Biology 2244A/B Lecture 2: STATS-L2
23 Jan 2020
School
Department
Course
Professor
Sampling Designs & Considerations -1
Sampling Frame: set of all individuals for which data possibly collected
1. Outside circle-population of interest (eg. whole human population)
→all individuals interested in studying
2. Middle circle-sampling frame (eg. university students in western)
→each one had a chance of being in study eg. sending email out to entire uni but
only some want to participate
*time, money, access make it difficult to have everyone so make a smaller sample
3. Inner circle-sample (eg. 30 participants) →those that are actually in the study
Random: outcome uncertain but there is a distribution of chance in each of the outcomes
eg. each side has 1/6 chance “regular distribution” →set probability of occurring, predictable
Fair Die: each side has equal chance
Loaded: some sides more likely then others (still random chance)
**random in context of sampling
→in context of random sampling/SRS it applies a quality (equal) chance but in any other
context, doesn’t have to be equal just a distribution of chance
Simple Random Sample (SRS): all individuals are equally likely to be chosen
eg. 12 unique individuals and picking sample of 4
Sampling with Replacement: once get a sample, put back and eligible to be chosen again
→can end up with sample that get same individual each time
Document Summary
Sampling frame: set of all individuals for which data possibly collected: outside circle-population of interest (eg. whole human population) All individuals interested in studying: middle circle-sampling frame (eg. university students in western) Each one had a chance of being in study eg. sending email out to entire uni but only some want to participate. *time, money, access make it difficult to have everyone so make a smaller sample. Inner circle-sample (eg. 30 participants) those that are actually in the study. Random: outcome uncertain but there is a distribution of chance in each of the outcomes eg. each side has 1/6 chance regular distribution set probability of occurring, predictable. Loaded: some sides more likely then others (still random chance) In context of random sampling/srs it applies a quality (equal) chance but in any other context, doesn"t have to be equal just a distribution of chance.
