Research Exam 2 UNIT 1 Sampling Element Sample Population Universe o Procedure when you select a sub population of elements from a total population o Goal infer characteristics of a population through measurements of a sub population of possible elements of that population a sample o A single member of a population or the unit of things chosen for a sample people events objects o Some part of a larger body of elements specifically selected through a variety of possible methods to represent the whole o A group of elements that are selected for study o All possible elements that can be chosen from based on some designated set of specifications o Specifications are specified by the researcher o A hypothetical group of elements o The abstract population that we want to generalize o Ex all inmates released from prison in the U S Why We Sample o Easier quicker less costly than measuring the population as a whole o It is also more accurate cuz it eliminates interviewer bias and fatigue 2 Broad Types of Sampling o Probability Samples that permit estimation of the likelihood of each element of the population being selected in the sample One can specify for each element of the population the probability that it will be included in the sample Simplest case each element has same probability of inclusion in sample Only with probability sampling can you determine the size of the sample needed for a given level of accuracy How different your results might be compared to the population o Non Probability No way of knowing or estimating the probability or likelihood that each element has of being included in the sample No assurance that every element has some change of being included Can t estimate risk of error thus can t do statistical inference Only useful for exploratory research or analysis of extreme cases Usually done for convenience economy or it s the only method possible Types of Probability Sampling o Simple Random Sampling Each element in the population has the same probability of being selected in the sample Must have a list of all elements in the population all felons in Florida o Systematic Sampling Determine how many cases you want or can sample then use a random number table or random number generator to select the cases for measurement Can now select cases randomly with statistical software packages such as SPSS Select every nth case as long as there is no pattern to how the cases are ordered Randomly select which case you begin with to make it a random sample Need to make sure there is no order to the listing of cases in your population Ex if you studying a population of married couples and the couples were listed in order of husband wife husband wife and you selected every even numbered case you would get only females o Stratified Random The population is divided into 2 or more strata based on relevant criteria A simple random sample is taken from each stratum and the sub samples are Strata are mutually exclusive groups within your population then joined to form the total sample Ex you might sample by crime type or on some offender characteristic such as gender age race You want to select within strata that you believe are relevant to your study 2 Forms of Stratified Random Sampling o Proportionate Stratified o Disproportionate Stratified Sample subjects are picked in same ratio that they exist in population If you have a stratum of crime types and robbery is 15 of your population you would randomly select within your robbery stratum and it would comprise 15 of your total sample Take a larger than proportionate number of certain groups to assure the appearance of a sufficient number of cases for comparative purposes of a group that is small in population Ex you might need to oversample on low frequency crime types such as murders Why do we stratify It increases the level of accuracy of sample over simple random sampling The things we infer from the sample have a lower probability of being incorrect This can be proven mathematically o Cluster Sampling selected for study Population is divided into clusters and then a probability sample of cluster is The researcher then samples the elements within each of the clusters chosen Less time consuming and requires fewer resources cuz the geographical region is smaller More error with cluster doe You may have an atypical cluster that causes more error than an atypical element from a simple random sample 2 Types of Cluster Sampling Single Stage Cluster Sampling o Select one set of clusters o Ex you are studying sentencing policy in Florida you could select a random sample of the 20 judicial circuits in the state Multi Stage Cluster Sampling o When you sample some number of clusters within broader clusters o Ex Each judicial circuit has 1 or more counties that comprise the circuit So you might sample some number of counties within some number of the 20 judicial circuits in Florida Types of Non Probability Sampling o Accidental Samples No set basis for case selection Take cases that fall into your hand Ex asking people coming our of a movie theater questions about crime and No way of knowing the level of bias and if the results can be generalized to a punishment large population o Quote Sampling A non probability stratified sample Judgmentally select elements to fit a distribution equal to what the population s distribution is on key variables It s difficult to sample on more than a few dimensions The actual proportions in the population may differ from what you believed it to be o Purposive Samples are studying o Snowball Sampling AKA judgmental sampling The researcher selects cases they consider representative of the population they want to generalize about based on informed opinion One strategy is to pick cases that are judged to be typical of the population you Choosing a 1st subject then another one based on introduction from the 1st etc Ex if you are studying homeless people gang members people who visit brothels you could start with one person in these types of populations asking them for a name of another person etc Population Perimeter o The true value or statistic for an entire population o Statistic mean median or correlation o Ex if you are measuring the level of Fear of Crime in a community and selected every citizen in the community you could derive a measure of the average Fear of Crime based on the responses to a questionnaire administered to the entire population That is a Population Parameter of the level of Fear of Crime
View Full Document
Unlocking...