Reference no: EM133676816
Problem
In this project, you have the opportunity to study a question (or a set of questions) of your interest related to human capital analysis. There is no restriction on what data to use, what statistical software to run, or what statistical tests you may use to answer the question. There are only three requirements/guidelines:
1) The question you propose to study needs to be interesting and meaningful in practice, meaning that (a) there is no obvious answer to the question without any data analysis, and (b) answering the question could enhance the practice of human capital management. It also needs to be related to at least one of the topics covered in the course (e.g., diversity, engagement, turnover, performance, recruitment, etc.).
2) To answer the question, you need to tap into more than one dataset. There are various ways to do so. You could join two datasets together according to a common identifier variable. You could also use different datasets to study different aspects of the problem (e.g., each dataset for a different industry). For example, if you are interested in studying how job experience affects job performance, you could tap into one dataset, "Chapter 6 Individual Turnover", to analyze the relationship between "AppraisalRating" and "LengthOfService", and another dataset, say "Chapter 7 with performance 2014 2015 and Sick2014", for a more comprehensive view that includes not only job tenure and performance rating but also other variables such as job strain.
3) Since our course focuses on analytical methods, your answer to the question should include not only conceptual arguments but also statistical evidence. In general, I would expect at least five statistical tests/procedures in your analysis. These could be different procedures or variations of the same procedure. For example, when studying job tenure, you could treat it as a numerical variable and run linear regression with job tenure being an independent variable. You could also treat it as a categorical variable and run another linear regression after dummy coding (like we do in Week 5 to address the non-linear effect of age).