Measurement

Definition and Purpose of Measurement

• Measurement Defined: A procedure whereby a researcher assigns numbers or symbols to the empirical properties of variables according to a prescribed set of rules.

• Objective of Measurement:

  • To quantify abstract phenomena in social research.

  • To concretize theoretical concepts so they can be tested empirically.

The Theoretical, Operational, and Empirical Continuum

• Three Levels of Conceptual Abstraction:

  • Theoretical Level: Concerns abstract constructs and their conceptual definitions. At this level, researchers establish a hypothetical causal relationship between independent and dependent abstract constructs.

  • Operational Level: Bridges theoretical concepts and concrete empirical indicators. Moves from conceptual definitions to operationalization procedures.

  • Empirical Level: Consists of actual indicators or measures used to test empirical hypotheses between independent and dependent variables.

• Processes of Translation:

  • Conceptualization: The transition from an abstract construct down to a formal conceptual definition.

  • Operationalization: The transition from a conceptual definition down to a specific indicator or measure.

  • Abstraction Gradient: Moves continuously from More Abstract (at the theoretical level) to More Concrete (at the empirical level).

Theory, Concepts, and Variables

• Theory:

  • Definition: A set of interrelated statements intended to explain some aspect of social life.

  • Role in Research: Helps define concepts and variables, and aids in formulating hypotheses regarding expected relationships among variables.

• Concepts:

  • Definition: Abstract ideas that classify general observations or experiences.

  • Examples: Social status, power, intelligence, satisfaction.

  • Operational Process: Abstract concepts must be defined and operationalized into measurable variables.

• Variables:

  • Definition: A concept that varies, meaning it possesses two or more distinct values or categories.

  • Key Property: Variables have been operationalized so that precise rules of measurement are established.

  • Examples: Income, race, gender, sexual orientation, IQ, marital status, crime rate, death rate, SAT score, GPA, Gross National Income.

Conceptual and Operational Definitions

• Conceptual Definition:

  • Definition: Abstractions articulated into words; states explicitly what a researcher means by a concept.

• Operational Definition:

  • Definition: A specific set of instructions detailing how to measure a variable that has been conceptually defined.

  • Function: Explains how data will actually be gathered and calculated in practice.

Concrete Examples of Operationalization

• Example 1: Trade

  • Conceptual Definition: The buying and selling of goods and services.

  • Operational Definition: Trade is measured as the sum of all imports and exports of goods and services of a country as a proportion of that country's gross domestic product (GDP) in current $USD$:

$\text{Trade} = \frac{\text{Imports} + \text{Exports}}{\text{GDP}}$

• Unit of Variation: Varies by one unit of current $USD$

  • Example 2: Democracy

• Conceptual Definition: A political system that:

  1. Replaces leaders through free and fair elections.

  2. Features the active participation of citizens.

  3. Protects human rights.

  4. Follows the rule of law.

• Operational Definition: A composite measure combining political rights and civil rights:

  • Political Rights: The degree to which a nation is governed by democratically elected representatives and maintains fair, open, and inclusive elections.

  • Civil Liberties: A scale measuring freedom of the press, freedom of assembly, general personal freedom, freedom of private organizations, and freedom of private property.

  • Scale of Measurement: Rated on a numeric scale from $1$ to $7$, where $7$ represents Highly Democratic and $1$ represents Low Democracy.

Variables and Their Attributes

• Operational Choices in Design:

  1. Determining the range of variation.

  2. Specifying the number of attributes.

  3. Defining the number of indicators (if applicable).

• Variable Attributes:

  • Definition: The specific conditions, values, or categories contained within a variable.

  • Rule of Alignment: Decisions regarding the variation of attributes must directly align with the theory, research questions, targeted population, and study purpose.

  • Examples of Attribute Structures:

  • Gender Variable ($4$ possible attributes/conditions): Male, Female, Nonbinary, Prefer Not to Answer.

  • Class Variable ($3$ possible attributes/conditions): Upper Class, Middle Class, Lower Class.

Independent and Dependent Variables

• Independent Variable (IV):

  • Represents the predictor or presumed cause.

• Dependent Variable (DV):

  • Represents the outcome or presumed effect.

• Directional Flow:

$\text{Independent Variable (IV)} \rightarrow \text{Dependent Variable (DV)}$

• Causal Dependency: The value of the dependent variable is hypothesized to depend upon the value of the independent variable.

Formulation and Evaluation of Hypotheses

• Hypothesis:

  • Definition: A formal statement predicting the expected relationship among variables in a study, derived from social theories or prior empirical literature.

  • Requirements for a Good Hypothesis: Must state both the expected relationship and the specific direction of that relationship relative to the variable attributes.

• Evaluating Hypothesis Quality:

  • Statement 1: "Gender is positively related to support for gender equality."

  • Evaluation: BAD ($\text{BAD} \boldsymbol{\times}$). Gender is a categorical variable without an intrinsic numerical direction; the hypothesis must explicitly compare specific attributes to a reference category.

  • Statement 2: "Women are more likely to support gender equality than men."

  • Evaluation: GOOD ($\text{GOOD} \boldsymbol{\checkmark}$). Correctly compares specific attributes (women vs. men).

  • Statement 3: "Age is related to attitudes toward same-sex marriage."

