Comprehensive Study Notes on Intro to Statistics: Terminology, Data Classification, Study Design, and Sampling

Course Guidelines and Classroom Policies

  • Classroom Attendance and Integrity Policy:

    • Attendance requires full physical and mental presence for the duration of the entire class period.
    • Signing the attendance roster and leaving before the conclusion of class is explicitly classified as cheating and is strictly prohibited.

  • Class Schedule and Time Management:

    • Class concludes at 05:10.
    • Instructional material wraps up at 05:05 to reserve 5 to 10 minutes at the end of class to address technical setups such as lockdown browsers.

  • Preparation Expectations:

    • Review material from prior lectures before attending class to maintain continuity.
    • Questions regarding prior material are actively encouraged at the beginning of each session.

Fundamentals of Statistics

  • Definition and Purpose of Statistics:

    • Statistics is the process of generating numerical data to reveal accurate narratives and support decision-making.
    • Data serves as the objective baseline for conclusions, preventing reliance on guesswork or emotional intuition.
    • Core Principle: Without data, there are no statistics.

  • Terminology Distinction:

    • Statistic (singular): A single calculated numerical value derived from a dataset (e.g., a mean or a mode).
    • Statistics (plural): The field of study or a collective group of calculated values.

  • Two Main Branches of Statistics:

    • Descriptive Statistics:

    • Involves the collection, organization, summarization, and presentation of data.

    • Focuses strictly on describing the specific dataset at hand without making broader generalizations.

    • Examples of descriptive statistics include the mean, median, mode, range, and standard deviation.

    • Covered extensively in Chapters 2 and 3.

    • Inferential Statistics:

    • Involves taking sample data, performing analysis, and drawing conclusions or inferences about the broader population from which the sample was drawn.

    • Covered in Chapters 7, 8, 9, and 10.

    • Key components of inferential statistics include:

      • Hypothesis Testing: Answering comparative questions about groups (e.g., comparing test performance between male and female students).
      • Estimations: Using sample values to estimate unknown population parameters (e.g., constructing confidence intervals rather than relying on a single point value to estimate true average GPA).
      • Determining Relationships: Evaluating associations between variables (e.g., analyzing the relationship between class absences and final grades, or food intake and body weight).
      • Predictions: Utilizing established relationships between variables to forecast future outcomes.

Classification of Data: Quantitative vs. Qualitative

  • Quantitative Data:

    • Composed of numerical values derived from counts or measurements.
    • Represents measurable quantities where mathematical calculations are meaningful.
    • Examples include height, age, and weight.

  • Qualitative Data:

    • Composed of words, labels, names, descriptions, or non-numeric observations.
    • Also referred to as categorical data or attribute data because it places observations into distinct categories.
    • Examples include eye color (blue, green, brown) and political affiliation (Democrat, Republican).

  • Non-Quantitative Numbers (Qualitative Numbers):

    • Certain data represented by numerical digits are classified as qualitative because the numbers serve strictly as non-numeric labels or identifiers rather than measurable quantities.
    • Mathematical operations (such as addition or averaging) on these numbers are meaningless.
    • Examples include:
    • Jersey numbers on athletic uniforms (e.g., jersey number 2424 used for Kobe Bryant acts as a label replacing a name).
    • Student Identification (ID) numbers.
    • Telephone numbers.
    • Social Security numbers.

Types of Quantitative Data: Discrete vs. Continuous

  • Discrete Data:

    • Consists of values that are countable, finite, or take on whole numbers only.
    • Values cannot be meaningfully divided into fractions or decimals in the context of the variable.
    • Examples include:
    • Number of children in a household (e.g., 22 children, never 1.51.5 children).
    • Number of pennies.
    • Number of enrolled students (e.g., 103103 students).
    • Number of shoes, cars, or houses.

  • Continuous Data:

    • Consists of an infinite number of possible values within a given range, including whole numbers, fractions, and decimals.
    • Obtained through the process of measurement rather than direct counting.
    • Examples include:
    • Height measurements (e.g., 5.1 feet5.1\,\text{feet}).
    • Weight measurements (e.g., 90.5 lbs90.5\,\text{lbs} or 90 lbs90\,\text{lbs}).
    • Volume of liquids consumed (e.g., 2 liters2\,\text{liters} or 1.5 liters1.5\,\text{liters} of milk).
    • Caloric intake (e.g., 100 calories100\,\text{calories}, 200 calories200\,\text{calories}, or 50.5 calories50.5\,\text{calories}).

