IBM HR Analytics - Employee Attrition & Performance

Thank you to Kaggle user pavansubhash for contributing this dataset to experiment with. This is presented as just a topic exploration and skill-building, so all code is visible - this is not in a condensed "Final Presentation" format nor is it fully exhaustive.

The Data in Question

This fictional dataset, created by IBM data scientists, represents employee data from one business including numerous data points regarding the employees and. The goal is to analyze this data and gain insight into patterns and factors that affect employee performance within and attrition from the business.

The data file includes:

See Appendix for information on codified data.

This is not an exhuastive list and I will pare down the available data to make this project more concise and readable.

Age - The employee's age at time of data collection. - integer

Attrition - Whether this employee left or stayed employed - string, "Yes" or "No"

BusinessTravel - A category representing how often an employee travels for work - string

Department - The company's internal department that an employee works in - string

DistanceFromHome - The employee's daily commute one-way, in miles- integer

Education - A codified representation of an employee's highest attained education - int

EducationField - The employee's general studied field - string

EmployeeNumber - A unique identifier for the employee - int

EnvironmentSatisfaction- A codified representation of the employee's satisfaction with their work environment from an assessment - int

Gender - The employee's gender - string

HourlyRate - The employee's hourly wage in dollars - int

JobLevel - A codified representation of the employee's rank in the organization - int

JobRole - The employee's title - string

JobSatisfaction - A codified representation of the employee's satisfaction from an assessment - int

MaritalStatus - The employee's marital status - string

MonthlyIncome - The employee's monthly income - int

NumCompaniesWorked - The number of companies the employee has worked at previously - int

Over18 - A flag showing the employee's age over 18 years - string, "Yes" or "No"

OverTime - A flag showing the employee's eligibility for overtime pay - string, "Yes" or "No"

PerformanceRating - A codified representation of the employee's assessed work performance - int

RelationshipSatisfaction - A codified representation of the employee's satisfaction of their personal relationship - int

StockOptionLevel - Unknown - int

TotalWorkingYears - The employee's accumulated years of experience at any employer - int

TrainingTimesLastYear - The number of trainings this employee completed in the last year - int

WorkLifeBalance - A codified representation of the employee's work-life balance from an assessment - int

YearsAtCompany - The employee's accumulated years of experience at this employer - int

YearsInCurrentRole - The employee's accumulated years worked in their current role - int

YearsSinceLastPromotion - The employee's accumulated years in this particular role without a promotion - int

YearsWithCurrManager - The employee's accumulated years working with their current manager - int

Let's get started

There is a lot of data here that needs to be categorized, labeled, or re-categorized.

import os
import pandas as pd
import numpy as np
import plotly.express as px
import seaborn as sns
import seaborn.objects as so
import matplotlib as mpl
import matplotlib.pyplot as plt
pd.options.mode.chained_assignment = None

#List files in directory
print('The files in our directory are:')
for file_path in os.scandir('Data/'):
    if file_path.is_file():
        print(file_path.name)

The files in our directory are:
IBM-HR-Employee-Attrition - Workbook.xlsx
IBM-HR-Employee-Attrition.csv

#import the data
import pandas as pd
pd.options.mode.chained_assignment = None

#Specify our file names
imb_hr_csv = 'Data/IBM-HR-Employee-Attrition.csv'
#import Alfred data, and add filename column
df_ibm_hr = pd.read_csv(imb_hr_csv, low_memory=False) #Read file into dataframe

#Verify the data fields in each file
print('File Name:',imb_hr_csv)
df_ibm_hr.info()

File Name: Data/IBM-HR-Employee-Attrition.csv
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 1470 entries, 0 to 1469
Data columns (total 35 columns):
 #   Column                    Non-Null Count  Dtype 
---  ------                    --------------  ----- 
 0   Age                       1470 non-null   int64 
 1   Attrition                 1470 non-null   object
 2   BusinessTravel            1470 non-null   object
 3   DailyRate                 1470 non-null   int64 
 4   Department                1470 non-null   object
 5   DistanceFromHome          1470 non-null   int64 
 6   Education                 1470 non-null   int64 
 7   EducationField            1470 non-null   object
 8   EmployeeCount             1470 non-null   int64 
 9   EmployeeNumber            1470 non-null   int64 
 10  EnvironmentSatisfaction   1470 non-null   int64 
 11  Gender                    1470 non-null   object
 12  HourlyRate                1470 non-null   int64 
 13  JobInvolvement            1470 non-null   int64 
 14  JobLevel                  1470 non-null   int64 
 15  JobRole                   1470 non-null   object
 16  JobSatisfaction           1470 non-null   int64 
 17  MaritalStatus             1470 non-null   object
 18  MonthlyIncome             1470 non-null   int64 
 19  MonthlyRate               1470 non-null   int64 
 20  NumCompaniesWorked        1470 non-null   int64 
 21  Over18                    1470 non-null   object
 22  OverTime                  1470 non-null   object
 23  PercentSalaryHike         1470 non-null   int64 
 24  PerformanceRating         1470 non-null   int64 
 25  RelationshipSatisfaction  1470 non-null   int64 
 26  StandardHours             1470 non-null   int64 
 27  StockOptionLevel          1470 non-null   int64 
 28  TotalWorkingYears         1470 non-null   int64 
 29  TrainingTimesLastYear     1470 non-null   int64 
 30  WorkLifeBalance           1470 non-null   int64 
 31  YearsAtCompany            1470 non-null   int64 
 32  YearsInCurrentRole        1470 non-null   int64 
 33  YearsSinceLastPromotion   1470 non-null   int64 
 34  YearsWithCurrManager      1470 non-null   int64 
dtypes: int64(26), object(9)
memory usage: 402.1+ KB

