Sampling, Data, and Levels of Measurement

Fundamentals of Statistics, Probability, and Key Terms

  • Definitions and Core Concepts

    • Statistics: The science of planning studies and experiments, obtaining data, and then organizing, summarizing, presenting, analyzing, interpreting, and drawing conclusions based on the data.

    • Descriptive Statistics: Methods for organizing and summarizing data through graphical representations and numerical values (such as an average).

    • Inferential Statistics: Methods that take a result from a sample, extend it to the population, and measure the reliability of the result.

    • Population: The entire group of individuals, things, or objects to be studied.

    • Sample: A subset of the population that is being studied.

    • Individual: A person or object that is a member of the population being studied.

    • Sampling: Selecting a portion (the sample) of the larger population to study for the purpose of gaining information about the population.

    • Parameter: A numerical summary of a population. For example, if 48.2%48.2\% of all students on a campus own a car, 48.2%48.2\% is a parameter because it describes the entire campus population.

    • Statistic: A numerical summary based on a sample. For example, if 46%46\% of a sample of 100100 students own a car, 46%46\% is a statistic because it describes that specific sample. A sample statistic is used to estimate a population parameter.

    • Representative Sample: A sample that contains the essential characteristics of the population. One of the main concerns in statistics is evaluating how accurately a statistic estimates a parameter.

    • Variable: A characteristic or measurement that can be determined for each member of the population (usually denoted by capital letters such as XX and YY).

    • Data: The actual values of the variables, which may be numbers or words.

    • Mean: The arithmetic average representing the center of the data. The terms "mean" and "average" are used interchangeably.

    • Proportion: The part out of the whole total, calculated using the formula:

proportion=parttotal\text{proportion} = \frac{\text{part}}{\text{total}}

  • Structural Relationships and Memory Hooks

    • Population pairs with Parameter (both begin with the letter P).

    • Sample pairs with Statistic (both begin with the letter S).

    • Measuring every member of a population is usually too slow, too expensive, or physically impossible. Therefore, a sample is measured to compute a statistic, which is then used to estimate the unseen parameter.

  • Illustrative Examples

    • Example 1: A community college employs 8686 full-time faculty members. To gain faculty opinions about an upcoming building project, the college president obtains a simple random sample of 99 faculty members.

      • Population: All 8686 full-time faculty members at the community college.

      • Sample: The 99 faculty members selected for the study.

      • Application: Data collected from the 99 faculty members is used to make inferences about the entire population of 8686 faculty members.

    • Example 2: Wind turbine maintenance companies hire employees to drive around wind farms and inspect reporting turbines. There are 114114 such employees nationwide. A random sample of 5050 employees shows a mean salary of $31,986\$31{,}986.

      • Sample: The 5050 employees selected.

      • Population: The total 114114 employees nationwide.

      • Statistic: The mean salary of the sample of 5050 employees ($31,986\$31{,}986).

      • Parameter: The mean salary of all 114114 employees nationwide.

    • Example 3 (Classifying Parameters vs. Statistics):

      • Following the 2006 national midterm election, 18%18\% of the governors of the 5050 United States were female: Parameter (describes the entire population of 5050 state governors).

      • The average score for a class of 5555 students taking a statistics midterm exam was 68%68\%: Parameter (describes the entire population of that specific class of 5555 students).

      • In a national survey of high school students (grades 99 to 1212), 25%25\% of respondents reported that someone had offered, sold, or given them an illegal drug at school: Statistic (describes a sample of survey respondents).

      • In a national survey on substance abuse, 66.4%66.4\% of respondents who were full-time college students aged 1818 to 2222 reported using alcohol within the past month: Statistic (describes a sample of survey respondents).

  • Section 1.1 Practice Problems & Solutions

    • Problem 1: A university has 4,0004{,}000 students. A researcher surveys 200200 of them and finds that 5050 own a car.

      • (a) Population: All 4,0004{,}000 students at the university.

      • (b) Sample: The 200200 students who were surveyed.

