Lecture #4 - Sampling

probability sampling

scientific sampling

non probability sampling

non scientific sampling

element/unit

case in the population

population

the group you want to learn about

sample

a number of units selected from the population from whom data will be collected

census

collected data from all units in the population, but not a sample

probability sample

this is a sample selected using a random process. people have a specific probability to be selected

non probability sample

a sample selected using a non random method

sampling frame

list of all of the units in your population from which a probability sample will be randomly selected.

representative sample

contains the same distribution of characteristics as we see in the population. we need this in order to generalize our findings

sampling error

errors estimation that occur as a result of differences between the characteristics of the sample and those of the population

non response

occurs if an element selected for the sample does not supply the required data

sampling related error

not using a random method to pick a sample. example: you want to interview university students so you go stand on the concrete beach at western. however, this is not a representative sample of western university

sampling frame

inaccurate or incomplete list.

non response error

occurs such that those who participate in the study differ in some important way from those who do not. for example, population is rural communities but indigenous reserves refuse to cooperate

central limit theorem

samples that are close to the true population will be picked more often than samples that are less similar

law of large numbers

picking larger samples always brings them closer to the true population

sampling ratio

what that chance is

sample weight

how many cases in the population each sample unit represents

simple random sample

each unit has equal chance of selection. most basic type of probability sample. example: we have a population of 100 students in research methods and want to sample 10 students to do in depth interviews.

stratified random sample

we divide our population into subgroups (strata) in our sampling frame. each subgroup is sampled separately but all use the same sampling ratio. ensures groups in the population are proportionally represented. but these samples are often not possible

multi stage cluster sample

used for sampling large populations. populations for which there is no adequate sampling frame.

sample size

very important for increasing sample quality and lowering sampling error. absolute size is more important than relative size. as sample size increases, sampling error decreases

heterogeneity of the population

generally the greater the heterogeneity of the population on the characteristics of interest, the larger the sample size will need to be to capture all the differences

convenience sample

cases are included because they are readily available. example: going to a pre natal class to find a sample of pregnant couples

snowball sample

a form of convenience sampling. researcher makes contact with a few people and then they introduce the researcher to more people. this is a strategy often used for sex workers and homeless populations

quota sample

makes a list of the population that have certain characteristics (age, sex, ethnicity, etc.). then uses convenience sampling to collect a sample that matches the proportions found in the population. also known as filling the quota. almost new used in social scientific research

grounded theory (inductive research)

sampling considered an emergent process, not a separate stage. data collection is determined by whatever theoretical or conceptual issues emerge as the study progresses