Exhaustive Study Notes on Introductory Statistics: Scales of Measurement, Real Limits, Statistical Notation, and Computational Operations

Administrative Logistics, Course Schedule, and Learning Resources

• Homework Deadlines and Submission Guidelines:

  • The Chapter 1 homework deadline was extended from Friday to today at 11:59 PM.

  • Submissions can be completed via digital scan uploaded to Canvas by 11:59 PM, or submitted directly as a physical paper copy at the end of class to the instructor or the Teaching Assistant, Elise.

  • When completing homework, all intermediate mathematical steps must be fully written out (e.g., showing $1 + 3 + 5 + 8 = 17$ rather than just providing the final value $17$). Demonstrating the complete process is required to verify procedural understanding, ensure standard grading criteria across all students, identify specific points of computational error, and qualify for partial credit on examinations.

• Office Hours and Extra Credit Opportunities:

  • Students can earn $2$ points of extra credit per visit by attending designated TA office hours (Kate's office hours, Kinsley's office hours, or V's office hours).

  • Extra credit visits can be completed every single day office hours are held.

  • Instructor and Head TA (Elise) office hours are available for academic support, but do not award extra credit points.

• Lab Schedule and Software Integration:

  • Current lab sessions focus on reviewing core theoretical concepts introduced during lecture.

  • Starting next week, lab sessions transition to hands-on applications using SPSS (Statistical Package for the Social Sciences).

  • SPSS is an advanced, high-power computer program for data wrangling and statistical analysis.

  • All calculations are performed manually by hand during lecture to ensure deep theoretical understanding of why specific operations are executed, after which SPSS allows rapid automated computation.

  • Proficiency in SPSS can be added as a specialized computational skill on professional resumes.

Review of Fundamental Concepts: Constructs and Variable Types

• Hypothetical Constructs:

  • Constructs are internal attributes, characteristics, or mechanisms that are known to exist but cannot be observed or measured directly.

  • Examples include intelligence, honesty, anxiety, depression, and Post-Traumatic Stress Disorder (PTSD).

  • Because there is no single universally agreed-upon direct measurement method for constructs, researchers must construct explicit operational definitions.

  • Example measurement frameworks for intelligence include standard course tests and quizzes, general IQ tests, the Wechsler Intelligence Scale, and the Stanford-Binet Intelligence Scales.

• Operational Definitions:

  • An operational definition specifies a precise, standardized procedure or set of operations for measuring and defining a hypothetical construct.

• Discrete vs. Continuous Variables:

  • Discrete Variables:

    • Consist of separate, indivisible categories where no intermediate values exist between units.

    • Data are counted strictly in whole units ($1, 2, 3, 4, \dots$).

    • Example: Class size (e.g., strictly $35$ individuals; fractional values such as $34.5$ are impossible).

    • Example: Counting whole milk cartons consumed (e.g., exactly $4$ cartons).

  • Continuous Variables:

    • Variables that can take on an infinite number of possible values along a continuum.

    • Can be divided into an infinite number of fractional parts.

    • Example: Measuring liquid volume poured into a glass (e.g., $1.876\,\text{ounces}$).

    • Example: Measuring weight or height.

Real Limits of Continuous Variables

• Theoretical Definition and Purpose:

  • Because continuous variables can take on infinite precision, measuring devices establish discrete cutoff points based on their designated unit of measurement.

  • Real limits are the boundaries established for scores that are obtained from a continuous variable.

  • The interval defined by real limits is bounded by two values: the Lower Real Limit ($LRL$) and the Upper Real Limit ($URL$).

• Mathematical Formula for Real Limits:

  • For any reported value of a continuous variable measured to a specific unit of precision:         $\text{Half-Unit} = \frac{\text{Unit of Measurement}}{2}$         $\text{Lower Real Limit (LRL)} = \text{Reported Score} - \text{Half-Unit}$         $\text{Upper Real Limit (URL)} = \text{Reported Score} + \text{Half-Unit}$

• Calculated Examples of Real Limits:

  • Example 1 (Nearest 1 Pound Increment):

    • Reported value: $150\,\text{lbs}$.

