Difference between Probability Sampling and Non-Probability Sampling - Essay Example

? RESEACH METHODS Difference between probability sampling and non-probability sampling Probability sampling is a method of statistical sampling that uses any of the random sampling techniques to arrive at the target population. This is a sampling method that presents an equal opportunity for the different units of the population targeted an equal chance of selection or choice. This means that when this method is used in a piece of research every member of the population has a chance of being surveyed for the research conducted. This type of sampling method gives us the ability of calculating the sampling error since we know the probability through which each sample was arrived at. A valid example of application of the method would be when we want to seek the opinion of married men about an issue; the sampling population is all the married men who have an equal probability of being researched on (Ross & PekoI?z, 2007). Non-probability sampling on the other hand is a sampling method, which presents itself in different shapes and sizes. This method does not present an equal chance for each member or unit of the population of investigation to be surveyed i.e. it is biased. This method is that which most researchers employ since it is less tedious and quite cheap. Non- probability sampling method requires that a hypothetical model is designed that fits the data defined, the validity of the model defined may also be extremely difficult to verify. This is because there is the definition of the sample for investigation is done based on convenience through a selection of a unit, which fits within the block selected. An example of this takes place when survey is passed to someone and the person asked to send the same to a friend to expand the target population (Ross & PekoI?z, 2007). Unobtrusive research This is a form of non- reactive research method and it is employed on data collection majorly in social research. This form of research do not in any way interfere with the human research subjects because they are not the direct source of the required data for analysis. The required data is obtained indirect ways through relevant methodologies to get the necessary data from the research subjects. In social research, unobtrusive research should not be viewed as a substitute for the reactive methods such as questionnaires and interviews but instead additional measures to collect. This one method can easily be used to eliminate the biases that are presented by the reactive measures. In social research, the method defines the aspects qualitatively though at some times it can be employed quantitatively (Esterberg, 2002). Unobtrusive research method is mostly preferred because; it has few unethical considerations, it is also less costly. The method exists in three different types that are; content analysis, analysis of existing statistics, case studies. Content analysis is the analysis that involves the analysis of documents in a text. Such analysis can be either qualitative or quantitative in nature with an objective of identifying the patterns through the text. Content analysis a wide area of research; it entails thematic analysis, indexing and qualitative analysis. An example of this type of research is the analysis of photos, films and paintings. Analysis of existing statistics will entail the researcher going through valid and viable previously collected and analyzed data, which may be from government or relevant agencies. Analysis of the existing statistics is used as the main source of data for making general assumptions of individual elements to generalize on the population (Esterberg, 2002). Case studies otherwise called the analysis of comparative or historical research involve the examination of the society over a given period of time and comparing them with other societies. An example is the research that saw the transformation of society from savagery through barbarism to civilization. US anthropologist, Lewis Morgan, conducted this research. Pre-experimental research design vs. classical experimental research design A research design entails the arrangement of the layout for collection and analysis of data in ways that are relevant in the process. Pre-experimental research design is a form of design that lacks the control of its internal validity. This form of design is the easiest to form in most of the cases. The major reason being it is non-random in providing its assignments. It does not also provide the framework of controls over threats to internal validity. This design is easy to implement but due to the absence of comparison, the findings may not reflect the actual findings. Participants in this design are placed for study skills and counseling (Lawson, 2010). Classical experimental designs on the other hand do focus on the experimental group of focus and alongside this; they establish a control group for the specified research in question. The independent variable for the case will be attached to the experimental component and not the control aspect. Both of the groups are compared to the same dependent variable though. The support behind such design is that all validly true experiments must have valuable components like control, sampling through randomization and results are subject to manipulation to fit the preferred outcomes (Lawson, 2010). The classical experimental design consists of comparable groups, which are two, the experimental and the control group. The two groups are equivalent and at the same level. This process operates in three ways that is; proof of co variation, which entails presentation of evidence for correlation. The relationship of the two variables must only be confined between the two no third party interference. Lastly, the occurrences should only occur within the specified timeline (Lawson, 2010). The design has four components; - comparison which results into proof of covariance; manipulation results into creation of time order; thirdly, controls which helps in the determination of the covariance as no spurious and lastly, establishment of the timing for which the events would occur (Lawson, 2010). Four threats of internal validity 1) Maturation- this is the process by which the participants change over time which may affect the outcome of the research. 2) History- during the study a series of event, which were not planned for, can destabilize the research process. These are an event, which are outside the study but influences the outcomes. 3) Testing- this only have an effect when it takes place through a pretest, this act has the effect of affecting posttest in the future i.e. if practice test is used to form exams. 4) Instrumentation- the function of instruments is normally too accurately and consistently measure constructs. Other instruments may lack support for validity and reliability hence produce erratic results hence a threat to internal validity. For example if rating systems are used, errors are likely to occur. (Neuman, 2009). References Esterberg, K. G. (2002). Qualitative methods in social research. Boston: McGraw-Hill. Lawson, J. (2010). Design and analysis of experiments with SAS. Boca Raton: Chapman & Hall/CRC. Neuman, W. L. (2009). Understanding research. Boston, MA: Pearson/Allyn and Bacon. Ross, S. M., & PekoI?z, E. A. (2007). A second course in probability. Boston: ProbabilityBookstore.com. Read More