Correlational research “is a research method that gives the researcher the opportunity to describe the relationship between two measured measure variables; whether two variables are correlated” (Sherri, 2011, p.148). It should be known that the two variables under discussion cannot be manipulated. As such, the researcher computes a correlation coefficient which documents the observed relationship in numerical terms (Cresewell, 2011). This enables the researcher to make predictions from one variable to another. If a researcher learns that the two variables are related, he or she can predict one variable to the other with a high degree of accuracy. As a non experimental study, correlational research has a number of advantages. First, it is particularly significant when it is impossible to conduct experimental research due to ethical or safety issues (Cresewell, 2011). Secondly, it is useful when an investigator is interested in measuring many variables with the aim of finding if those variables are correlated (Cresewell, 2011). The degree of relationship or correlation between variables differs in magnitude. While some variables can be strongly related, others may have a moderate relationship. In addition, some variables may not be related at all.
Explanatory design and prediction design models are widely used in correlational research. According to Pandita (2012 p.1), “explanatory design examines the correlation of two or more variables while in predictive design, the capability of the prediction is the main aim of the research”. Generally, predictive design has two fundamental principles: predictor variable and the criterion variable. A predictor variable refers to a variable which the researcher uses to gauge the value of the other variable (criterion variable); criterion variable refers to the variable whose value is determined by a predictor variable. In this case, the investigation normally has two variables: predictor and criterion. Thus, the researcher uses the predictor variable to determine the value of the criterion variable.
Explanatory design has a number of characteristics. First, its main aim is to elucidate the correlation between two or more variables (Rashid, 2012). Secondly, a correlational research that employs explanatory design requires that data collection be done at one time (Rashid, 2012). Thirdly, such a design concentrates on a single group (Rashid, 2012). Fourth, this approach calls for the computation of at least two scores (Rashid, 2012). Fifth, the researcher has to determine correlational statistical test strength and direction of correlation (Rashid, 2012). Sixth, this type of design demands that the researcher has to make conclusions basing on the statistics alone (Rashid, 2012).
Explanatory design focuses on answering why two or more things are related. Social researchers argue that it is one thing to describe high unemployment rates in a given country, to examine trends over time and compare the current rates with other countries (Rashid, 2012). It is “quite a different thing however, to explain why there is a high unemployment rate in that particular country and why unemployment rates are significantly high in some countries and extremely low in others” (Rashid, 2012, p.3). Thus, the approach of the study will depend on the research question. A research question can be descriptive or explanatory in nature. A perfect example is when a researcher wants to elucidate why in country X unemployment rates are significantly high unlike in country Y where such rates are extremely low. The researcher in this case is seeking to find the relationship between high unemployment rates in country X and low unemployment rates in country Y. This will require synthesizing causal explanations.
On the other hand, “in predictive design the investigator has to categorically indicate that the goal of the investigation is to determine prediction capability of a given variable” (Pandita, 2012, p.1). Thus, this design demands that the investigator must have a predictor variable and a criterion variable. As a result, this design is significant in studies which aim at forecasting a given phenomenon. These two designs give correlational research several notable characteristics. First, correlational research is known to display scores (Rashid, 2012). This can be in the form of scatter plots or matrics. Secondly, in correlational research there are associations between scores (Rashid, 2012). These associations are described in terms of form, strength and direction. Thirdly, correlational research involves the analysis of multiple variables (Rashid, 2012). This can be in the form of multiple regression or partial correlation. Causal relationships vary and they may be simple, moderate or complex.
In correlational research, people should distinguish between causation and correlation. Rashid (2012, p. 3) argues that “they are not the same thing; however, one phenomenon follows the other and it does not mean one phenomenon causes the other”. Let us examine two examples. There is a correlation between the number of rioters and the magnitude of damage they cause. This means that if the rioters are many, the magnitude of damage will be high and vise versa. However, the number of rioters and the magnitude of damage will both be due to a third factor such as the tools used to cause destruction. In the second example, it has been observed that recently, the divorce and crime rates are very high. However, this does not imply that high crime rate is as a result of divorce. An increase in divorce and crime rates may be due to another factor. In addition, people should be able to differentiate between prediction and explanation. This means that a researcher can observe correlation. On the contrary, a researcher cannot observe cause. Thus, he or she can only infer cause basing on the correlation between the variables under investigation. Such inferences are indirectly linked to the observations made by the investigator and are thus fallible. As such, investigators should be cautious so as at to avoid incorrectly stating that a relationship is causal. Thus, when an investigator uses correlational research design he or she is must avoid making invalid inferences.
In conclusion, this paper has noted that correlational research “is a research method that gives the researcher the opportunity to describe the relationship between two measured measure variables; whether two variables are correlated” (Sherri, 2011, p.148). Explanatory design and prediction design models are widely used in correlational research. According to Pandita (2012 p.1), “explanatory design examines the correlation of two or more variables while in predictive design, the capability of the prediction is the main aim of the research”. In correlational research, people should distinguish between causation and correlation. They are not the same thing; however, one phenomenon follows the other and it does not mean one phenomenon causes the other. Inferences are indirectly linked to the observations made by the investigator and are thus fallible. Thus, when an investigator uses correlational research design he or she is must avoid making invalid inferences.
References
Cresewell, J. (2011). Educational Research: Planning, Conducting, and evaluating quantitative and qualitative research. New York: Pearson.
Pandita, R. (2012). Correlational Research. New Delhi: Web.
Rashid, A. (2012). Research methods in Education. New Delhi: Web.
Sherri, L. (2011). Research Methods and Statistics: A Critical Thinking Approach. New York: Cengage Learning.