      • (c) Proportion owning a car: Fraction = 50200=14\frac{50}{200} = \frac{1}{4}, Decimal = 0.250.25, Percent = 25%25\%.

      • (d) Parameter or Statistic: Statistic, because the value is calculated from a sample of 200200 students rather than the entire student body.

    • Problem 2: A statistics class has exactly 2525 students. The instructor records that 1515 of them are first-year students.

      • (a) Proportion of first-year students in the whole class: 1525=0.60\frac{15}{25} = 0.60 or 60%60\%.

      • (b) Parameter or Statistic: Parameter, because the 2525 students constitute the entire population of the class.

      • (c) Random selection of 55 students yields 22 first-year students: Proportion = 25=0.40\frac{2}{5} = 0.40 or 40%40\%. This value is a Statistic because it describes a sample of 55 students.

    • Problem 3: A factory makes 1,0001{,}000 light bulbs each day. A quality inspector tests 5050 of them and finds 22 defective.

      • (a) Population: All 1,0001{,}000 light bulbs produced by the factory in a day. Sample: The 5050 light bulbs tested.

      • (b) Defective proportion: Decimal = 250=0.04\frac{2}{50} = 0.04, Percent = 4%4\%.

      • (c) Expected defective bulbs out of 1,0001{,}000: 1,000×0.04=401{,}000 \times 0.04 = 40 defective bulbs.

    • Problem 4: Determine whether each underlined value is a parameter or a statistic:

      • (a) 60%60\% of the 2525 students in this classroom are wearing sneakers: Parameter (refers to all members of the population in the classroom).

      • (b) In a poll of 500500 registered voters, 40%40\% said they plan to vote early: Statistic (refers to a sample of 500500 voters).

      • (c) The mean weight of all 1212 puppies in a litter is 4 lb4\,\text{lb}: Parameter (refers to the entire population of the litter).

      • (d) A sample of 2020 boxes of cereal had a mean weight of 18 oz18\,\text{oz}: Statistic (refers to a sample of 2020 boxes).

    • Problem 5: Reason for taking a sample instead of testing the whole population:

      • (a) Testing whether every firework in a shipment explodes correctly: Testing requires exploding the firework, which destroys it. Testing the entire population would destroy all inventory.

      • (b) Finding the average height of every adult in Texas: The population size is far too large, making it logistically impossible, excessively expensive, and time-consuming to measure every person.

Data Types, Sampling Methods, and Sources of Bias

  • Classifications of Data

    • Qualitative (Categorical) Data: Results from categorizing or describing attributes or characteristics of a population (e.g., gender, zip code, eye color, hometown). Relies on sensory observation (sight, smell, touch, hearing). Example: "The flower is yellow."

    • Quantitative Data: Results from counting or measuring attributes of a population (e.g., height, weight, GPA, time). Relies on measurement tools (rulers, beakers, thermometers). Example: "The leaf is 7 cm7\,\text{cm} long."

    • Example 4 (Qualitative vs. Quantitative):

      • (a) Social Security Number: Qualitative (functions as an identifying label rather than a numerical count/measurement).

      • (b) Temperature: Quantitative (a measured value).

      • (c) Number of days during the past week a student posted on social media: Quantitative (a counted quantity).

      • (d) Zip code: Qualitative (a categorical geographic code).

  • Quantitative Sub-classifications

    • Discrete Random Variable: Has either a finite or countable number of values. Plotted on a number line with space between each point.

    • Continuous Random Variable: Has infinitely many values over a continuous range without gaps, interruptions, or jumps. Includes whole numbers, fractions, decimals, and irrational numbers.

    • Example 5 (Discrete vs. Continuous):

      • (a) Number of text messages received in a week: Discrete

      • (b) Volume of water lost each day through a leaky faucet: Continuous

      • (c) Length (in seconds) of a country song: Continuous

      • (d) Number of sequoia trees in a randomly selected acre of Yosemite National Park: Discrete

      • (e) Temperature on a randomly selected day in Denton, Texas: Continuous

      • (f) Internet connection speed in kilobytes per second: Continuous

      • (g) Points scored in an NCAA basketball game: Discrete

      • (h) Air pressure in pounds per square inch (psi\text{psi}) in an automobile tire: Continuous

  • Sampling Techniques

    • Simple Random Sampling: Randomly selects nn subjects from a population such that every subject has an equal chance of selection (e.g., drawing names from a hat).

