Key Takeaways

  1. People who applied online tended to be better hires, as measured by attrition and sales
  2. The Sales department has the highest percentage of disengaged employees
    • Their employees also take the fewest average vacation days
  3. Pay between new hires and current employees is not equitable when accounting for gender
  4. Women are less likely to be rated high performers than men

Executive Summary

This exercise illustrates my ability to analyze raw data to address specific business questions related to people analytics, diversity, equity, and inclusion. I answer critical human resource (HR) questions, provide insights into recruitment and equity issues, and make recommendations using fabricated HR data from six relational datasets. I manipulate, merge, visualize, and perform statistical tests on the HR data to answer business questions. The questions driving this analysis are:

  1. Which recruiting source yields the best hires?
  2. What is driving low employee engagement?
  3. Are new employees earning more than current employees?
  4. Are performance ratings being given fairly?

The fabricated HR data is helpful to answer the questions above. However, the data is limited in terms of its demographic data. For example, this data does not contain age, race/ethnicity, ability, or education information. Furthermore, I cannot answer other crucial questions focused on legal and regulatory issues or effective diversity practices. Nonetheless, my first analysis with these data is to examine recruitment efforts. My data analysis suggests that people who applied online produced the best hires as measured by attrition and sales. Conversely, the search firm produced the worst hires as measured by attrition and sales.

Gallup defines engaged employees as involved in, enthusiastic about, and committed to their work and workplace. Gallup categorizes workers as “engaged” based on their responses to key workplace elements it has found predict important organizational performance outcomes. Employees in the sales department were more likely to report disengaging from work. Through careful analysis of the data, I discovered that employees in the sales department took significantly fewer vacation days, on average, than the rest of the company. This isn’t to suggest not going on vacation causes disengagement; however, it may be beneficial to encourage sales employees to use their vacation days.

According to labor and employment attorney Karen Denney, pay equity means compensating employees the same when they perform similar job duties while accounting for other factors, such as their experience level, job performance, and tenure. As a result, it is essential to disaggregate data based on race, gender, age, and other social axes of difference to determine whether an organization is paying its employees equitably. This analysis showed pay between new hires and current employees is not equitable when accounting for gender. Hiring managers and other key stakeholders should implement policies and practices to address this issue. Furthermore, my analysis revealed there are other existing inequitable issues regarding gender. Women are less likely to be rated high performers than men; this may exacerbate pay gaps since performance ratings can affect bonuses and promotions. Furthermore, women are less likely to have a salaried or managerial position than men. An initiative should be undertaken to ensure performance ratings and leadership roles are given fairly.

Based on this analysis, the executive team and directors should consider discontinuing to pay a search firm to help with recruitment. Costs can be reduced in this area since many of the “best” hires find the company online. The organization should reallocate the money saved from not using a search firm to compensate women fairly. Before new hires are onboarded, the hiring managers must review their compensation package to ensure new hires are getting paid equitably and are not underemployed. Lastly, current employees must take advantage of their vacation days. Low engagement appears to be driven by departments whose employees take fewer vacation days. All work and no play make the employee disengaged.

References:

GalluP

Pay Equity

Which recruiting source yields the best hires?

Table 1: Average Attrition Percentage for Sales Team by Recruiting Source
Recruiting Source Attrition Percentage
Applied Online 24.6%
Campus 28.6%
Referral 33.3%
Search Firm 50.0%

Table 2: Average Sales by Recruiting Source
Recruiting Source Average Sales
Search Firm 0.887
Campus 0.908
Referral 1.023
Applied Online 1.059

What is driving low employee engagement?

Which department has the lowest engagement?

Table 3: Employee Engagement by Department
Department Min Median Mean Max
Finance 1 3 3.24 5
Engineering 1 3 3.15 5
Sales 1 3 2.81 5
Note:
1= low engagement, 5 = high engagement
Table 4: Disengagement by Department
Department Percent Disengaged Average Salary Average Vacation Days
Engineering 20.60% $73,576.35 12.21
Finance 19.05% $76,651.66 11.48
Sales 32.96% $75,073.57 9.22
Note:
Disengagement is defined as having an engagement score of 1 or 2

Are the engagement results statistically significant?

Below are the results of two different statistical tests (chi-square analysis and a t.test, respectively) to determine whether employees in the sales department are significantly more disengaged than employees not in the sales department. The results show that employees in the sales department are significantly more disengaged than employees not in the sales departments.

Are new employees earning more than current employees?

According to the results below, new employees earn more than current employees.

When employee salary is disaggregated by job-level, it reveals that new employees do not earn more than current employees

Are new hourly employees getting paid more?

According to this test, there is not a statistically significant difference in pay for new hourly employees versus current hourly employees.

Multiple Linear Regression

Here I use a multiple linear regression to test the difference in pay between new hires and current employees. The power of linear regression is that it can combine rigorous analysis to test the difference between groups with adding a filter (accounting for certain variables), and test again. By adding the additional variable directly into the regression, I get a significant result that takes additional information/variables into account.
According to the model above, new hires are not paid significantly more than current employees when accounting for job level. However, the model below reveals that when also accounting for gender, new hires are paid more than current employees.

Are performance ratings being given fairly?

Table 5: Gender distribution of employees who received performance ratings
Gender Percent (Total)
Female 51.9% (1526)
Male 48.1% (1414)
Total 100.0% (2940)
Table 6: Average Performance Rating by Gender
Gender Average Performance Rating
Female 2.75
Male 2.92
The analysis below reveals that one gender identity is more likely to be rated a high performer than another gender identity.

The analysis below shows a statistical difference in job-level distributions between men and women. Men are more likely to be salaried and be managers than women. Women are more likely to be hourly employees.

Logistic Regression

The logistic regression reveals that men are more likely to be rated high performers than women.
The logistic regression reveals that men are not more likely to be rated high performers than women when accounting for job-level.