Share
Explore BrainMass

Conceptual Model Building, Implementation, and Prediction

Assume you own a software company. When you bring on new employees your hiring process is 1) to give them an technical aptitude test, 2) bring them in for an interview (conducted by a team of three hiring managers), and 3) put them through a 6-week training course. At the end of the course you have the option to let the employee go with no further costs or keep them on board. You keep a record of all your employees, including their interview score (an average score from 1-5 based on three hiring manager's input from the interview), their aptitude test score, the number of training classes they missed, and their annual performance score. These records are found in the included file "employee-perf.csv". You are hiring and want to identify good candidates before having to keep them on the payroll for a full year.

1. Conceptual Model Building: Show a multiple regression model (conceptual) that may be used to predict employee performance based on hiring and training data. If you use variable names (like X, Y, Z, etc.) please be sure to state what each variable name denotes.

2. Model Implementation in R: Implement your model in R using the "employee-perf.csv" data file. Answer the following questions:
a. Other than the fact that there are not a lot of observations in the data, does your model appear to have good validity? Why or why not?
b. How much variance of an employee's annual performance can be explained by your model?

3. Prediction: Suppose you have the following candidates that have just made it through the training course and you can either retain them or let them go. Their information is as follows:

Candidates Aptitude test; Interview score; Missed training classes
Steve jobs 84; 3.55; 6
Bill Gates 37; 4.72; 1
Sergey Brin 86; 4.61; 3

a. What is the predicted annual performance score for each of these candidates? What is the 99% confidence range for each of these predicted scores?
b. Who (if anyone) would you hire? Who (if anyone) would you let go? Why?

4. Insight: By examining different aspects of the model, what (if any) cost-cutting policies or changes should you consider? Why or why not?

Attachments

Solution Preview

The full solution is in the attached document. I hope this helps!

---------------------------------
The following text is included only for searching purposes:
Assume you own a software company. When you bring on new employees your hiring process is 1) to give them an technical aptitude test, 2) bring them in for an interview (conducted by a team of three hiring managers), and 3) put them through a 6-week training course. At the end of the course you have the option to let the employee go with no further costs or keep them on board. You keep a record of all your employees, including their interview score (an average score from 1-5 based on three hiring manager?s input from the interview), their aptitude test score, the number of training classes they missed, and their annual performance score. These records are found in the included file ?employee-perf.csv?. You are hiring and want to identify good candidates before having to keep them on the payroll for a full year.

1. Conceptual Model Building: Show a multiple regression model (conceptual) that may be used to predict employee performance based on hiring and training data. If you use variable names (like X, Y, Z, etc.) please be sure to state what each variable name denotes.

The multiple regression model would be as follows:
Annual Performance = Intercept + B1*(Aptitude Test Score) ...

Solution Summary

The expert examines conceptual model building, implementation and prediction.

$2.19