    • Convenience Sampling: Uses data that is easy to obtain. It is not a valid scientific sampling method and routinely produces biased data.

    • Systematic Sampling: Selects a random starting point and then chooses every kthk\text{th} element in the population sequence (e.g., choosing every 3rd3\text{rd} person after starting at the 2nd2\text{nd} person).

    • Stratified Sampling: Subdivides the population into at least two distinct subgroups (strata) sharing characteristics, and then draws a sample from each subgroup ("Some from all").

    • Cluster Sampling: Divides the population into sections or clusters, and then randomly selects all members from some of those clusters ("All from some").

    • Key Distinction: Stratified sampling takes some items from all groups; Cluster sampling takes all items from some groups.

  • Example 6 (Sampling Method Scenarios)

    • Scenario A: Instructors evaluating student study techniques in research methods:

      • (a) Drawing names from a hat: Simple Random Sampling

      • (b) Selecting 1010 students randomly from each class section: Stratified Sampling

      • (c) Selecting all students from 22 specific class sections: Cluster Sampling

    • Scenario B: Intel quality-control manager selects every 8th8\text{th} chip coming off the line starting with the 3rd3\text{rd} until reaching 140140 chips: Systematic Sampling

    • Scenario C: State Athletic Commission randomly selects 5050 high schools and tests all varsity baseball players at those selected schools for HGH: Cluster Sampling

    • Scenario D: A representative divides her constituency into low-income, middle-income, and upper-income households, then takes a simple random sample from each class: Stratified Sampling

    • Scenario E: Marketing firm conducts a poll by randomly selecting individuals from a list of known product users: Simple Random Sampling

  • Sampling Bias and Sources of Bias

    • Bias: Occurs when the design or execution of a study systematically favors certain outcomes over others. A sample is biased if it fails to represent the target population.

    • Sampling Bias: The technique used to select individuals favors one part of the population over another (e.g., under-representing certain demographics).

    • Nonresponse Bias: Exists when individuals selected for the sample who fail to respond hold systematically different opinions from those who respond (common with low response rates).

    • Response Bias: Exists when survey answers do not reflect the true feelings or behaviors of respondents (e.g., intentional lying, poor wording, self-reporting sensitive actions).

    • Example 7 (Identifying Sources of Bias):

      • (a) New Canadians are under-represented in a voting intention poll: Sampling Bias

      • (b) An email poll on teen nicotine use gets a 10%10\% response rate, with 8%8\% reporting use: Nonresponse Bias

      • (c) A survey asks subjects to self-report daily alcohol consumption face-to-face: Response Bias

  • Section 1.2 Practice Problems & Solutions

    • Problem 6: Classify variables as qualitative or quantitative:

      • (a) Color of a car in a parking lot: Qualitative

      • (b) Number of cars in a parking lot: Quantitative

      • (c) Student's letter grade in a course: Qualitative

      • (d) Time (in minutes) to drive to campus: Quantitative

      • (e) Person's phone number: Qualitative

      • (f) Temperature in Denton at noon: Quantitative

    • Problem 7: Classify quantitative variables as discrete or continuous:

      • (a) Number of siblings a student has: Discrete

      • (b) Weight of a backpack: Continuous

      • (c) Number of eggs in a carton: Discrete

      • (d) Distance a person walks in a day: Continuous

      • (e) Number of text messages sent in an hour: Discrete

      • (f) Amount of milk poured into a glass: Continuous

    • Problem 8: Identify sampling method used:

      • (a) Teacher writes 3030 names on slips, mixes in a bowl, draws 66: Simple Random Sampling

      • (b) Teacher lines students up alphabetically and selects every 5th5\text{th} student: Systematic Sampling

