SOCI 217 Sociological Research Methods Week 6 Readings Social Research Methods, Chapter 7: Quantitative Sampling Introduction While it would be ideal to study a whole population, resources do not allow for that to happen easily so researchers rely on sampling The object of sampling is to get a representative one so that it can be used to make inferences about the population as a whole ->a biased sample is one that is not representative, usually due to researcher error Sampling is incredibly important because it influence the information acquired influences how the researcher perceives the data However, it is impossible to completely eliminate bias, so scientists focus on three specific sources of bias keep their research as unbiased as possible 1. Not using a random method to select a sample can create room for researcher's judgment to influence who is chosen 2. Inadequate sampling frame/subjects can exclude important elements form the sample 3. Nonresponse error, which is not always easy to account for as it is hard reach everyone and hard to know who is actively avoiding research Sampling terms and concepts: Element or unit: a single case of the population sample · In social science, a unit is usually a person however in other fields different things can be sampled o Cities, years, businesses Population: the group of cases who are of concern to the study Sampling frame: the elements for which cases are chosen to be part of a sample Sample: The elements selected for investigation, a subset of the population. The method of selection may involve probability or non-probability sampling Representative sample: one that shows an accurate picture of the larger population in question (best supported by probability sampling methods)
Probability sampling: uses methods that allow researchers to know the chance of each individual being selected for the sample Non-probability sampling: uses methods that indicate certain cases may be more likely than others to be chosen Sampling error: when an error of estimation causes there to be a difference between the the larger population and the sample (can even happen with probability) Non-response error: when for any reason a ppt does not supply the required information for a study and thus skews the data Census: data collected from a whole population (census means all or nearly all) Sampling error Though probability calculations present the most probable outcome, real-life outcomes almost always stray from that outcome more or less Types of probability sample There are several different types of probability sampling Simple random sampling is the most basic from of probability sampling where each unit of a population has an equal chance of being selected · No-room for bias · Does not rely on availability Computer generated tables filled with random numbers are useful for random sampling . Ask TA about the table from textbook on 149 confusing as heck ? "Without replacement" The ratio for sample sizes is formulated as n/N · n is the sample · N is the larger population Systematic sampling is selected directly