The performances of employees in terms of sales made for two IT companies, in company A employees are expected to work for five hours while in company B employees work for ten hours, are compared.
ANOVA is an abbreviation of One- way analysis of Variance and is used in the comparison of means of more than two samples using the Fisher- Snedecor distribution (F distribution). ANOVA technique is also applicable in numerical data. Usually the ANOVA is used for testing the Null hypothesis that the group samples are taken from a population with the same value of the mean. For example the null hypothesis in the example for the two companies above is; ‘There is no significant relationship between the amount of employee working hours and the sales made per day’
For the results of a one way ANOVA to be considered valid three main assumptions have to be given (Girden, P8) . The first assumption is that the population variance is equal. In the above example it has to be assumed that the gender, age, religious affiliation and individual employee motivation does not have any impact on the working efficiency of the employees.
The results for a specific group are identical and independently distributed in a normal random sample. In this example after the experiment it should be assumed that the results for company ‘A’ and the correspondent result of company ‘B’ is identical and independently distributed in a normal random sample.
Lastly the response (dependent) variable should have a normal distribution or an approximately normal distribution. In the above example the sales made per day should assume a normal distribution.
An experiment can simply be described as a test or test series with making of objective changes to the input variable so as to identify and observe reasons for change in the output variable (Montgomery, p1). Experiment design and analysis in research refers to a given task design aimed at describing or explaining information variation under hypothesized conditions in order to reflect variation. Experimental design is not only involved with selection of suitable predictors and the related outcomes but also delivery planning of the experiment subject to statistically optimal conditions within the available resources. The major concerns for experimental design are establishing the validity, reliability and replicability of the specific experiment.
Methods for experimental design have been widely applied in many disciplines. Experimentation can be viewed in two ways, either as a part of scientific research or as a way for learning how systems and process work- in management. Experimental design is therefore imperative tool both in research and management.
In management experimental design proves effective in improving process yield, reducing variability and closer conformance to nominal or required target, to required targets or nominal, lessening the time for development and reduction in the overall cost expenditure(Montgomery, p8).
In research experimental designs and analysis is useful in the formulation of hypothesis. The hypothesis directs the researcher formulating the objectives of the study and also is relevant in formulating the research questions. An ad hoc analysis which refers to the hypothesis identified following results from the test and is useful in explaining reasons for contrary results.
Work Cited
Girden, Ellen R. Anova: Repeated Measures. Newbury Park, Calif. [u.a.: Sage Publ, 1999. Print.
Montgomery, Douglas C. Design and Analysis of Experiments. Hoboken, NJ: Wiley, 2008. Print.
