6B Flashcards
What is a sample in research?
A subset of cases from a larger population used to draw conclusions about that population.
What are the two key characteristics of a good sample?
- Large enough to provide reliable estimates.
- Representative enough to generalize findings.
What are the two core questions when recruiting a sample?
- How many cases should be recruited?
- How will these cases be selected?
What determines the required sample size?
- Effect size – Smaller effects require larger samples.
- Desired statistical power – Larger samples reduce the risk of missing a real effect.
Does the required sample size depend on population size?
No, sample size depends on effect size and statistical power, not on population size.
What is a convenience sample?
A sample based on availability, such as recruiting participants from friends, social media, or self-selected online polls.
What are the strengths and limitations of convenience sampling?
Strengths – Cheap, easy, allows larger sample sizes.
Limitations – Highly unrepresentative, difficult to generalize results.
What is quota sampling?
A method where researchers set quotas for certain characteristics (e.g., gender, political preference) to ensure representation in the sample.
What are the strengths and limitations of quota sampling?
Strengths – More representative than convenience sampling at a lower cost.
Limitations – Only ensures representativeness on chosen characteristics, requires knowledge of population distributions.
What is a probability sample?
A sample where all members of the population have an equal chance of selection, ensuring representativeness.
What are the strengths and limitations of probability sampling?
Strengths – Gold standard for representativeness.
Limitations – Expensive, requires a complete population register, affected by non-response bias.
What is non-response bias?
When certain groups systematically refuse to participate, making the sample unrepresentative (e.g., politically uninterested people are less likely to respond to election surveys).
What is weighting in survey research?
A statistical adjustment that assigns different weights to respondents to correct for underrepresentation of certain groups.
When should data be weighted?
- When analyzing mean levels (e.g., percentage of voters) rather than relationships.
- When examining variables affected by non-response bias (e.g., voter turnout).
- When using unrepresentative samples (e.g., convenience samples).
What is an example of data weighting?
If non-voters are underrepresented, their responses can be given higher statistical weight to better reflect the actual population.
What is a criticism of extreme data weighting?
If the original sample is highly unrepresentative, weighting may not fully correct for bias and can even introduce new biases.
What are the three main survey modes?
- Structured interviews (in-person)
- Paper questionnaires (mail surveys)
- Web-based surveys (online surveys)
What are the advantages and disadvantages of structured interviews?
Advantages – Allows clarification of questions, accessible to illiterate respondents.
Disadvantages – Expensive, time-consuming, risk of social desirability bias.
What are the advantages and disadvantages of paper questionnaires?
Advantages – No interviewer bias, can reach offline populations.
Disadvantages – Expensive, inconvenient for respondents.
What are the advantages and disadvantages of web-based surveys?
Advantages – Cheap, efficient, easy to administer.
Disadvantages – Excludes people without internet access (mostly elderly).
Which survey mode is now the most common?
Web-based surveys – Research shows they produce results as reliable as traditional methods.
What is an expert survey?
A survey where experts rate a topic instead of collecting responses from the general public.
What are the strengths and limitations of expert surveys?
Strengths – Cost-effective, provides insight into topics where public surveys are impractical.
Limitations – Subjective, potential for bias, results can be circular.
What is quantitative content analysis?
A method for systematically analyzing text or media by converting qualitative data into numerical data (coding).
