According to (Cresewell, 2011), research surveys are grouped into two main categories namely descriptive and analytical.
Analytic Survey Research
According to Oppenheim (2000), the analytic, relational survey is set up with the aim of exploration associations between variables. Its design resembles laboratory experiments. An analytic survey tends to focus on finding associations and explanations rather than description and enumeration. Oppenheim (2000) argues that an analytic survey tends to ask ‘why’ and ‘what goes with what’ rather than ‘how many or ‘how often’. An analytical survey, according to Oppenheim (2000, p.23), “attempts to describe and explain why situations exist”.
Analytic surveys have four different types of variables and it is important for the investigator to understand these variables. Experimental variables fall in the first category. Experimental variables are also called ‘causes or predictors’. Other experts refer to them as independent or explanatory variables. According to Oppenheim (2000), the analytic type of survey is designed in a manner that allows for the variation of these factors systematically so that there effects can be observed. As such, in an analytic survey, several variables work both in isolation and in various combinations. Dependent variables fall into the second category. These variables must be measured particularly carefully and group difference tested for statistical significance. Controlled variables fall in the third category. Oppenheim (2000) suggests that as a source of variation, these should be eliminated so as to fulfill the condition of ‘other things being equal’ when the effects or correlates of experimental variables are stated. Oppenheim (2000) recommends that variables should be controlled by employing the exclusion principle. For example, by having only males in a sample, gender is excluded as a source of variation; or, by interviewing all respondents on the same day, this eliminates the variation that occurs when they are interviewed on different days or weeks. This can also be achieved through randomization. For example, by randomizing the order in which alternatives are given to the respondents, the researcher eliminates ordinal and serial effects as sources of variation.
The last category is that of uncontrolled variables. These are free floating variables and theoretically, they fall into two groups: confounded variables and error. Confounded variables, which are also known as, correlated biases, have hidden influence of unknown size, on the results. For example, in psychological research, genetic factors may influence the outcome of results. This influence is of unknown magnitude. Essentially, this means that knowledge and understanding of the phenomena under investigation are still incomplete in important ways; there are variables, other than the experimental and controlled variables, but confounded within them, that can affect the results and hence can produce serious misinterpretations. On the other hand, such uncontrolled variables are of significant help because they can lead to the development of new hypotheses, so that, eventually, their effect may be controlled (Cresewell, 2011).
Inevitably, any research design cannot be free from error. Such error variables are randomly distributed or, at any rate, distributed in such a way not to hinder the results. Generally, it is not easy to distinguish between confounded variables and pure error. In analytical surveys, as in experiments, the influence of uncontrolled variables is made as small as possible (Berger, 2010). If the presence of confounded variables is suspected, a good deal of statistical analysis may be required to uncover their identity, and ultimately their impact must be studied systematically in a new enquiry. Oppenheim (2000) argues that for the design of an analytic reach to be effective, the four types of variables should be measured as careful as possible. In deciding to measure such a variable, the investigator must define that variable in both theoretical and practical terms.
Descriptive Survey Research
Descriptive surveys are widely used and they demand high standards of precision. The main purpose of descriptive survey research is to explain what exists at the moment. Descriptive survey research are concerned with high populations. In this particular context, population denotes all those who fall in the category of interest. As such, when designing a descriptive survey research it is extremely significant to put into consideration the relationship between the sample chosen and the population under investigation. Experts argue that it is important to describe the characteristics of both the sample and the targeted population in order to have a clear understanding of their relationship (Ariola, et al., 2006).
Drawing a representative sample should be done with caution. It is advisable that a representative sample of any population should be so drawn that every member of that population has a specified non zero probability of being included in the sample (Richey, & Klein, 2007). In practice, “this usually means that every, member of the population has a statistically equal chance of being selected” (Oppenheim, 2000, p. 40). The best way of achieving this goal is by using completely random sampling method. The investigator should use appropriate methods to select a representative sample.
When choosing a survey design for a descriptive investigation, an investigator should ask himself or herself if it is possible to study the complete population or it will be necessary to draw a sample. For a complete population study, it might be problematic to define the population. It is important to have a clear demarcation of where the population begins and ends (Richey, & Klein, 2007). If this is not done in the correct manner, the investigator risks including sources of biases in his study.
In conclusion, this paper has focused on the analysis of the two major types of research surveys. First, this paper has noted that the analytic, relational survey is set up with the aim of exploration associations between variables. Its design resembles laboratory experiments. An analytic survey tends to focus on finding associations and explanations rather than description and enumeration. Oppenheim (2000) argues that an analytic survey tends to ask ‘why’ and ‘what goes with what’ rather than rather than ‘how many or ‘how often. Analytic surveys have four different types of variables and it is important for the investigator to understand these variables. For the design of an analytic reach to be effective, the four types of variables should be measured as careful as possible. In deciding to measure such a variable, the investigator must define that variable in both theoretical and practical terms.
Secondly, this paper has noted that Descriptive surveys are widely used and they demand high standards of precision. The main purpose of descriptive survey research is to explain what exists at the moment. When designing a descriptive survey research, it is extremely significant to put into consideration the relationship between the sample chosen and the population under investigation. Experts argue that it is important to describe the characteristics of both the sample and the targeted population in order to have a clear understanding of their relationship (Ariola, et al., 2006).
References
Ariola, E. et al. (2006). Principles and Methods of ResearcH. London: Oxford Press.
Berger, A. (2010). Media and Communication Research Methods: An Introduction to Qualitative and Quantitative Approaches. Sage: New York.
Cresewell, J. (2011). Educational Research: Planning, Conducting, and evaluating quantitative and qualitative research. New York: Pearson.
Oppenheim, A. (2000). Questionnaire Design, Interviewing and Attitude Measurement. London: Continuum International Publishing Group.
Richey, R., and Klein, J. (2007). Design and Development Research: Methods, Strategies, and Issues. London: Routledge.