Learning Objectives
448 words
2 min read
Visual companion
Python
Type and operator map
Python Week 1: the first filter for runtime behavior
View
Revision summary
What this note is really saying
Short form
# Learning Objectives - Understand probability and non-probability sampling - Determine appropriate sample size - Evaluate sampling errors and bias - Basic probability theory - Week 2: Research design ## 1. Probability Sampling (Each element has known, non-zero chance of selection) **Simple Random Sampling (SRS):**...
Learning Objectives
- Understand probability and non-probability sampling
- Determine appropriate sample size
- Evaluate sampling errors and bias
- Basic probability theory
- Week 2: Research design
1. Probability Sampling (Each element has known, non-zero chance of selection)
Simple Random Sampling (SRS): Every element has equal chance. Like lottery. Gold standard, but requires complete sampling frame.
Stratified Sampling: Population divided into strata (groups) based on characteristic (e.g., age, region). Random sample from each stratum. Ensures representation of all subgroups. More efficient than SRS.
Cluster Sampling: Population divided into clusters (geographic areas). Randomly select clusters, then sample ALL elements in selected clusters OR sub-sample. More practical for large geographic areas.
Systematic Sampling: Select every kth element after random start. Efficient, but periodic patterns can bias results.
2. Non-Probability Sampling (Selection based on judgment/convenience)
Convenience Sampling: Available participants. Quick, cheap, but highly biased. Use only for exploratory research.
Quota Sampling: Like stratified but non-random. Interviewers fill quotas for subgroups. Common in market research.
Snowball Sampling: Participants refer others. Used for rare or hard-to-reach populations (e.g., executives, niche communities).
Purposive Sampling: Researcher selects participants based on judgment. Used in qualitative research for information-rich cases.
3. Sample Size Determination
Formula for proportion:
Z = Z-score (1.96 for 95% confidence, 2.58 for 99%) p = estimated proportion (0.5 if unknown, gives maximum sample) q = 1 - p e = margin of error (0.05 = +/- 5%)
Example: 95% confidence, 5% margin, p=0.5 n = (1.96^2 * 0.5 * 0.5) / 0.05^2 = 384.16 (need 385)
Q1: Compare stratified and cluster sampling.Stratified: divide by characteristic, sample from ALL strata (ensures representation). Cluster: divide geographically, sample CLUSTERS, then study all in selected clusters (practical for large areas). Q2: When to use snowball sampling?When population is rare, hidden, or hard to reach (drug users, executives of specific companies, people with rare diseases). Participants recruit others in their network. Q3: What sample size is needed for 95% confidence, 3% margin, p=0.5?n = (1.96^2 * 0.5 * 0.5) / 0.03^2 = 3.8416 * 0.25 / 0.0009 = 0.9604/0.0009 = 1067.1. Need ~1068 respondents. Q4: What is sampling error?Difference between sample statistic and population parameter due to chance. Larger samples reduce sampling error. Cannot be eliminated entirely (unlike bias). Q5: What is the difference between sampling error and non-sampling error?Sampling error: random variation due to sampling. Quantifiable, decreases with sample size. Non-sampling error: all other errors (measurement, non-response, processing). Harder to measure, often larger than sampling error. Join Discord PreviousResearch Design & Survey MethodologyNextQualitative Research Methods