Levels of Measurement

  • 1. Nominal Level:

    • Characterized by data that consist of names, labels, or categories only.
    • Exclusively qualitative data.
    • Data cannot be arranged in an ordering or ranking scheme.
    • Calculations cannot be performed on nominal data.
    • Examples include gender (female, male) and race.

  • 2. Ordinal Level:

    • Data can be arranged in a specific order or rank (e.g., small, medium, large; good, better, best).
    • While data can be ranked, precise mathematical differences between ranks do not exist or are meaningless.
    • Examples include:
    • Course evaluation ratings (e.g., rating a professor as average, good, or extremely good indicates relative standing, but the exact difference in quality between ranks cannot be quantified).
    • Letter grades.
    • Military ranks (e.g., private, corporal, captain).

  • 3. Interval Level:

    • Data can be ordered, ranked, and precise mathematical differences between data values are meaningful.
    • Does not possess a true zero; the value zero is arbitrary and acts as a placeholder rather than indicating the total absence of the quantity.
    • Because there is no true zero, ratios between values are not meaningful (e.g., 80∘F80^\circ\text{F} is not twice as hot as 40∘F40^\circ\text{F}).
    • Mathematical operations of addition and subtraction are valid:     Difference=70∘F−60∘F=10∘F\text{Difference} = 70^\circ\text{F} - 60^\circ\text{F} = 10^\circ\text{F}
    • Examples include temperature scales (Fahrenheit), IQ scores, and calendar years.

  • 4. Ratio Level:

    • Possesses all characteristics of the interval level (labeling, ranking, and meaningful differences).
    • Includes a true zero, where zero represents the absolute absence of the quantity being measured.
    • Ratios between numerical values are fully meaningful.
    • Example calculation:     \20 \times 2 = \4040     A person with $40\$40 has exactly twice as much money as a person with $20\$20.
    • Examples include bank account balances (where $0\$0 means absolute absence of money), height, weight, and counts of items.

Effective Study Habits and Memory Retention

  • Immediate Review Protocol:

    • Spend 10 to 30 minutes immediately following class working on homework problems that align with the material covered that day.
    • PowerPoints should be utilized as a structural guide to direct targeted readings in the textbook.

  • Cognitive Retention Science:

    • Allowing 3 consecutive days to pass without reviewing newly presented material leads to significant memory decay.
    • Returning to unreviewed material after extended gaps causes content to appear unfamiliar, increasing study time and academic stress.

Statistical Study Designs

  • 1. Observational Studies:

    • Measures specific characteristics or observes subjects without modifying them or applying any interventions/treatments.
    • Examples:
    • Counting vehicle makes (e.g., Toyota, Tesla, Mercedes) at a street intersection for 1 hour.
    • Recording student weights using a scale without applying dietary changes.
    • Categories of Observational Studies:
    • Cross-Sectional Study: Data are observed, measured, or collected at one specific point in time (e.g., recording student GPAs at the present moment or tracking institution graduation rates for a single year).
    • Retrospective Study: Data are collected from past time periods using historical records, archives, or retrospective interviews (e.g., examining past drug mortality rates among heart attack patients).
    • Prospective (Longitudinal / Cohort) Study: Data are collected by following a specific group (a cohort sharing common characteristics) forward in time across future intervals.
    • Cohort Example: Tracking Indiana University (IU) freshman entry cohorts (e.g., cohorts entering in Fall 2019, 2020, 2021, 2023, or 2024) across 4-year, 6-year, and 8-year intervals to observe graduation and dropout rates.
    • National Student Clearinghouse Research Insights: Research demonstrates that undergraduate students pursuing 4-year bachelor's degrees take an average of 6 years to complete their degree, while students pursuing 2-year associate degrees average 4 years to complete, primarily due to increased employment obligations during enrollment.

  • 2. Experimental Studies:

    • Applies a specific treatment to subjects (referred to as experimental units when non-human, or subjects when human) and observes the resulting effects.
    • Lurking Variables:
    • A variable that influences the relationships among study variables but is omitted from the study design.
    • Example: A strong correlation observed between rising ice cream consumption and increased drowning incidents is confounded by the lurking variable of summer temperature.
    • Prevention: Conducting thorough reviews of literature published within the preceding 5 years identifies necessary control variables to include in the experimental design.
    • Elements of Experimental Design:
    • Replication: Repeating an experiment or treatment across a sufficiently large sample size of subjects to ensure reliable outcomes.
    • Blinding: A technique where subjects do not know whether they are receiving a true treatment or a placebo.
      • Placebo: An inactive or fake treatment (e.g., a sugar pill) administered as a control baseline.
      • Single-Blind: The subject does not know whether they received the active treatment or placebo, but the administering physician knows.
      • Double-Blind: Neither the subject nor the administering physician knows who received the active treatment or placebo. Only the developing entity (e.g., pharmaceutical company Eli Lilly) retains the key.
    • Randomization: Assigning subjects to distinct treatment groups completely by chance to remove bias and construct structurally similar comparison groups.