      • (c) School has 44 grade levels; researcher selects 1010 students from each grade: Stratified Sampling

      • (d) School has 2020 classrooms; researcher picks 33 classrooms randomly and surveys every student in them: Cluster Sampling

      • (e) Reporter interviews first 1212 people walking out of a store: Convenience Sampling

    • Problem 9: Identify source of bias:

      • (a) Internet habit survey conducted purely online: Sampling Bias (excludes people without internet access)

      • (b) Survey mailed to 1,0001{,}000 households with 4040 replies: Nonresponse Bias

      • (c) Students asked face-to-face by instructor how many classes they skipped: Response Bias

    • Problem 10: Campus of 800800 students evenly split into 44 majors (200200 per major); sample size 4040:

      • (a) Stratified sample with equal numbers per major: 404=10\frac{40}{4} = 10 students from each major.

      • (b) Fraction of total campus in sample: 40800=120\frac{40}{800} = \frac{1}{20}, which equals 5%5\%.

      • (c) Surveying all 200200 students in a single major is not stratified because it takes zero students from the other three majors. This method is Cluster Sampling (or Convenience Sampling relative to the whole campus).

Frequency Distributions and Levels of Measurement

  • Frequency and Frequency Tables

    • Frequency: The total count of observations falling within a specific category or class.

    • Frequency Distribution: A listing of all categories along with their corresponding frequency counts.

    • Relative Frequency: The proportion or percentage of the total observations belonging to a category:

relative frequency=frequency of the classtotal number of observations\text{relative frequency} = \frac{\text{frequency of the class}}{\text{total number of observations}}

*   **Relative Frequency Distribution**: A listing of all categories along with their relative frequencies. The sum of all relative frequencies must equal 11 (or approximately 11 if rounding occurs).
*   **Cumulative Frequency Distribution**: The sum of the frequencies for a class and all preceding classes. The final entry always equals the total sample size.
  • Example 8 (Constructing Frequency Tables)

    • Data set of pocket cash collected from students:

      • \0\text{--}\44: Frequency = 55, Relative Frequency = 539≈0.1282\frac{5}{39} \approx 0.1282, Cumulative Frequency = 55

      • \5\text{--}\99: Frequency = 33, Relative Frequency = 339≈0.0769\frac{3}{39} \approx 0.0769, Cumulative Frequency = 88

      • \10\text{--}\1414: Frequency = 88, Relative Frequency = 839≈0.2051\frac{8}{39} \approx 0.2051, Cumulative Frequency = 1616

      • \15\text{--}\1919: Frequency = 66, Relative Frequency = 639≈0.1538\frac{6}{39} \approx 0.1538, Cumulative Frequency = 2222

      • \20\text{--}\2424: Frequency = 1010, Relative Frequency = 1039≈0.2564\frac{10}{39} \approx 0.2564, Cumulative Frequency = 3232

      • \25\text{--}\2929: Frequency = 77, Relative Frequency = 739≈0.1795\frac{7}{39} \approx 0.1795, Cumulative Frequency = 3939

      • Total Frequency = 3939, Total Relative Frequency = 1.00001.0000

  • Four Levels of Measurement

    • Nominal: Categories consist of names, labels, or qualities only. Data cannot be arranged in an ordered scheme, and arithmetic calculations are meaningless (e.g., eye color, zip code, gender, jersey numbers).

    • Ordinal: Data can be arranged in a meaningful order or rank, but precise differences between data values cannot be measured or are meaningless (e.g., letter grades A/B/C, survey ratings poor/fair/good, race finishing positions).

    • Interval: Data can be ordered, and differences between values are equal and meaningful. However, there is no true zero point (zero does not mean "none"), so ratios are meaningless (e.g., temperature in ∘F^\circ\text{F} or ∘C^\circ\text{C}, calendar years).

    • Ratio: Data can be ordered, differences are meaningful, and there is a true zero point (zero represents the complete absence of the attribute). Ratios are fully valid (e.g., height, weight, age, income, distance, counts).

