Chapter 5: Sampling & Generalizability

Key Terms & Concepts:

• Sample: A subset of individuals from a larger population used to make inferences about the whole.

• Sampling Frame: A list or database from which a sample is drawn (e.g., a list of registered voters).

• Element: The individual unit of analysis in a sample (e.g., a person, a household, a school).

• Sampling Error: The difference between sample results and the true population values due to chance.

• Sampling Unit: The entity selected at each stage of sampling (e.g., individuals, groups, organizations).

• Representative Sample: A sample that accurately reflects the characteristics of the population.

Generalizability:

• Sample Generalizability: The extent to which findings from a sample apply to the larger population.

• Cross-Population Generalizability (External Validity): The extent to which findings apply across different populations or settings.

Types of Sampling Methods:

Probability Sampling (Random Selection)

1. Simple Random Sampling: Every element has an equal chance of being selected.

2. Systematic Random Sampling: Selecting every kth element from a list (e.g., every 10th person).

3. Stratified Random Sampling: Dividing the population into strata (e.g., age groups) and randomly sampling within each stratum.

   • Proportionate: Sample proportions match the population proportions.

   • Disproportionate: Some strata are oversampled for better representation.

4. Cluster Sampling: Groups (clusters) are randomly selected, then individuals within them are sampled.

Non-Probability Sampling (Non-Random Selection)

1. Convenience Sampling: Selecting individuals based on ease of access.

2. Purposive (Judgmental) Sampling: Selecting individuals who fit a specific criterion.

3. Snowball Sampling: Participants recruit others (useful for hard-to-reach populations).

4. Quota Sampling: Ensuring the sample meets a predetermined demographic quota.

Chapter 6 & 7: Causation & Experimental Designs

Criteria for Establishing Causation:

1. Empirical Association: A relationship exists between variables.

2. Temporal Order: The cause must precede the effect.

3. Non-Spuriousness: No alternative explanations (confounding variables) exist.

Identifying Cause & Effect:

• Example: If studying the effect of sleep on test scores:

  • Cause: Amount of sleep

  • Effect: Test performance

Components of a True Experiment:

1. Independent & Dependent Variables (Cause & Effect).

2. Pretest & Posttest Measures (Before & After).

3. Random Assignment (Participants randomly assigned to groups).

4. Experimental & Control Groups (Treatment vs. No treatment).

Quasi-Experimental Designs (When Randomization Isn't Possible):

• Nonequivalent Control Group Design: Uses pre-existing groups instead of random assignment.

• Before-and-After Design: Measures the same group before and after a treatment.

• Time-Series Design: Multiple observations before and after the intervention.

Internal Validity & Threats:

• Internal Validity: The degree to which a study accurately shows a causal relationship.

• Threats to Internal Validity:

  • Selection Bias (non-random groups).

  • History Effects (outside events affecting results).

  • Maturation (natural changes over time).

  • Testing Effects (learning from previous tests).

  • Instrumentation (changes in measurement methods).

  • Regression to the Mean (extreme scores shifting toward average).

Longitudinal Research Designs:

1. Trend Studies: Data collected at different times from different samples.

2. Panel Studies: The same individuals are followed over time.

3. Cohort Studies: A specific group is studied over time (e.g., people born in 2000).

Chapter 8: Surveys & Data Collection

Why Use Surveys in Social Science?

• Collect data efficiently from large populations.

• Standardized questions ensure comparability.

• Allow for statistical analysis of relationships.

UCR & NCVS (Crime Data Sources):

• Uniform Crime Report (UCR): Official police-reported crime data.

  • Strengths: Nationwide data, good for tracking trends.

  • Weaknesses: Underreporting due to unreported crimes.

• National Crime Victimization Survey (NCVS): Self-reported victimization survey.

  • Strengths: Captures unreported crimes.

  • Weaknesses: Memory issues, exaggeration, or misreporting.

What Makes a Good Survey Question?

• Avoid:

  • Leading questions (e.g., "Don't you agree that…?")

  • Double-barreled questions (asking two things at once).

  • Ambiguous wording.

  • Social desirability bias (people responding in a way they think is acceptable).

Benefits of Combining Questions into an Index:

• Increases reliability.

• Captures broader concepts.

• Reduces the impact of individual question bias.

Types of Surveys (Strengths & Weaknesses):

1. Mail Surveys:

   • Strengths: Cost-effective, anonymous.

   • Weaknesses: Low response rates.

2. Telephone Surveys:

   • Strengths: Fast, easier to clarify responses.

   • Weaknesses: Declining response rates, bias toward people with landlines.

3. Face-to-Face Interviews:

   • Strengths: High response rate, detailed answers.

   • Weaknesses: Expensive, interviewer bias.

4. Online Surveys:

   • Strengths: Quick, inexpensive.

   • Weaknesses: Sample bias (only internet users).

5. Mixed-Mode Surveys:

   • Strengths: Combines methods to increase responses.

   • Weaknesses: Can be complex and costly.