pd.set_option('display.max_columns', None) #Display all of the columns
display(df_ibm_hr)

	Age	Attrition	BusinessTravel	DailyRate	Department	DistanceFromHome	Education	EducationField	EmployeeCount	EmployeeNumber	EnvironmentSatisfaction	Gender	HourlyRate	JobInvolvement	JobLevel	JobRole	JobSatisfaction	MaritalStatus	MonthlyIncome	MonthlyRate	NumCompaniesWorked	Over18	OverTime	PercentSalaryHike	PerformanceRating	RelationshipSatisfaction	StandardHours	StockOptionLevel	TotalWorkingYears	TrainingTimesLastYear	WorkLifeBalance	YearsAtCompany	YearsInCurrentRole	YearsSinceLastPromotion	YearsWithCurrManager
0	41	Yes	Travel_Rarely	1102	Sales	1	2	Life Sciences	1	1	2	Female	94	3	2	Sales Executive	4	Single	5993	19479	8	Y	Yes	11	3	1	80	0	8	0	1	6	4	0	5
1	49	No	Travel_Frequently	279	Research & Development	8	1	Life Sciences	1	2	3	Male	61	2	2	Research Scientist	2	Married	5130	24907	1	Y	No	23	4	4	80	1	10	3	3	10	7	1	7
2	37	Yes	Travel_Rarely	1373	Research & Development	2	2	Other	1	4	4	Male	92	2	1	Laboratory Technician	3	Single	2090	2396	6	Y	Yes	15	3	2	80	0	7	3	3	0	0	0	0
3	33	No	Travel_Frequently	1392	Research & Development	3	4	Life Sciences	1	5	4	Female	56	3	1	Research Scientist	3	Married	2909	23159	1	Y	Yes	11	3	3	80	0	8	3	3	8	7	3	0
4	27	No	Travel_Rarely	591	Research & Development	2	1	Medical	1	7	1	Male	40	3	1	Laboratory Technician	2	Married	3468	16632	9	Y	No	12	3	4	80	1	6	3	3	2	2	2	2
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
1465	36	No	Travel_Frequently	884	Research & Development	23	2	Medical	1	2061	3	Male	41	4	2	Laboratory Technician	4	Married	2571	12290	4	Y	No	17	3	3	80	1	17	3	3	5	2	0	3
1466	39	No	Travel_Rarely	613	Research & Development	6	1	Medical	1	2062	4	Male	42	2	3	Healthcare Representative	1	Married	9991	21457	4	Y	No	15	3	1	80	1	9	5	3	7	7	1	7
1467	27	No	Travel_Rarely	155	Research & Development	4	3	Life Sciences	1	2064	2	Male	87	4	2	Manufacturing Director	2	Married	6142	5174	1	Y	Yes	20	4	2	80	1	6	0	3	6	2	0	3
1468	49	No	Travel_Frequently	1023	Sales	2	3	Medical	1	2065	4	Male	63	2	2	Sales Executive	2	Married	5390	13243	2	Y	No	14	3	4	80	0	17	3	2	9	6	0	8
1469	34	No	Travel_Rarely	628	Research & Development	8	3	Medical	1	2068	2	Male	82	4	2	Laboratory Technician	3	Married	4404	10228	2	Y	No	12	3	1	80	0	6	3	4	4	3	1	2

1470 rows × 35 columns

What data do we need to wrangle?

Data represented in strings is very useful for us as humans to draw charts and understand easily, but if we want to do scientific analysis on most of this data, we need to transform the string fields to integers that represent those strings. For instance, Attrition as a "Yes" or a "No" is much easier to discern when reading a chart, but a computer will not understand it to analyze it in the way we want to. We can transform a "No" to a 0 and a "Yes" to a 1, then we can run algorithms like regressions on those easily.

We can also categorize Attrition the opposite way, "Yes" being 0 and "No" being 1. Whichever perspective we take we just have to ensure that we understand that while we're analyzing the data. In this case we will stick with 1 meaning "Yes".

