Communication Research: Sampling - Assignment Example

Communication Research: Sampling Sampling or sample size refers to the "selection of a part of the aggregate statistical material with a view to obtain information about the whole" (Das, 2009: 488). It is used in quantitative research studies as a tool for measuring, identifying, estimating, and presenting information collected in a numerical format and analyze the same to derive a logical conclusion or substantiate the research hypotheses. Sampling helps the researcher in establishing a definite link / relationship between the conceptual aspects of the study with the empirical data (Keyton, 2011). Given the context and implications of use of sampling in a research study it is highly imperative to ensure that the sample size chosen is appropriate and adequate in order to derive effective results from the study. Since the analysis of the data relies heavily on the sampling methods used as well as on the sample size it is highly crucial to ensure that the chosen method is apt for the research study. However determining an appropriate sample size is highly tedious process and the researchers are generally found to be vulnerable to random sampling errors. These sampling errors encountered by researchers refer to ascertaining the appropriate sample size and disregarding the response and non-response bias (Wunsch, 1986). The key advantages of sampling is that it allows the researchers to use numbers as a tool to assign value to a given phenomenon and derive meaning out of the numerical data, collected by them (Keyton, 2011). Quantitative research enables the researchers to use/ select small groups or sample size and make effective assumptions / observations or draw appropriate inferences about the larger group and in the process enables them to effectively conclude their study with minimum costs involved (Holton and Burnett, 1997). Researchers use various tools and formulas to determine the most appropriate sample size for their study in order to minimize the risk of errors; however despite such efforts researchers continue to remain vulnerable to random sampling errors and the use of questionable approaches to select the sample size appropriate for their field of study. The key purpose of any study is to achieve multiple objectives rather than derive conclusions regarding a single parameter. Hence choosing an appropriate sample size is a highly complex issue. Furthermore the choice of a sample size is dependent on a range of factors including the homogeneity of the target population, the availability of time, costs involved, as well as personnel available at the disposal of the researchers to aid and assist them in collecting the relevant data (Churchill and Iacobucci, 2009). Validity is one of the key elements in ascertaining and ensuring an appropriate sample size (Keyton, 2011). According to Ryan (2013: p. 33) the term validity with reference to sample size refers to "the confidence interval that contains true parameter value". Cook and Campbell (1979) define the term validity as the "best available approximation of the truth or falsity of propositions, including propositions about cause" (cited in Dattalo, 2007: p. 11) The key goal of sampling is to ensure use of accurate and practical tools to enable effective outcomes from the chosen sample population. There is ample evidence which suggests the inability or failure on the part of the researchers to select an appropriate sample size has often resulted in inconclusive findings due to the failure of the sample population to reject the null hypothesis. The determination of the appropriate sample size hence requires detailed planning and explanatory studies on the part of the researcher to minimize the risk of failure. It also requires the researchers to ensure that they are armed with adequate information that affords them access to detailed information about their study design as well as the primary variables of interest. One of the most effective ways to reduce sampling errors is to choose a sample size larger than the intended sample size, since the observations and conclusions are generalized on the basis of the sample population, the larger sample size helps in minimizing the risks of failure and increases the precision or power of estimation. Although this method may not guarantee effective conclusions merely on the basis of larger sample size but it does however eliminate the probability of bias and prevent overoptimistic estimation on the part of the researcher (Sim and Wright, 2000). Example: Determining a sample size Assuming that the key objective of our study is to examine and ascertain the responses and opinions of a group of university students with regard to the gun control debate. The aim of the study is to examine the proportion of university students who are in favor of gun possession and those who are against it; the key question that needs to be addressed is "How large a sample size do we need?" to answer derive a reliable outcome. In order to estimate the appropriate sample size it is crucial to estimate / ascertain the margin of error. Assuming the margin of error is less than 2.5% the next step is to ascertain the confidence intervals. Typically confidence intervals range from 90 per cent to 99 per cent. For this particular study, the margin assumed is 95 per cent1. Furthermore on the basis of previous / historical studies on similar topics, results observed from online polls regarding public opinion, let us suppose the public opinion is 30 per cent. The formula for calculating sample size: ME = margin of error (i.e. 0.025) Z = z score (1.96) P = previous judgment (0.3) N = sample size (?) Thus, based on the above calculation the estimated sample size i.e. total number of students required for this study is 1291 or approximately 1300 students in total. The problem of selecting and determining an appropriate sample size vary in complexity depending on the type of research study. In such circumstances simulation methods can be effectively applied to address and resolve such problems. There is no permanent solution to address the wide range of problems associated with selecting an appropriate sample size hence the researcher must ensure that certain key questions are addressed prior to selecting a sample population. These include the key objectives and goals of the study, the type of methodology used i.e. qualitative or quantitative; the response variables and the tools planned to be used for measuring the same; the associated risks attached with the chosen sampling; estimate regarding the non-response rates; key sources of variation; as well as time frame among others. Determining an effective sample size entails effective and thorough planning on the part of the researchers and the process is almost invariably involves encountering risks and obstacles. Resolving and addressing such risks require the implementation of scientific techniques and obtaining suitable information and data regarding the study beforehand. Sample size is not always solely dependent on scientific data collection methods but involves an element of ethicality as well. References: Churchill, G. A., Iacobucci, D., (2009). Marketing research: Methodological foundations. Thousand Oaks, CA: Cengage Learning Publication, p. 313-315 Das, N. G., (2009). Statistical methods. McGraw Hill Publication Dattalo, P., (2007). Determining sample size: Balancing power, precision, and practicality: Balancing power, precision, and practicality. London, UK: Oxford University Press. Holton, E. H., Burnett, M. B., (1997). Qualitative research methods, In Swanson, R. A., Holton, E. F., (Eds). Human resource development research handbook: Linking research and practice. San Francisco: Berrett-Koehler Publishers. Ryan, T. P., (2013). Sample size determination and power. London, UK: John Wiley & Sons Publication. Sim, J., Wright, C., (2000). Research in health care: Concepts, designs and methods. Cheltenham, UK: Nelson Thornes Publication, p. 349-353 Wunsch, D. (1986). Survey research: Determining sample size and representative response. Business Education Forum, vol. 40 (5): p. 31-34. Read More