CA3011 Communication Arts Research A. Parichart W. A. Chulamani C. Sampling Objectives To explain and differentiate the population and sample To describe the types of sampling procedures; nonprobability sampling, and probability sampling Population and Sample Population One goal of scientific research is to describe the nature of a population- a group or class of subjects, variables, concepts, or phenomena. In many situations, however, an entire population cannot be examined due to time and resource constraints. The usual procedure in these instances is to take a sample from the population. Sample A sample is a subset of the population that is representative of the entire population. A sample that is not representative of the population , regardless of its size, is inadequate for testing purposes because the results cannot be generalized to the population from which the sample was drawn. A Venn Diagram Used in the Process of Sample Selection Population Sample Types of Sampling Procedures Probability VS Nonprobability Sampling Probability sampling uses mathematical guidelines whereby each unit’s chance for selection is known. Nonprobability sampling does not follow the guidelines of mathematical probability. However, the most significant characteristics distinguishing the two types of samples is that probability sampling allows researchers to calculate the amount of sampling error present in a research study: nonprobability sampling does not. Types of Nonprobability Sampling 1. Convenience Sampling A convenience (available) sampling is a collection or readily accessible subjects, elements, or events for study, such as a group of students enrolled in a subject. The nonprobability sampling can be problematic because they contain unknown quantities of error. Researchers need to consider the pros an cons of available samples. E.g. students in the class, shoppers in the mall The available samples do not represent the population . Example Mall intercept studies are criticized because only the people who are at the mall the time of the study have a chance to participate. No one outside the mall has such an opportunity. Research using an unqualified volunteer sample is bad science because there is no way to know who participated in the research study. The results from any study using an unqualified volunteer sample should be consider highly questionable. 2. Purposive Sample Purposive sample, which includes respondents, subjects, or elements selected for specific characteristics or qualities and eliminates those who fail to meet these criteria. • Purposive samples are used frequently in mass media studies when researchers select respondents who use a specific medium and asked specific questions about that medium. Ex. teenagers between the age of 18 to 23 Ex. The sample is those who consumer online media A purposive sample is not representative of the general population. 3. Quota Sampling Subjects are selected to meet a predetermined or known percentage. Ex. A researcher interested in finding out how DVD owners differ from non DVD owners in their use of television may know that 40% of a particular population owns a DVD. The sample the researcher selects, therefore, would be composed of 40% DVD owners and 60% nonDVD owners (to reflect the population characteristics) Example: Search engine market share 4. Snowball Sampling A researcher randomly contacts a few qualified respondents and then asks these people for the names of friends, relatives, or acquaintances they know who may also qualify for the research study. It is recommended for the academic research for a reason that the sample may be completely biased. The sample may consist of respondents who are from a particular club or group. Types of Probability Sampling 1. Simple Random Sampling Where each subject, element, event, or unit in the population has an equal chance of being selected. Researchers often use a table of random numbers to generate a simple random sample. A representative sample may not result in all casess 2. Systematic Random Sampling Every nth subject, unit, or element is selected from a population. Ex. To obtain a sample of 20 from a population of 100, or a sampling rate is 1/5, a researcher randomly selects a starting point and a sampling interval. If the number 11 is chosen as the starting point, the sample will include the 20 subjects or items numbered 11, 16, 21, 26, and so on. Systematic Random Sampling The accuracy of the systematic random sampling depends on the adequacy of the sampling frame, or the complete list of members in the population. - In some projects, researchers want to guarantee that a specific subsample of the population is adequately represented, and no such guarantee is possible using a simple random sampling. -The list can be biased. 3. Stratified Sample is the approach used to get adequate representation of sample (strata or segment) may include almost any variable; age, gender, religion, income level, or even individuals who listen to specific radio stations or read certain magazines. Stratified sampling ensures that a sample is drawn from a homogeneous subset of the population-that is , from a population that has similar characteristics. This requires the researcher to have a complete list of the population. Proportionate VS Disproportionate Stratified Sampling Proportionate stratified sampling includes strata which sizes based on their proportions in the population. Ex. If 30% of the population is adults ages 18-24, then 30% of the total sample will be subjects in this age group. • The procedure is designed to give each person in the population and equal chance of being selected. Proportionate VS Disproportionate Stratified Sampling Disproportionate stratified sampling is used to oversample or over represent a particular stratum. The approach is used because that stratum is considered important for marketing, advertising, or other similar reason. Example In a telephone study of 400 respondents, the station management may wish to have the sample represented as follows: 70% in the 25-34 group, 20% in the 35-49 group, and 10% in the 50-54 group. This distribution would allow researchers to break the 25-34 group into smaller subgroups such as males, females, fans of specific stations, and still have reasonable sample sizes. 4. Cluster Sampling With cluster sampling, the state can be divided into districts, countries, or zip code areas, and groups of people can be selected from each area. Cluster sampling may create the error. Ex. A zip code area may contain mostly residents of a low socioeconomic status who are unrepresentative of the rest of the state. This is suitable when there is no way to obtain the list of the population. Multistage Sampling In many national studies, researchers use a form of cluster sampling called multistage sampling, in which individual households or people are selected. Ex. First, a cluster of countries in the U.S. is selected. Researchers then narrow this cluster by randomly selecting a country, district. Next, individual blocks are selected within each area. Finally, a convention such as “ the third household from the northeast corner’ is established. • Applying the selection formula in the stages just described can thus identify the individual households in the sample. Issues to Consider when Using Probability and Nonprobability Sampling Four Issues when deciding to use Probability or Nonprobability Sampling Purpose of the study. Some studies are not designed to generalize the results to the population but rather to investigate the variable relationships or collect exploratory data to design questionnaires or measurement instruments. Nonprobability sampling is appropriate in these situations. Cost versus value. A sample should produce the greatest value for the least investment. If the cost of probability sampling is too high in relation to the type and quality of info. Collected, then nonprobability sampling is usually satisfactory. Four Issues when deciding to use Probability or Nonprobability Sampling Time constraints. In many cases, researchers collecting preliminary info. operate under time constraints imposed by sponsoring agencies, management directives, or publication guidelines. Since probability sampling is often time consuming, a non probability sample may meet the need temporarily. Amount of acceptable error. In preliminary studies or pilot studies, where error control is not a prime concern, a non-probability sample is usually adequate. Example 1 The comparison of the perception between the Communication Arts students and Business students towards Dove advertising campaign in 2013. Example 2 The study of the impact of the celebrity endorser on the purchase intention Example 3 The study of the value derived from the media of the people in Thailand Reference Wimmer, R. & Dominick, J. (2011). Mass Media Research: An Introduction (9th ed.). Belmont, CA: Thompson Wadsworth. Thank you for your attention In-class Group Advising Develop the research topic which is relevant to the communication area Define the population and sample Define and describe the types of the nonprobability and probability sampling method used