Fall 8 Frequency Claims Part 2 Chapter 7 - Tagged

Survey and Observational Research

Part 2

Today

  • Topics include:

    • Population vs. Sample

    • External Validity of Frequency Claims

    • Sampling Techniques

Population vs. Sample

Definition

  • Population of Interest: The entire set of people the researcher is interested in.

  • Sample: A smaller set of individuals taken from the population to represent it.

Example

  • If XX% of students in a class drink coffee daily, this raises the question of whether XX% of the entire FSU student population drinks coffee daily.

  • External validity is key when evaluating frequency claims to ensure that sample results are generalizable to the population.

External Validity of Frequency Claims

Importance

  • Assessing which type of survey has better external validity:

    • A representative sample of 1,000 people vs. a biased sample of 1,000,000 people.

  • Representative samples generally have higher external validity, making them more reliable for frequency claims.

  • Focus on how the sample was obtained rather than just the sample size.

Sampling Techniques

Overview

  • The sampling technique used influences the external validity of a frequency claim.

  • Representative Sample: All members of the population have an equal chance of being included (uses probability sampling).

  • Biased Sample: Some members have a greater chance of being selected (uses biased sampling).

Probability Sampling Techniques

  1. Simple Random Sampling: Equal chance for all in the population (e.g., names drawn from a hat).

  2. Systematic Sampling: Choose a starting point and select every nth participant (e.g., every 4th person).

  3. Cluster Sampling/Multistage Sampling:

    • Divide the population into arbitrary groups, then randomly select some groups.

    • Multistage: Further random sampling from selected groups.

  4. Stratified Random Sampling/Oversampling:

    • Select specific demographics and randomly select within each category proportionately.

    • Oversampling increases the representation of smaller groups (e.g., political orientation).

Application Examples

  • The researcher obtains a database to randomly select registered voters.

  • If the aim is to match proportions of political party affiliation, stratified random sampling would be appropriate.

  • Oversampling may be used to ensure minority demographics are adequately represented.

Biased Sampling Techniques

Overview

  1. Convenience Sampling: Using readily available participants (e.g., polling FSU students).

  2. Purposive Sampling: Selecting specific demographics and recruiting at targeted locations (e.g., Starbucks).

  3. Snowball Sampling: Existing participants help recruit future participants (e.g., independent voters).

  4. Quota Sampling: Setting targets for specific demographics, usually through non-random sampling methods until quotas are met.

Application Examples

  • Finding registered voters through recommendations in snowball sampling.

  • Using convenience sampling via public spaces or events to gather participants.

Practice Scenarios

Activity

  • Review survey scenarios based on ghosting behaviors and identify samples and populations:

    • Plenty of Fish survey of 800 adults shows 80% of millennials have experienced ghosting.

    • YouGov survey of 1,000 adults indicates lower rates of ghosting.

    • Creating a sample from 235 million adults in the U.S. by asking about ghosting experiences, explore different collection methods (simple random, cluster, stratified, systematic, convenience).