Comm Research Week 6 Sampling and Surveys Part 1 Lecture (Oct 9, 2025)
Surveying Difficult Topics
Topics covered in surveys can include:
Medical topics
Mental health issues such as depression
Subjects related to violence
Challenges associated with studying these topics:
Reluctance of individuals to participate in surveys concerning sensitive issues.
Population Identification Challenges
Focus on a specific population (e.g., US adults with diabetes).
Difficulties faced:
Identifying the target population can be problematic.
The exact size of the target population may be unknown.
Some critical points to consider regarding sampling:
The population from which samples are drawn can sometimes be unknown.
Estimation methods may be required in such cases.
Historical Context of Survey Bias
Historical examples of survey bias include:
Surveys conducted during the FDR era, where non-respondents were often supporters of FDR, leading to biased results.
Sampling methods such as those using telephone directories may exclude certain populations (e.g., households without telephones).
Drawing from specific sub-populations can yield different findings (e.g., health information seeking behaviors in the Hispanic migrant population in LA County vs broader national samples).
Sampling Techniques and Challenges
Importance of sample alignment:
Example: Matching survey sample statistics from Georgia and South Carolina with general population statistics.
Discussion of the randomness in samples:
Discussed likelihood of making correct inferences (50-50 chance) based on random selection practices.
Survey Design Challenges
Presentation of different sampling approaches:
Random digit dialing for certain populations due to non-response bias.
Example of using random sampling methods to obtain data:
A survey aiming to understand respondents’ financial situations with a bounded amount between 0 and 9 dollars.
Distrust among respondents leading to general underrepresentation of specific groups.
Significant errors observed in polling results (3-5% inaccuracies) can lead to misleading conclusions, particularly in elections.
Sampling Methodology Fundamentals
Definition of key sampling terms:
Representative Sample: A subset of the population that accurately reflects the larger group.
Random Sampling: Each individual in the population has an equal chance of being selected.
Importance of probability theory in sampling:
Probability theory facilitates inferences about population characteristics based on sample data.
Enables the estimation of how accurately sample estimates represent the overall population.
Accuracy and Margin of Error
Explanation of margin of error:
Example:
A sample of 2,000 drawn from a total population of 100,000,000 can be indicative of voting preferences within a close approximation (such as +/- 3%).
The concept of weighting in polling:
Adjusting sample responses based on demographic characteristics to match the population accurately.
The necessity of ensuring adequate representation:
Survey samples must accurately reflect age, education, and racial demographics of the broader population.
Non-Random Sampling Techniques
Types of non-random sampling mentioned:
Convenience Sampling: Involves participants available for the study, often leading to unintentional bias.
Volunteer Sampling: Participants voluntarily choose to join the study, which can introduce self-selection bias.
Quota Sampling: Researchers segment a population to ensure representation of specific characteristics.
Snowball Sampling: Existing study subjects recruit future subjects from among their acquaintances.
Examples of how bias can emerge in non-random samples, especially in student-driven surveys or voluntary studies.
Ethical Implications and Institutional Review Boards (IRBs)
Discussion of ethics in research, particularly about using convenience sampling techniques that solicit students:
Importance of avoiding coercion in student surveys, as seen in relationship to extra credit incentives.
Considerations regarding representativeness of volunteer samples in broader populations.
Research Findings Reporting Limitations
Common issues with research findings reporting in media:
Media often simplifies complex polling data, ignoring nuances such as sampling error and population representativeness.
Potential misrepresentation of statistics, leading to public misunderstanding about survey accuracy and implications.
Conclusion
Necessity for critical evaluation of survey methodologies:
A reminder to carefully assess surveys and research findings, particularly regarding shortcomings due to sampling biases, methodology, and representative concerns.
Acknowledging limitations and potential biases in data interpretation is essential for accurate understanding.