week 3

Survey Design

• Survey design selects best data collection methods (structured interviews, self-administered questionnaires).
• Interviews offer control and probing, but introduce interviewer bias.
• Questionnaires allow anonymity and standardization, lacking interaction.
• Open-ended questions allow free-form answers, helpful for exploratory research, difficult to quantify.
• Close-ended questions offer pre-defined choices, easier to standardize and analyze.
• Ambiguous wording and double-barreled questions should be avoided.
• Maintain brevity in questions to sustain respondent focus.
• Place sensitive questions later to build rapport.
• Consider response patterns and social desirability biases.

Survey Administration

• Telephone interviews are quick and cost-effective but lack depth.
• Face-to-face interviews provide richer data but are resource-intensive.
• Online surveys are convenient but may have engagement issues.
• Pre-testing helps minimize recall bias and ambiguity in questions.
• Vignettes can encourage honest responses to sensitive topics.
• Consider utilizing secondary analysis (existing data) but acknowledge its potential limitations.

Quantitative Sampling

• Crucial for selecting representative samples allowing findings to be generalised to a larger population.
• Sampling selects a subset instead of the whole population for study.
• Proper techniques minimize bias and enhance findings' validity.
• Probability sampling (random selection) is key.
• Simple random sampling gives every member an equal chance of selection.
• Systematic sampling selects every kth member from a list.
• Stratified sampling divides the population into subgroups for random sampling from each.
• Multi-stage cluster sampling randomly selects groups first, then individuals within those.

Non-Probability Sampling

• Less rigorous, doesn't involve random selection, often used when probability sampling is impractical.
• Convenience sampling uses readily available participants, introducing potential bias.
• Snowball sampling uses initial participants to recruit others, creating dependence and bias.
• Quota sampling selects participants based on specific quotas reflecting population characteristics.

Sampling Challenges

• Sampling error occurs if the sample inaccurately represents the population.
• Larger sample sizes generally improve reliability and accuracy, but come with increased costs.
• Higher response rates are crucial; strategies like follow-ups or incentives improve response rates.

Lecture 3: Populations and Sampling

• A population is a complete group for research, like all university students.
• A sampling element is a single case within a population.
• A sample is a representative subset of a population.
• Clear population definition is crucial; e.g., a study on married couples should exclude unmarried individuals.
• Probability sampling (random selection), like simple random sampling, minimizes bias.
• Non-probability sampling methods, like convenience sampling, are often used due to resource/time constraints. These methods come with the risk of bias.
• Clearly defining the population and appropriate sampling methods (probability or non-probability) are important in a survey.