In statistics interaction occurs when the relationship among three ormore variables is been considered. In most cases, interactionrelationship is assessed when simultaneous influence of two variableson a third variable is not additive. Interaction is considered duringregression analysis. The presence of interaction between variableshas great implication on the interpretation of statistical models ofanalysis. In the even an interaction occurs, relationship between twovariables, the third valuable is seen as the ‘dependent variable’which depends on the interacting variables. Interaction presentsgreat problem in predicting consequences of altering one variablethat interact if the variables involved are difficult to measure(Brambor and Clark, 2006).
The effects of smoking and inhaling asbestos fibers lead to increasedrisk of cancer among people for smokers and none smokers. However,the risks of getting cancer are higher for smokers who are exposed toasbestos fibers.
Variables that interaction
In this case, the variables are smoking, asbestos fibers, smokers andnone smokers. Smoking and inhaling asbestos fibers are risk factorsthat lead to cancer for both smokers and non smokers in the society.In this case, interaction occurs between the joint effect of inhalingasbestos and smoking (Brambor and Clark, 2006). This means thatindividuals who smoke are more vulnerable to cancer than non smokerswho inhale asbestos fibers. The interaction variables smoking andinhaling asbestos fibers cause an additive effect on the thirdvariable which is ‘smoking.’ Dependent variables such as cancerdepend on the value of interacting variables ‘asbestos fibers andsmoking.’ In this case, it may be difficult to assess individualeffect of changed variable like moderating smoking or asbestos fiber.Precise prediction of the effect the interacting variables ‘smokingand asbestos’ has on cancer risk, can be achieved if the variablesare separated and measured individually (Brambor and Clark, 2006).
Brambor, T. Clark, W. R. (2006)."UnderstandingInteraction Models: Improving Empirical Analyses".Political Analysis14 (1):63–82. Accessed at:http://pan.oxfordjournals.org/content/14/1/63.short