This assignment is a Research Design and Data Collection
Please see the library and tutoring services section under course home to use Argosy University online library resources.
Assignment Grading Criteria:
1) Identify the correct variable and their impact and control of them.
2) To describe the research design chosen and why.
3) Noted the data collection technique and adequately explained why it was chosen.
4) Explain the correlation between satisfaction and absenteeism.
5) Organization, usage and mechanics: Introduction, transition, and conclusion; grammar, spelling, and sentence structure.
6) APA elements: In text citations and references, paraphrasing, and appropriate use of quotations and other elements of style (such as tone, audience, and word choice). Thank you.
I need to submit my response on wednesday, October 9, 2013.
The response should be at least three pages length.
It is 3 pages double-spaced, excluding the reference page. Thank you.
Where my answer is? How can I do to see that?
I can't see my response yet.
Testing the relationship between employee satisfaction and absenteeism
Prior to performing an experiment it is critical to have a well designed and thought out question. It is imperative that the hypothesis being tested is clear and that the method being used addresses the question being asked. Furthermore, it is essential that the variables being used are defined, and that there is an understanding of how the collected data will be analyzed. Below are the steps and justifications for my experimental and statistical approaches for the given assignment/question.
The hypothesis, high employee satisfaction is correlated with low employee absenteeism, can be tested using a linear regression. After the linear regression is modeled, a line of best fit can be modeled and a Pearson correlation coefficient can be calculated (McDonald 2009). To perform this analysis, the variables being used must be defined. In this analysis, I would make the independent variable employee absenteeism and measure this variable as number of days missed per employee over a 1 year time period. My dependent variable would be employee satisfaction. I would measure employee satisfaction using a numeric scale of 1-10, where 1 is low satisfaction and 10 is high satisfaction (Evans and Mathur 2005). While my dependent variable is qualitative in nature, it can and will be expressed quantitatively for this analysis.
I expect that there will noise in the dataset from extraneous variables. One extraneous variable may be gender as males and females have different perceptions of satisfaction. I will account for this by performing 2 separate analyses, one for females and one for males. Another extraneous variable may be that employees have not worked long enough for the company to know what satisfaction levels can be reached. To control for this, I will only study employees that have worked for the company for one year or more. And, because absenteeism varied by a set amount of time in which a person can be absent, I will look at absenteeism over a 1 year time period.
To collect the data for this study, I would use an online computer survey. This survey would provide anonymity. It would ask only that the person completing the survey identify their gender. The scale of satisfaction with a definition of what each number equates to would be presented, and the person completing the survey could select only one number. The survey would be available for a set amount of time; so all employees have time to complete it (Evans and Mathur 2005, Van Selm and Jankowski 2006).
Assuming that I get a correlation of r = -.70, this tells me a lot about the strength and direction of the correlation between satisfaction and absenteeism. First, the negative value indicates that is an inverse relationship, such that as satisfaction increases, absenteeism decreases, or vice versa. The r = .70 can be squared, giving you an r-sq of 0.49. This means that about 50% of the variations between these two variables can be explained from this analysis. To me, this suggests that the relationship isn’t particular strong but sort of neutral in a sense. I would consider an r-sq of .08 and higher between two variables as showing a strong relationship (McDonald 2009).
When I conduct this study there are some potential problems I might encounter, but these problems can and will be minimized. I may find that my data is not normally distributed, making it difficult to meet the assumptions of a linear regression. To meet the assumptions if my data is not normally distributed, I could perform a natural log transformation. If my data are still not normally distributed, I could choose a non-parametric test such as a Spearman rank (McDonald 2009). Another issue I may run into is having too small of a sample size. To increase my sample size, I could collect data from multiple offices of the company under study, or if this is not possible, I could pool my gender data. If this occurs, I would choose to run three analyses: female, male, female and male.
The results of this study can provide credible and valuable insight into how employee satisfaction and absenteeism are related to one another. From these findings, we will have a better understanding of how employees view their workplace in relation to the amount of time they physically spend working.
Evans, J. R., & Mathur, A. (2005). The value of online surveys. Internet Research, 15(2), 195-219.
McDonald, J.H. (2009). Handbook of Biological Statistics (2nd ed.). Sparky House Publishing, Baltimore, Maryland.
Van Selm, M., & Jankowski, N. W. (2006). Conducting online surveys. Quality and Quantity, 40(3), 435-456.