    • Unit of measurement: $1\,\text{lb}$.

    • Calculated half-unit: $\frac{1}{2} = 0.5\,\text{lbs}$.

    • $\text{LRL} = 150 - 0.5 = 149.5\,\text{lbs}$.

    • $\text{URL} = 150 + 0.5 = 150.5\,\text{lbs}$.

    • Real Limit Interval: $[149.5, 150.5]$.

  • Example 2 (Nearest Half-Pound / $0.5\,\text{lb}$ Increment):

    • Reported value: $150\,\text{lbs}$.

    • Unit of measurement: $0.5\,\text{lbs}$.

    • Calculated half-unit: $\frac{0.5}{2} = 0.25\,\text{lbs}$.

    • $\text{LRL} = 150 - 0.25 = 149.75\,\text{lbs}$.

    • $\text{URL} = 150 + 0.25 = 150.25\,\text{lbs}$.

    • Real Limit Interval: $[149.75, 150.25]$.

  • Example 3 (Nearest 1 Inch Increment):

    • Reported value: $68\,\text{inches}$.

    • Unit of measurement: $1\,\text{inch}$.

    • Calculated half-unit: $\frac{1}{2} = 0.5\,\text{inches}$.

    • $\text{LRL} = 68 - 0.5 = 67.5\,\text{inches}$.

    • $\text{URL} = 68 + 0.5 = 68.5\,\text{inches}$.

    • Real Limit Interval: $[67.5, 68.5]$.

  • Example 4 (Nearest Half-Inch / $0.5\,\text{inch}$ Increment):

    • Reported value: $68\,\text{inches}$.

    • Unit of measurement: $0.5\,\text{inches}$.

    • Calculated half-unit: $\frac{0.5}{2} = 0.25\,\text{inches}$.

    • $\text{LRL} = 68 - 0.25 = 67.75\,\text{inches}$.

    • $\text{URL} = 68 + 0.25 = 68.25\,\text{inches}$.

    • Real Limit Interval: $[67.75, 68.25]$.

• Practical Statistical Significance:

  • Establishing precise real limits directly alters subsequent quantitative computations, such as calculating the exact range of continuous data sets.

Scales of Measurement: The NOIR Framework

• Overview of the NOIR Acronym:

  • The framework categorizing types of measurement scales is abbreviated by the acronym NOIR:

    • N = Nominal

    • O = Ordinal

    • I = Interval

    • R = Ratio

• 1. Nominal Scale:

  • Definition: Consists of a set of categories that have different names, labels, or words. Nominal measurement classifies data into distinct qualitative categories without any quantitative distinction, numerical value, or intrinsic order.

  • Etymology: Derived from Greek and Latin roots (nom / nombre), meaning "name".

  • Properties: No mathematical operations or rankings can be performed across categories.

  • Examples: Academic major (Psychology, Biology, English, Chemistry), race, biological sex/gender, occupation, marital status (single, married, divorced), favorite reality television show.

• 2. Ordinal Scale:

  • Definition: Consists of a set of categories organized in an ordered directional series or ranking. An ordinal scale identifies whether one measurement is greater than or less than another.

  • Properties: Ranks items in sequence, but does not depict the precise quantitative distance or interval between ranks (intervals between ranks are unequal or unknown).

  • Examples: Class rank (Freshman, Sophomore, Junior, Senior), competition finish order ($1\text{st}$, $2\text{nd}$, $3\text{rd}$ place), commercial size labels (Cold Stone Creamery sizes: "Like It", "Love It", "Gotta Have It"), socioeconomic status (Low, Medium, Upper), T-shirt sizes (Small, Medium, Large, Extra Large), letter grades ($A$, $B$, $C$, $D$, $F$).