  • Evaluation: BAD ($\text{BAD} \boldsymbol{\times}$). Fails to specify the nature or direction of the relationship.

  • Statement 4: "Age is negatively associated with support for same-sex marriage. As age increases, there will be less support for same-sex marriage."

  • Evaluation: GOOD ($\text{GOOD} \boldsymbol{\checkmark}$). Explicitly states the negative direction of the association.

• Comprehensive Theoretical-Empirical Model (Global Inequality Example):

  • Theoretical Basis: Dependency Theory (Frank 1967), which posits that economic dependency stems from unequal economic exchanges between nations.

  • Conceptual Proposition: International debt deteriorates living conditions within developing countries.

  • Research Hypothesis: Increased total bilateral and multilateral debt load will be associated with increased poverty in low-income countries.

  • Independent Variable: Total debt load, operationalized as:

$\text{Debt Load (\%)} = \frac{\text{Total Bilateral and Multilateral Debt}}{\text{Total Exports of Goods and Services}} = \frac{\text{TDS}}{\text{XGS}}$

• Dependent Variable: Poverty, operationalized as the headcount living at or below the national poverty line expressed as a percentage of the total population.

Null Hypothesis Testing and Logic

• The Null Hypothesis ($H_0$):

  • Definition: Asserts that there is no relationship or no significant difference between the specified variables.

  • Research Hypothesis ($H_1$): States the expected relationship or difference.

• Comparative Examples:

  • Example 1: Environmental Regulation

  • Research Hypothesis ($H_1$): Countries that adopt the Basel Ban Amendment on Toxic Waste ($B$) will experience lower pollution ($P$) than those that do not adopt the amendment.

  • Null Hypothesis ($H_0$): There is no difference in pollution levels ($P$) between countries that have adopted the Basel Ban Amendment ($B$) and those that have not.

  • Example 2: Sociological Demographics (Durkheim)

  • Research Hypothesis ($H_1$): Catholics ($C, \mu_1$) have significantly larger families than Protestants ($P, \mu_2$).

    • Notation: $H_1: C > P$ or $H_1: \mu_1 > \mu_2$

  • Null Hypothesis ($H_0$): There is no significant difference in family size between Catholics and Protestants.

    • Notation: $H_0: C = P$ or $H_0: \mu_1 = \mu_2$

• Logic of Null Hypothesis Testing:

  • Falsification Principle: Scientific methodology focuses on falsifying untrue hypotheses rather than definitively proving true ones.

  • Strength of Support: Rejecting the null hypothesis (and eliminating alternative explanations) strengthens the empirical support for the proposed research hypothesis.

Relationships, Direction, and Correlation

• Directionality of Relationships:

  • Positive (Direct) Relationship: Both variables change in the same direction ($\uparrow\uparrow$ or $\downarrow\downarrow$).

  • Negative (Indirect) Relationship: Variables move in opposite directions ($\uparrow\downarrow$).

• Correlation:

  • Definition: Occurs when two variables covary or are systematically associated.

  • Core Principle: Correlation does not prove causation. Establishing a correlation does not inherently demonstrate that the independent variable directly causes the change in the dependent variable, though it may represent a component of a causal pathway.

• Correlation Coefficients:

  • Quantifies the strength and direction of an association on a numerical scale.

  • Values closer to $1$ or $-1$ denote stronger associations.

  • Strong Positive Correlation Example: $r^2 = 0.9991$

  • Moderate Negative Correlation Example: $r^2 = -0.5777$

Control Variables and Spurious Relationships

• Control Variables:

  • Definition: Potential explanatory variables included in an analysis to account for potential confounding effects on the dependent variable.

• Spuriousness:

  • Definition: A mathematical or statistical relationship in which two variables ($X$ and $Y$) have no direct causal connection, yet it is falsely inferred that they do because both are caused by an unmeasured third variable ($Z$).

  • Impact of Control: When controlling for the third variable ($Z$), the original association between $X$ and $Y$ vanishes.

• Illustrative Examples of Spuriousness:

  • Example 1: Firefighters and Damage

  • Observed Association: Number of Firefighters ($X$) correlates positively with Amount of Fire Damage ($Y$).

  • Control Variable: Size of Fire ($Z$).

  • Explanation: A larger fire causes both a higher Number of Firefighters to be dispatched and a greater Amount of Fire Damage. Controlling for Size of Fire eliminates the spurious link between firefighters and damage.

  • Example 2: Ice Cream and Crime

  • Observed Association: Ice Cream Sales ($X$) correlate positively with Crime Rate ($Y$).

  • Control Variable: Temperature ($Z$).

  • Explanation: Warmer weather causes both an increase in ice cream consumption and an increase in outdoor public activity/crime rates. Controlling for temperature reveals the direct relationship to be spurious.

Measures of Association

• Definition: Metrics that quantify the degree to which two or more variables covary.

• Scale and Magnitude:

  • Values range from $0$ (complete absence of relationship) to $1$ or $-1$ (perfect linear relationship).

  • Correlation measures such as the coefficient of determination ($R^2$) quantify the proportion of variance shared by variables.

• Linear Fit Examples:

  • Perfect Fit: $R^2 = 1$

  • Strong Empirical Fit: $R^2 = 0.748$ (e.g., plotting Number of Hours of Study per Day ($X$) against Number of Excellent Grades ($Y$))