated Sample Mean:\n      \overline{x} = \frac{90562.50}{375} = 241.5\text{ (or \$241,500)}\n * Squared deviation weighted sums:\n * (49.5 - 241.5)^2 \times 59 = (-192)^2 \times 59 = 36864 \times 59 = 2174976\n * (149.5 - 241.5)^2 \times 136 = (-92)^2 \times 136 = 8464 \times 136 = 1151104\n * (249.5 - 241.5)^2 \times 79 = (8)^2 \times 79 = 64 \times 79 = 5056\n * (349.5 - 241.5)^2 \times 41 = (108)^2 \times 41 = 11664 \times 41 = 478224\n * (449.5 - 241.5)^2 \times 18 = (208)^2 \times 18 = 43264 \times 18 = 778752\n * (599.5 - 241.5)^2 \times 42 = (358)^2 \times 42 = 128164 \times 42 = 5382888\n * Sum of squared terms: \sum (x - \overline{x})^2 f = 9971000.\n * Estimated Sample Standard Deviation:\n      s = \sqrt{\frac{9971000}{375 - 1}} = \sqrt{\frac{9971000}{374}} \approx 163.3\text{ (or \$163,300)}\n\n* **Coefficient of Variation (CV):**\n * Describes the standard deviation of a data set as a percentage of its mean.\n * Allows for direct comparison of variation across data sets measured in completely different units or with substantially different magnitudes.\n * Population Formula:\n    CV = \frac{\sigma}{\mu} \times 100\%\n * Sample Formula:\n    CV = \frac{s}{\overline{x}} \times 100\%\n * Worked Example (Basketball Team Heights versus Weights):\n * Team Heights (N = 12,ininches):, in inches):72, 74, 68, 76, 74, 69, 72, 79, 70, 69, 77, 73.\n * Mean \mu \approx 72.8\text{ inches},Standarddeviation, Standard deviation\sigma \approx 3.3\text{ inches}.\n * Calculation:\n        CV_{\text{height}} = \frac{3.3}{72.8} \times 100\% \approx 4.5\%\n * Team Weights (N = 12,inpounds):, in pounds):180, 168, 225, 201, 189, 192, 197, 162, 174, 171, 185, 210.\n * Mean \mu \approx 187.8\text{ pounds},Standarddeviation, Standard deviation\sigma \approx 17.7\text{ pounds}.\n * Calculation:\n        CV_{\text{weight}} = \frac{17.7}{187.8} \times 100\% \approx 9.4\%\n * Comparison: Because 9.4\% > 4.5\%$$, the team's weights demonstrate greater relative variability than their heights.