Chapter 5

Got it! Below are very in-depth notes for pages 239-275 of The Practice of Research in Criminology and Criminal Justice by Ronet D. Bachman and Russell K. Schutt.

In-Depth Notes on Sampling (Pages 239-275)

Overview of Sampling

Definition & Importance

Sampling is a systematic process of selecting a subset of individuals from a larger population to analyze and make inferences about the whole group. It allows researchers to draw conclusions efficiently without having to study every single member.

• Goal of Sampling: Ensure that the sample is representative of the population to maximize external validity (generalizability).

• Why Sampling Matters:

• Studying an entire population is often impractical.

• Resource constraints (time, money, manpower) make full-population research difficult.

• Proper sampling methods ensure scientific rigor and reduce bias in research findings.

Key Concepts in Sampling

1. The Population & Sample

• Population: The entire group about which a researcher wants to make inferences.

• Example: All juveniles arrested for violent crimes in the U.S. in 2023.

• Sample: A subset of individuals from the population selected for study.

• Example: A random selection of 500 juveniles arrested for violent crimes in 2023.

2. Sampling Frame

• Definition: The actual list from which a sample is drawn.

• A perfect sampling frame includes every element of the target population.

• Problem: Many sampling frames are incomplete or outdated.

• Example of Sampling Frames:

• List of all inmates in a state prison system.

• National Crime Victimization Survey (NCVS) respondents.

3. Elements & Units of Analysis

• Elements: The individuals or objects being studied.

• Units of Analysis: The level at which data is collected (e.g., individuals, groups, cities).

Types of Sampling Methods

1. Probability Sampling (Scientific, Representative)

Each element has a known and nonzero chance of selection. This ensures randomness and allows researchers to calculate sampling error.

A. Simple Random Sampling (SRS)

• Definition: Every element in the population has an equal probability of being selected.

• Method: Use of random number tables, computer-generated randomization, or lotteries.

• Example: Selecting 200 students from a list of 10,000 students using a random number generator.

• Advantages:

• High representativeness.

• Simple and easy to implement with a complete list.

• Disadvantages:

• Requires a complete sampling frame.

• Can be inefficient for large populations.

B. Systematic Sampling

• Definition: Every nth element is selected after randomly choosing a starting point.

• Method: If you need a sample of 100 from a list of 1,000, you select every 10th name.

• Potential Bias: Hidden patterns in the list can cause unintentional bias (e.g., lists with cyclical orderings).

C. Stratified Sampling (Proportional or Disproportional)

• Definition: Population is divided into strata (subgroups), and samples are drawn separately from each group.

• Types:

• Proportional Stratified Sampling: The sample matches the proportion of each subgroup in the population.

• Disproportional Stratified Sampling: Some strata are oversampled to ensure adequate representation.

• Example:

• Studying racial disparities in the criminal justice system → ensure proportional representation of Black, White, and Hispanic defendants.

• Advantage: Ensures better representation than simple random sampling.

• Disadvantage: Requires detailed knowledge of population demographics.

D. Cluster Sampling (Multi-Stage Sampling)

• Definition: Researchers divide the population into clusters (e.g., schools, precincts) and randomly select whole clusters.

• Use Case: When a full sampling frame is unavailable.

• Example:

• Instead of surveying all police officers in the U.S., randomly select 10 police departments and survey every officer in those departments.

• Advantage: Cost-efficient and does not require a complete list.

• Disadvantage: Higher sampling error due to clustering.

2. Non-Probability Sampling (Non-Random, Non-Generalizable)

Used when probability sampling is impractical, often in exploratory or qualitative research.

A. Convenience Sampling (Haphazard Sampling)

• Definition: Selecting participants based on ease of access.

• Example: Interviewing inmates at one prison instead of a random selection across all prisons.

• Problem: High risk of selection bias.

B. Purposive Sampling (Judgmental Sampling)

• Definition: Deliberate selection based on researcher judgment.

• Example: Studying gang members → purposively selecting known gang members for interviews.

C. Snowball Sampling

• Definition: Existing participants refer new participants, forming a chain.

• Use Case: Hard-to-reach populations (e.g., undocumented immigrants, sex workers).

• Problem: Sample can become homogeneous, leading to bias.

D. Quota Sampling

• Definition: Researchers set quotas to match population characteristics.

• Example: If 40% of the U.S. population is female, researchers ensure 40% of the sample is female.

• Problem: Not truly random, often leads to hidden biases.

Sampling Errors & Biases

1. Sampling Error

• Definition: The difference between sample results and the true population characteristics due to chance.

• Minimization: Larger samples reduce sampling error.

2. Coverage Error

• Occurs When: The sampling frame misses key segments of the population.

• Example: Only surveying landline users excludes younger populations who primarily use mobile phones.

3. Nonresponse Bias

• Occurs When: Certain groups systematically refuse to participate.

• Example: People with criminal records may avoid surveys about past offenses.

• Solution: Follow-up contacts, incentives.

Inferential Statistics & Sampling

1. Sampling Distributions

• Definition: The theoretical distribution of sample statistics if multiple samples were taken.

• Use: Helps estimate population parameters.

2. Standard Error

• Definition: Measures variability between different samples from the same population.

3. Confidence Intervals

• Definition: A range within which the true population parameter likely falls (e.g., “We are 95% confident that the true proportion is between 45% and 55%”).

Practical Applications in Criminal Justice Research

1. Police Use of Force Studies

• Proper sampling ensures representative data on excessive force cases.

2. Crime Victimization Surveys

• Probability sampling is essential for reliable victimization rate estimates.

3. Evaluating Crime Prevention Programs

• Randomized controlled trials (RCTs) use probability sampling to compare outcomes before and after interventions.

Final Summary

• Probability Sampling ensures representative, generalizable results.

• Non-Probability Sampling is used for exploratory research but carries higher risk of bias.

• Errors & Biases (e.g., nonresponse, coverage error) must be minimized to ensure valid research findings.

These are the detailed and thorough notes for pages 239-275. Let me know if you need anything clarified or expanded!