Project II. Human Resources Application

Management at 1st Financial Corporation has come to the conclusion that predicted sales is the key element in their recruiting strategy. They are very interested in modeling sales by their employees so that they can begin a recruiting strategy that has the potential for increasing their profits. They therefore need to focus on predictors that they will have access to prior to hiring the employee.

  1. They believe that employees who see a good future in the company will have better motivation and, hence, will result in higher sales. So they want to use the three questions about future employment possibilities (ADVNCOK, PROMFAIR and FUTOPPS) as predictors[1]. Rather than including their results separately, they believe that they should form a scale by adding the results of the three questions together to form a Guttman Scale. You will have to generate this scale (name it TOTAL) by adding answers to the three questions together. For this scale to be reliable, it must be true that all the questions that make up the scale have positive simple correlations with each other. They want you to check to be sure that this is true (i.e. these correlations are indeed positive).
  2. They have also decided to give a pre-employment exam based on their previous results of analyses of their worker satisfaction survey. This survey showed that their best workers (who produce the highest sales figures) enjoyed working with customers (ENJYCUST).They want to use a question like the ENJYCUST question. They want to compare applicants who will enjoy working with customers (ENJYCUST>4) with applicants who are less enthusiastic (ENJYCUST<5). You must therefore include a dummy variable that will allow us to compare current employees who score greater than 4 on this question with the employees who score less than 5.
  3. In addition to results for the preemployment test variables, they are also interested in the following worker characteristics:
  1. Gender of the worker
  2. Ethnic group of the worker
  3. Education level of the worker
  4. Work experience of the worker ( measured by WRKYRS) (you should allow for a quadratic relationship of sales to this variable)
  5. Age of the worker (compare workers under 25 with workers 25 and older)

They expect you to predict the sales of workers based on a model that uses all these characteristics at the same time. We call this the ``full multiple regression model'' (FULL MODEL). When you have this information, it should allow you to obtain point estimates of the various slope coefficients of all the explanatory variables.

  1. With this information, it should also allow you to determine which characteristics are significantly related to sales when you control for other characteristics. Please let them know which characteristics are important in this analysis and which characteristics turn out to be insignificant. To explore this, you should perform an overall ``full model F-test'' for the significance of the full multiple regression model. Then you need to perform an individual t-test on the significance of each individual explanatory variable and decide which of the explanatory variables in the FULL MODEL could be dropped to arrive at the REDUCED MODEL. After that, you should perform a ``partial F-test'' on the group of explanatory variable(s) that is (are) suggested by the individual t-tests to be dropped from the FULL MODEL to arrive at the REDUCED MODEL. This should allow you to decide whether you should use the FULL MODEL or REDUCED MODEL as your FINAL MODEL for prediction. You can perform all the tests using a 5% level of significance.
  2. They are also interested in the following questions:
  1. How much of the variation in sales is explained by all the characteristics in the FULL MODEL?
  2. Should they pay higher wages on average for college educated applicants in the FULL MODEL based on your result using a 5% level of significance? Suppose that they do need to pay slightly higher wages for college educated persons as a result of using a much higher level of significance in performing the individual t-test, how much higher should they be willing to go on average? Please provide both an estimate and a 30% confidence interval for this estimate.
  3. Using the FULL MODEL again and assuming that all the dummy variables for the various ethnic groups are significant as a result of a much higher level of significance used in performing the individual t-tests, how much more or less will they be willing to pay on average for an Asian employee as compared to a Caucasian employee? Also how much more or less will they be willing to pay on average for an African-American employee as compared to a Hispanic employee?
  4. How much more will they be willing to pay on average for an applicant who scores two more points in the TOTAL score using the FULL MODEL?
  5. Please estimate the expected sales for the following person: a 24 year old white male with 3 years of work experience and a college education; he reports that he definitely enjoys working with customers and he scores "10" on the TOTAL scale.

As usual, you need to include a printout of the results that you obtained in an appendix to your report. This should include the means and standard deviations of all variables that you used as well as all your regression results. These tables must be formatted to be easily read. Remember, all the rules listed under the Project Report Guidelines still apply.

Appendix

Below we provide the specific questions associated with the data your team has been provided. The variable name coincides with the label at the top of each column in your excel spread sheet.

FEELJOB. How do you feel about your job?

1. I like my job a great deal

2. I am somewhat satisfied with my job

3. I don't strongly like or dislike my job

4. I am somewhat dissatisfied with my job

5. I don't like my job at all

PLACEWRK. How would you rate 1st Financial as a place to work

1 One of the best

2 Above Average

3 Average

4 Below Average

5 One of the worst

ATTR1YR. How likely will you be working here 1 year from now?

1 Not at all likely

4 Probably

6 Very likely

DAYSMISS. How many days of work did you miss in the last month?

0 none

1 one

2 two

3 three

4 four

5 five

6 six

7 seven

8 eight or more

ADVNCOK. I feel my advancement here has progressed satisfactorily

1 Strongly disagree

2 Disagree

3 Somewhat disagree

4 Somewhat agree

5 Agree

6 Strongly agree

PROMFAIR. In general, I think the way 1st Financial promotes fairly

1 Strongly disagree

2 Disagree

3 Somewhat disagree

4 Somewhat agree

5 Agree

6 Strongly agree

FUTOPPS. My future opportunities here look good

1 Strongly disagree

2 Disagree

3 Somewhat disagree

4 Somewhat agree

5 Agree

6 Strongly agree

ENJYCUST. I enjoy working with our customers.

1 Strongly disagree

2 Disagree

3 Somewhat disagree

4 Somewhat agree

5 Agree

6 Strongly agree

ETHNIC. Ethnicity

1 African-American/Black

2 American Indian/Native Alaskan

3 Asian/Pacific Islander

4 Caucasian

5 Hispanic/Latino

GENDER. Gender

1 Female

2 Male

AGE. Age

1 Under 25 years of age

2 25 to 34 years of age

3 35 to 44 years of age

4 45 to 54 years of age

5 55 years or older

EDUCATN. Education Level

2 High school diploma/GED

4 College

SCHEDULE. Work schedule

1 Part-time

2 Full-time

WRKEXPER. How many years of work experience did you have prior to this job (classified into various categories)?

1 No experience

2 Less than 1 year

3 1-2 years

4 3-5 years

5 6-9 years

6 10 or more years

RATEPROD. Productivity Rating

1 One of the best

2 Above average

3 Average

4 Below average

5 One of the worst

WRKYRS. How many years of work experience did you have prior to this job?

TYPE. Rating of supervisors of the type of worker

1 Good

2 Average

3 Poor

SALES. Dollar sales during the past year

[1] See the Appendix to this document for a description of the variables.