Difference between Causation and Correlation

The Difference between Causation and Correlation within the Context of DBA Doctoral Research Study Introduction For growth to occur it is important to understand the concepts of correlation and causation. Correlation can be differentiated from causation in general terms in that correlation assists in the prediction of future events since it indicates what is likely to occur. Causation on the other hand makes it possible to alter the future. Understanding the difference helps ensure that business decisions are made based on measurable variables and tangible facts.

When decisions are based on guesswork and assumption there is a high risk that success will be jeopardized (Bleske-Rechek, Morrison & Heidtke, 2015). Prior to making any decision it is fundamental to check that the decision has been made not on assumptions but on proven facts. This discourse analyses difference between correlation and causation with respect to the doctoral research context. The implications for professional practice when a researcher implies causation after using correlation analyses While correlation is necessary it is never sufficient for making a causal inference with confidence.

It is important to have an appropriate data collection method. In order to make a causal inference it is important to gather data through the control of peripheral variables and experimental means which are likely to mislead the outcome (Bleske-Rechek, Morrison & Heidtke, 2015). Following the gathering of data using this method, if it can be established that the variable that has been manipulated experimentally has some correlation with dependent variable and that the correlation is not necessary linear, then the condition are right for making causal inference.

This mean that when the gathering of data is done through experimental means and any misleading data is alleviated then the presence of a correlation implies there is causation. In order to make causal inference it is required that there is confidence from the outcomes of the ANOVA and t tests although not necessarily with outcomes of the regression or correlation techniques.

An experimental research often entails smaller experimental treatment numbers and that the data gathered from this research is evaluated conveniently with two groups ANOVA and t tests (Coogan, 2015).ANOVA and the t test are learned during experimental research studies. At times researchers mistake experimental methods with statistical techniques (Bleske-Rechek, Morrison & Heidtke, 2015). Using correlational design adds to the existing problem.

Whenever students are trained on using correlational design for the description of non-experimental data collection methods and warned against the challenges relating to inferring causality from the data, the mistake students make is confusing the technique of correlational statistics with the method of correlational data (Coogan, 2015). Using correlational technique for design description will make the entire research to become non-experimental when it is supposed to be observational. Correlation can be understood as association.

More accurately it can be understood as a measure to the depth with which 2 variables have a relationship. If for instance increase in value of a variable is associated with the increase in value of another variable then the two variables are said to have positive correlation (Bleske-Recheks, Morrison & Heidtke, 2015). For instance there is relationship between weight and height. A taller person is expected to be heavier. Where the increase in a variable is associated with decrease of another then there is negative correlation.

For instance there is a negative correlation between temperature and sea level. The higher you go the cooler it becomes. Where no relationship exists between two variables there the situation is referred to as zero correlation. For instance no relationship exists between intelligence level and quantity of coffee consumed. Correlation may be visually expressed. A scatter gram can be drawn where the figures of a given variable are plotted against the figures of another variable on the same graph.

Reason why the results of bivariate correlation analyses are considered weak in terms of internal validity The validity concept refers to the characteristics of instruments used during quantitative research for the purpose of operationally defining the importance of variables. In general sense the validity concept is associated with the general research study quality including the consistency, uniformity, the procedure and the internal validity. Internal validity can be easily defined as the confidence that can be conferred in the causes and effects of the study relationship.

Any experiment should answer the following question: is there a possibility that there is an alternative cause that can explain the results and observation? Extreme examples such as a physics experiment on heat effects on metal conductivity have high levels of internal validity (Bleske-Rechek, Morrison & Heidtke, 2015). It is possible for a researcher to alleviate any potential confusing variables and come up with strong controls for the isolation of other factors.

On the other end studying the correlation existing between income levels and the chances of smoking has a much lower internal validity level. Researchers may identify some relationship between smoking and low income but may not have certainty that either one of them influences the other. Ethnicity, parental smoking, education, profession, social status and advertising exposure are variables likely to influence smoking behavior (Coogan, 2015). These are possible causes of smoking that are hard to eliminate. Social research could be an important field of statistical data for the credulous.

How to extend or.