    • The True Zero Test: Does zero mean none of the variable exists? Zero pounds means no weight (Ratio scale). Zero degrees Fahrenheit does not mean no temperature (Interval scale); thus, 80∘F80^\circ\text{F} is not twice as hot as 40∘F40^\circ\text{F}.

  • Levels of Measurement Comparison Matrix

    • Nominal: Ordered? No | Equal gaps? No | True zero? No | Ratios valid? No

    • Ordinal: Ordered? Yes | Equal gaps? No | True zero? No | Ratios valid? No

    • Interval: Ordered? Yes | Equal gaps? Yes | True zero? No | Ratios valid? No

    • Ratio: Ordered? Yes | Equal gaps? Yes | True zero? Yes | Ratios valid? Yes

  • Section 1.3 Practice Problems & Solutions

    • Problem 11: Survey of 2020 students regarding pet ownership:

      • 00 pets: Frequency = 22, Relative Frequency = 220=0.10\frac{2}{20} = 0.10, Cumulative Frequency = 22

      • 11 pet: Frequency = 66, Relative Frequency = 620=0.30\frac{6}{20} = 0.30, Cumulative Frequency = 88

      • 22 pets: Frequency = 88, Relative Frequency = 820=0.40\frac{8}{20} = 0.40, Cumulative Frequency = 1616

      • 33 pets: Frequency = 44, Relative Frequency = 420=0.20\frac{4}{20} = 0.20, Cumulative Frequency = 2020

      • Total: Frequency = 2020, Relative Frequency = 1.001.00

      • (b) Percent owning exactly 22 pets: 0.40=40%0.40 = 40\%.

      • (c) Percent owning 22 or fewer pets: 1620=0.80=80%\frac{16}{20} = 0.80 = 80\%.

      • (d) Check: 0.10+0.30+0.40+0.20=1.000.10 + 0.30 + 0.40 + 0.20 = 1.00.

    • Problem 12: Quiz scores for 5050 students:

      • Score 0–40\text{--}4: Frequency = 55, Relative Frequency = 550=0.10\frac{5}{50} = 0.10, Cumulative Frequency = 55

      • Score 5–95\text{--}9: Frequency = 1010, Relative Frequency = 1050=0.20\frac{10}{50} = 0.20, Cumulative Frequency = 1515

      • Score 10–1410\text{--}14: Frequency = 2020, Relative Frequency = 2050=0.40\frac{20}{50} = 0.40, Cumulative Frequency = 3535

      • Score 15–1915\text{--}19: Frequency = 1515, Relative Frequency = 1550=0.30\frac{15}{50} = 0.30, Cumulative Frequency = 5050

      • Total: Frequency = 5050, Relative Frequency = 1.001.00

      • (b) Proportion scoring below 1010: 1550=0.30\frac{15}{50} = 0.30

      • (c) Proportion scoring 1010 or above: 1−0.30=0.701 - 0.30 = 0.70

    • Problem 13: Level of measurement classification:

      • (a) Student ID number: Nominal

      • (b) Finishing position in a 5K race (1st1\text{st}, 2nd2\text{nd}, 3rd3\text{rd}): Ordinal

      • (c) Temperature of a classroom in ∘F^\circ\text{F}: Interval

      • (d) Number of pages in a textbook: Ratio

      • (e) T-shirt size (small, medium, large): Ordinal

      • (f) Person's annual income in dollars: Ratio

    • Problem 14: Reasoning exercises:

      • (a) The statement that a 90∘F90^\circ\text{F} day is "twice as hot" as a 45∘F45^\circ\text{F} day is invalid because temperature in Fahrenheit is measured on an interval scale. It lacks a true zero point (0∘F0^\circ\text{F} does not mean zero heat), making ratio comparisons meaningless.

      • (b) The statement that a 90-lb90\text{-lb} dog is twice as heavy as a 45-lb45\text{-lb} dog is valid because weight is measured on a ratio scale, which possesses an absolute, true zero (0 lb0\,\text{lb} represents the complete absence of weight).