• 3. Interval Scale:

  • Definition: Consists of ordered categories that are all organized into numerical intervals of equal size. Equal differences between numbers on the scale reflect equal differences in magnitude.

  • Properties: Has equal intervals between values, but lacks an absolute/true zero point. A value of $0$ is arbitrary and does not indicate a complete absence of the variable being measured. Values below zero (negative numbers) are mathematically possible.

  • Examples: Temperature scales ($^\circ\text{F}$ or $^\circ\text{C}$). A temperature of $-20^\circ\text{F}$ is cold, but temperature still exists; $0^\circ\text{F}$ does not represent an absolute absence of heat.

• 4. Ratio Scale:

  • Definition: An interval scale with the additional structural feature of an absolute, true zero point.

  • Properties: A value of $0$ represents a complete absence of the variable being measured. Because zero is absolute, negative values are mathematically impossible. Ratios of numbers accurately reflect ratios of magnitude (e.g., $10\,\text{lbs}$ is twice as heavy as $5\,\text{lbs}$).

  • Examples: Height ($5\,\text{ft } 5.75\,\text{in}$; negative height is impossible), weight ($150\,\text{lbs}$; negative weight is impossible), age, physical count of objects (holding $0, 1, 2, 3, 4$ markers), physical currency/cash income ($0$ dollars in hand represents absolute absence of cash).

• Summary Table of Measured Variable Classifications:

  • Age: Continuous variable measured on a Ratio scale.

  • Physical Cash Income: Ratio scale.

  • Marital Status: Discrete variable measured on a Nominal scale.

  • Academic Grade ($A, B, C, D, F$): Measured on an Ordinal scale.

  • Favorite Television Show: Measured on a Nominal scale.

  • Number of Children: Discrete variable.

    • Note on Misapplied Averages: Studies reporting average household child counts as $2.3$ or $2.5$ reflect a misapplication of the arithmetic mean to discrete data. Discrete data distributions should be summarized using median or mode rather than fractional means.

Mathematical and Statistical Notation

• Precision Rules in Notation:

  • Statistical notation requires strict accuracy; case sensitivity (uppercase vs. lowercase) and formatting (italicized vs. standard) dictate distinct statistical concepts.

• Basic Variable and Sample Size Notation:

  • $X$: Capital letter serving as a placeholder for scores of the primary observed variable (e.g., height, test score).

  • $Y$: Capital letter serving as a placeholder for scores of a secondary observed variable (e.g., weight, GPA).

  • $N$: Uppercase letter representing the total number of scores/individuals in a complete Population.

  • $n$: Lowercase letter representing the total number of scores/individuals in a Sample.

• Summation Operator Notation:

  • $\Sigma$: Capital Greek letter Sigma (characterized by a pointy shape), representing the statistical operation of summation ("add up all scores that follow").

  • $\sigma$: Lowercase Greek letter Sigma, representing population standard deviation (introduced later in the course).

  • $\Sigma X$: Shorthand operation instructing the addition of all individual values listed under variable $X$.

• Order of Mathematical Operations (PEMDAS):

  • Calculations follow the standard order of operations prioritized by the acronym PEMDAS ("Please Excuse My Dear Aunt Sally"):

    1. Parentheses: Perform all computational operations enclosed inside parentheses first.

    2. Exponents: Calculate all squared terms or exponents.

    3. Multiplication and Division: Execute left to right.

    4. Addition and Subtraction / Summation: Perform final addition, subtraction, or summation operations.

Computational Tables and Step-by-Step Examples

• Structure of Computational Tables:

  • A computational table uses columns where each row represents a single research subject or observed score. Columns track intermediate calculation steps required by statistical formulas prior to final summation.