Sampling Techniques and Methodology

  • Unbiased vs. Biased Samples:

    • An unbiased sample is representative of the underlying population and allows for valid statistical generalization.
    • A biased sample fails to represent the population, rendering conclusions mathematically invalid for general claims.

  • Sampling Methods:

    • Random Sampling:

    • Selection of individual subjects where every member in the population has a known, non-zero chance of inclusion.

    • Accomplished by assigning every individual a distinct identification number (e.g., numbering students from 1 to 103) and generating random selections using computer software or random number tables.

    • Simple Random Sampling (SRS):

    • A specific condition where every possible sample of a given size nn drawn from a population size NN has the exact same probability of being selected as any other sample of size nn.

    • Systematic Sampling:

    • Elements are selected by numbering each value in the population and selecting every kthk^{\text{th}} individual.

    • Formula for Sampling Interval (kk):k=Nnk = \frac{N}{n}       Where NN represents population size (capital NN) and nn represents sample size (lowercase nn).

    • Systematic Calculation Example:       Given a population N=2000N = 2000 and a desired sample size n=50n = 50:       k=200050=40k = \frac{2000}{50} = 40       The starting individual is randomly chosen between position 1 and position 40. If the starting position chosen is 22, the subsequent selected positions in the population are:       Position 1=2\text{Position 1} = 2Position 2=2+40=42\text{Position 2} = 2 + 40 = 42Position 3=42+40=82\text{Position 3} = 42 + 40 = 82Position 4=82+40=122\text{Position 4} = 82 + 40 = 122Position 5=122+40=162\text{Position 5} = 122 + 40 = 162

    • Stratified Sampling:

    • The population is subdivided into at least two distinct, non-overlapping subgroups (strata) based on shared characteristics (e.g., gender, religion, political affiliation).

    • A random sample is drawn from each stratum to guarantee representation across all categories in the final sample.

    • Cluster Sampling:

    • The population is divided into separate sections or geographic groups called clusters (e.g., branch locations of high-credit banks across Bloomington: East, West, Ellettsville/Martinsville).

    • Entire clusters are randomly selected, and all individuals residing within the selected clusters are sampled.

    • Multistage Sampling:

    • Data selection constructed by combining multiple sampling methods across distinct stages of research (e.g., utilizing stratified sampling in stage one and simple random sampling in stage two).

    • Convenience / Suspect Sampling:

    • Data collection based on subjects that are easy to access (e.g., surveying immediate friends or nearby individuals).

    • Generates biased data that cannot be generalized to a broader population.

Misleading Statistics

  • Misleading statistics present inaccurate or distorted numbers, leading to false conclusions.
  • Primary sources of statistical misrepresentation include:
    • Suspect/biased samples.
    • Flawed or leading survey questions that manipulate respondent answers.
    • Misleading graphical representations.
    • Improper collection techniques.

Questions & Discussion

  • Classification of Calendar Time and Years:

    • Question: How are years classified within levels of measurement?
    • Response: Years fall under the interval level of measurement because there is no true zero point for absolute time (year zero does not denote the complete absence of time). Differences between years are meaningful, but ratio operations are invalid.

  • Respondus LockDown Browser Setup and Troubleshooting:

    • Licensing Fee: Requires a $10\$10 access fee paid by the student because Indiana University (IU) does not currently hold an active site license.
    • Access Location: Navigated via Access Pearson →\rightarrow MyLab →\rightarrow Assignments →\rightarrow Select the assessment designated with the LockDown Browser requirement.
    • Pop-Up Blockers: If assignment download links fail to open, browser security settings must be modified to unblock pop-up windows. Search device-specific instructions (Mac, Dell, HP) to enable pop-ups for Pearson.
    • Instructor Office Location: 461 S P M.
    • Student Support: Assistance is available via organized STATS lab drop-ins and scheduled Zoom sessions.