Sampling Methods and Generalizability

Participants in Research

Population Identification

• Researchers must define the population of interest.
• Examples:
  • Men and women over 70 in romantic relationships.
  • Bisexual women aged 25-40.

Sampling from the Population

• A sample is chosen from the target population.
• The sampling method determines the generalizability of findings.

Convenience Sampling

• Participants are selected based on their accessibility and willingness to participate.
• Example: Psychology 100 subject pool.
• Involves asking people to volunteer.
• Potential for volunteer bias.

Volunteer Bias

• Volunteers for sexuality research tend to be:
  • More likely to be male.
  • Younger.
  • More sexually experienced.
  • More comfortable with sexual topics.
  • More liberal in sexual attitudes.
• Kinsey's studies used convenience sampling.

Probability Sampling

• Each person in the population has a specific chance of being chosen.
• Types:
  • Random sampling.
  • Stratified random sampling.
• The National Health and Social Life Study (NHSLS) used probability sampling.

Random Sampling

• Every person has an equal chance of being chosen.
• Example: Attitudes toward marriage equality in New York City and Hoboken, New Jersey.
• Population of New York City: $8,538,000$. Population of Hoboken: $54,379$.
• Procedure: Place each person's name from both cities in a hat, shake, and draw 1,000 names.
• Problem: May not yield enough participants from smaller populations (e.g., Hoboken) for meaningful comparisons.

Stratified Random Sampling

• Divides the population into subgroups (strata) and randomly samples from each.
• In the marriage equality example, create separate hats for New York City and Hoboken.
• Draw 500 names from each hat.
• Ensures adequate representation from each group for comparisons.
• NHSLS used stratified random sampling.
• Useful for:
  • Comparing gay men and lesbians to heterosexuals.
  • Studying racial and ethnic minorities.

Representative vs Generalizable

• Comparison of studies on frequency of sex (at least once a week).
• Janus survey: Volunteer sample (convenience sample).
• GSSNORC (National Opinion Research Center): Probability sample.

Volunteer Bias Illustrated

• Age group 18-26:
  • Janus (convenience sample): $72\%$ of men report having sex once a week.
  • GSSNORC (probability sample): $57\%$.
• Age group 39-50:
  • Janus (convenience sample): $83\%$ of men.
  • GSSNORC (probability sample): $56\%$.
• Age group 65+:
  • Janus (convenience sample): $69\%$ of men.
  • GSSNORC (probability sample): $17\%$.
• Women over 65:
  • Janus (convenience sample): $74\%$.
  • GSSNORC (probability sample): $6\%$.
• Sample selection affects the generalizability of findings to the population.