• Worked Example 1 (Single Variable Calculations):

  • Given observed sample data set ($n = 4$):         $X = [9, 7, 6, 3]$

  • Computation 1: Sum of scores ($\Sigma X$)         $\Sigma X = 9 + 7 + 6 + 3 = 25$

  • Computation 2: Sum of squared scores ($\Sigma X^2$)

    • First, square each individual score ($X^2$):             $9^2 = 81, \quad 7^2 = 49, \quad 6^2 = 36, \quad 3^2 = 9$

    • Second, sum the squared values:             $\Sigma X^2 = 81 + 49 + 36 + 9 = 175$

  • Computation 3: Sum of subtracted constants ($\Sigma (X - 2)$)

    • First, subtract $2$ from each individual $X$ score ($X - 2$):             $9 - 2 = 7, \quad 7 - 2 = 5, \quad 6 - 2 = 4, \quad 3 - 2 = 1$

    • Second, sum the resulting differences:             $\Sigma (X - 2) = 7 + 5 + 4 + 1 = 17$

  • Computation 4: Sum of squared subtracted constants ($\Sigma (X - 2)^2$)

    • First, square each value from the $(X - 2)$ column:             $7^2 = 49, \quad 5^2 = 25, \quad 4^2 = 16, \quad 1^2 = 1$

    • Second, sum the squared differences:             $\Sigma (X - 2)^2 = 49 + 25 + 16 + 1 = 91$

• Worked Example 2 (Distinguishing $\Sigma X^2$ vs. $(\Sigma X)^2$):

  • Given observed sample data set ($n = 4$):         $X = [3, 1, 7, 4]$

  • Step 1: Compute $\Sigma X$         $\Sigma X = 3 + 1 + 7 + 4 = 15$

  • Step 2: Compute $\Sigma X^2$

    • Individual squares: $3^2 = 9, \quad 1^2 = 1, \quad 7^2 = 49, \quad 4^2 = 16$

    • $\Sigma X^2 = 9 + 1 + 49 + 16 = 75$

  • Step 3: Compute $(\Sigma X)^2$

    • Per PEMDAS, execute operations within parentheses first ($\Sigma X = 15$).

    • Square the resulting total:             $(\Sigma X)^2 = (15)^2 = 225$

• Worked Example 3 (Operations with Constants):

  • Given observed sample data set ($n = 4$):         $X = [3, 1, 7, 4]$

  • Computation 1: $\Sigma (X - 1)$

    • Subtractions: $(3 - 1) = 2, \quad (1 - 1) = 0, \quad (7 - 1) = 6, \quad (4 - 1) = 3$

    • Summation: $\Sigma (X - 1) = 2 + 0 + 6 + 3 = 11$

  • Computation 2: $\Sigma (X - 1)^2$

    • Squares of differences: $2^2 = 4, \quad 0^2 = 0, \quad 6^2 = 36, \quad 3^2 = 9$

    • Summation: $\Sigma (X - 1)^2 = 4 + 0 + 36 + 9 = 49$

  • Computation 3: $\Sigma X - 1$

    • Perform summation of $X$ first, then subtract $1$:             $\Sigma X - 1 = 15 - 1 = 14$

• Worked Example 4 (Two-Variable Computations: $\Sigma XY$):

  • Given paired data set ($n = 4$):         $X = [3, 1, 7, 4]$         $Y = [5, 3, 4, 2]$

  • Computation: Sum of products ($\Sigma XY$)

    • First, multiply adjacent values for each individual case ($XY$):             $3 \times 5 = 15, \quad 1 \times 3 = 3, \quad 7 \times 4 = 28, \quad 4 \times 2 = 8$

    • Second, sum all resulting products:             $\Sigma XY = 15 + 3 + 28 + 8 = 54$

• Worked Example 5 (Comprehensive Practice Problem):

  • Given observed sample data set ($n = 4$):         $X = [6, 2, 4, 2]$

  • Calculated Output Table:

    • $\Sigma X = 6 + 2 + 4 + 2 = 14$

    • $X^2 = [36, 4, 16, 4] \implies \Sigma X^2 = 36 + 4 + 16 + 4 = 60$

    • $(\Sigma X)^2 = (14)^2 = 196$

    • $(X - 2) = [4, 0, 2, 0] \implies \Sigma (X - 2) = 4 + 0 + 2 + 0 = 6$

    • $(X - 2)^2 = [16, 0, 4, 0] \implies \Sigma (X - 2)^2 = 16 + 0 + 4 + 0 = 20$

Research Design, Construct Operationalization, and Experimental Methodology

• Operationalizing Intangible Constructs (Example: Love):

  • Love is a hypothetical construct with no direct objective instrument for direct physical measurement.

