Communication Research Methods
Communication Research Methods COM 3311
University of Central Florida
Popular in Course
Popular in Communication
This 3 page Class Notes was uploaded by Jerrod Walter on Thursday October 22, 2015. The Class Notes belongs to COM 3311 at University of Central Florida taught by John Butler in Fall. Since its upload, it has received 23 views. For similar materials see /class/227518/com-3311-university-of-central-florida in Communication at University of Central Florida.
Reviews for Communication Research Methods
Report this Material
What is Karma?
Karma is the currency of StudySoup.
Date Created: 10/22/15
Chapter 11 Sampling Why Sample 0 When we sample it allows us to accurately infer characteristics about a large population 0 Cheaper and faster than asking a large population Population all groupsevents Parameter data gathered from the entire population Sample selection of events from a population Statistlc gathered data from a sample Data events gathered from a populationsample Population is a group defined by the researcher Bias is a difficulty when it comes to sampling It s the tendency for a sample to err so that it doesn t represent the population Essentials of Sampling Have to be representative and need to meet 2 criteria 0 They have to accurately represent the population 0 They have to be large enough to permit accurate analysis of the data Sample Size Depends on how much bias you re willing to tolerate Sampling Error is the margin of error associated with the sample Ex Team X leads Team Y 1020 plus or minus 3 That percentage is the sampling error It means there is a error already built in As the sample gets smaller the amount of sampling error gets bigger Confidence Interval a confidence interval of 95 means is that we are 95 confident that this sample accurately represents the population Standard Deviation measures how far from the average the data falls Statistical Effects of Small Samples 0 Small samples increase sampling error o A small sample can make it hard to detect significant relationships Volunteers vs General Population Volunteers have increased need for social approval They tend to have increased socioeconomic status They tend to have increased intelligence Forms of sampling SimpleRandom sampling everyone has an equal chance of being selected Not useful when subjects need to be broken up into subgroups or in a field study where it s impossible to identify everyone individually Systematicperiodic sampling not true random sampling Every ten people who walk through the mall every hour Stratified random sampling Problems with stratified random sampling 0 Uses large samples and gets expensive 0 You must know the size of the sample Cluster Sampling you look for naturally occurring clusters in the environment Examples cities with varying populations 0 Identify the clusters cities with 7500010000 0 Take samples from each of the clusters Forms of non random sampling 0 Convenience samples selection of events that s are readily available Typically have good internal validity but questionable external validity Quota sampling is a lot like stratified random sampling but you don t take a random sample you just pick people who meet your criteria Known group sampling when you know who is in your group Snowball sampling used to study deviant behaviors Advantages 0 Works for studying things that happen naturally Disadvatanges of non random sampling 0 Yield biased results 0 You can t compute the confidence interval 0 Can t generalize results to the population o If you get a refusal rate of over 25 then you have inaccurate data Problems of all samples 0 Subjects can refuse to participate May refuse to do the survey May provide incomplete or inaccurate data ls considerable evidence than non participants attitudes differ from participants attitudes Ethical issues in sampling 0 Sample size Must be large enough to detect relationships but not so large that it makes meaningless relationships sound substantive EX twins cancer and smoking Researcher who used just a few set of twins found no relation between smoking and twins but they only used 20 people and there was a large margin for error