  • To evaluate love scientifically, researchers must establish explicit operational definitions based on observable variables:

    • Behavioral metrics: Selfless acts of service, observable effort, physical contact (e.g., frequency of hugs or kisses).

    • Psychometric metrics: Standardized multi-item self-report intimacy scales (e.g., a 10-question questionnaire scoring emotional closeness).

    • Physiological metrics: Measures of autonomic nervous system activation (e.g., elevated heart rate, galvanic skin response/sweating, gastrointestinal responses commonly termed "butterflies" or being "twitterpated").

• Correlational Research Methodology:

  • Definition: Involves measuring two or more naturally occurring variables per individual to determine if a systematic relationship exists.

  • Limitations: Correlation does not demonstrate causality. A correlation between two variables ($X$ and $Y$) cannot prove that $X$ caused $Y$.

  • Advantages: Allows scientific evaluation of phenomena where active experimental manipulation would be unethical, hazardous, or impossible (e.g., evaluating the association between cigarette smoking and lung cancer incidence).

• Experimental Methodology (True Experiments):

  • Definition: The gold standard of scientific methodology designed explicitly to establish cause-and-effect relationships ($X \rightarrow Y$).

  • Essential Components of a True Experiment:

    1. Manipulation: The researcher actively manipulates one variable (the Independent Variable) by altering its values across experimental conditions.

    2. Control: The researcher exercises strict control over all extraneous variables to ensure that non-manipulated variables do not influence outcomes.

• Detailed Case Example: The Cold Pressor Pain Sensitivity Experiment:

  • Experimental Paradigm: Participants place their non-dominant hand into a bucket of ice-cold water while holding paper objects to test pain endurance.

  • Independent Variable (IV): Type of paper held during cold immersion.

    • Control Condition: Holding green construction paper cut to paper bill size.

    • Experimental Condition: Holding real paper currency (cash money).

  • Dependent Variable (DV): Duration of time (measured in seconds) that the participant can maintain their hand submerged in the freezing water.

  • Classification of DV: Time is a continuous variable measured on a Ratio scale.

Classroom Questions and Student Discussions

• Homework Formatting and Detail Requirements:

  • Question: Is it mandatory to transcribe full word-for-word text problems when submitting homework, or is listing computed steps sufficient?

  • Response: Transcribing full text descriptions is unnecessary. However, every intermediate computational step (e.g., explicitly showing additions, step-by-step subtractions, or column calculations) must be fully documented. Providing only a final number without showing intermediate calculations forfeits credit.

• Appropriate Application of Statistical Averages:

  • Question: How do studies arrive at values such as $2.3$ or $2.5$ children per family if children are discrete units?

  • Response: Fractional child counts occur when researchers incorrectly calculate an arithmetic mean on discrete categorical data. For discrete distributions, the median or mode provides an accurate measure of central tendency without producing impossible fractional values.

• Performance Anxiety and "Board Blindness":

  • Discussion: Performing mathematical computations publicly under social observation (e.g., writing on a classroom white-board) frequently triggers acute situational anxiety, inducing temporary cognitive processing lapses ("board blindness"). Utilizing designated assistants or structured computation tables mitigates these errors.

• Extra Credit Class Recognition:

  • Class participation points for identifying correctly categorized nominal variable categories during lecture exercises were awarded directly to students Garcia, Fraley, Peanuts, Writing, and Dalky.