We will need to categorize the following variables:

Attrition

BusinessTravel

Department

EducationField

JobRole

MaritalStatus

OverTime

Let's understand those data fields a little better.

df_ibm_hr_attrition = df_ibm_hr.groupby(df_ibm_hr.Attrition)['Attrition'].count().reset_index(name='Rows') #Create a new frame, counting rows grouped by year 
display(df_ibm_hr_attrition)

df_ibm_hr_businesstravel = df_ibm_hr.groupby(df_ibm_hr.BusinessTravel)['BusinessTravel'].count().reset_index(name='Rows') #Create a new frame, counting rows grouped by year 
display(df_ibm_hr_businesstravel)

df_ibm_hr_department = df_ibm_hr.groupby(df_ibm_hr.Department)['Department'].count().reset_index(name='Rows') #Create a new frame, counting rows grouped by year 
display(df_ibm_hr_department)

df_ibm_hr_educationfield = df_ibm_hr.groupby(df_ibm_hr.EducationField)['EducationField'].count().reset_index(name='Rows') #Create a new frame, counting rows grouped by year 
display(df_ibm_hr_educationfield)

df_ibm_hr_jobrole = df_ibm_hr.groupby(df_ibm_hr.JobRole)['JobRole'].count().reset_index(name='Rows') #Create a new frame, counting rows grouped by year 
display(df_ibm_hr_jobrole)

df_ibm_hr_maritalstatus = df_ibm_hr.groupby(df_ibm_hr.MaritalStatus)['MaritalStatus'].count().reset_index(name='Rows') #Create a new frame, counting rows grouped by year 
display(df_ibm_hr_maritalstatus)

df_ibm_hr_overtime = df_ibm_hr.groupby(df_ibm_hr.OverTime)['OverTime'].count().reset_index(name='Rows') #Create a new frame, counting rows grouped by year 
display(df_ibm_hr_overtime)

	Attrition	Rows
0	No	1233
1	Yes	237

	BusinessTravel	Rows
0	Non-Travel	150
1	Travel_Frequently	277
2	Travel_Rarely	1043

	Department	Rows
0	Human Resources	63
1	Research & Development	961
2	Sales	446

	EducationField	Rows
0	Human Resources	27
1	Life Sciences	606
2	Marketing	159
3	Medical	464
4	Other	82
5	Technical Degree	132

	JobRole	Rows
0	Healthcare Representative	131
1	Human Resources	52
2	Laboratory Technician	259
3	Manager	102
4	Manufacturing Director	145
5	Research Director	80
6	Research Scientist	292
7	Sales Executive	326
8	Sales Representative	83

	MaritalStatus	Rows
0	Divorced	327
1	Married	673
2	Single	470

	OverTime	Rows
0	No	1054
1	Yes	416

I think we can leave out some of those variables for now.

Let's get rid of those variables with multiple. The only exception is BusinessTravel as that is easily categorized in ascending order with Non-Travel, Travel-Rarely, and Travel-Frequently.

#Make a new dataframe with only useful data
df_ibm_hr_filtered = df_ibm_hr[[
                             'Age'
                            ,'Attrition'
                            ,'BusinessTravel'
                            #,'Department'
                            ,'DistanceFromHome'
                            ,'Education'
                            ,'EnvironmentSatisfaction'
                            ,'Gender'
                            ,'HourlyRate'
                            ,'JobLevel'
                            ,'JobSatisfaction'
                            #,'MaritalStatus'
                            ,'MonthlyIncome'
                            ,'NumCompaniesWorked'
                            ,'OverTime'
                            ,'PerformanceRating'
                            ,'RelationshipSatisfaction'
                            ,'StockOptionLevel'
                            ,'TotalWorkingYears'
                            ,'TrainingTimesLastYear'
                            ,'WorkLifeBalance'
                            ,'YearsAtCompany'
                            ,'YearsInCurrentRole'
                            ,'YearsSinceLastPromotion'
                            ,'YearsWithCurrManager'
                         ]]

df_ibm_hr_cat = df_ibm_hr_filtered
#Replace string data with integer data
df_ibm_hr_cat['Attrition'] = np.where(df_ibm_hr_cat['Attrition'] == 'Yes', 1, 0) #Categorize Attrition as "Yes" = 1, "No = 0
df_ibm_hr_cat['Gender'] = np.where(df_ibm_hr_cat['Gender'] == 'Female', 0, 1) #Categorize Gender alphabetically as Female = 0, Male = 1
df_ibm_hr_cat['OverTime'] = np.where(df_ibm_hr_cat['OverTime'] == 'Yes', 1, 0) #Categorize OverTime as "Yes" = 1, "No" = 0
df_ibm_hr_cat['BusinessTravel'] = df_ibm_hr_cat['BusinessTravel'].map({'Non-Travel': 0, 'Travel_Rarely': 1, 'Travel_Frequently': 2}) 
#df_ibm_hr_cat['BusinessTravel'] = np.where(df_ibm_hr_cat['BusinessTravel'] == 'Non-Travel', 0) #Categorize BusinessTravel
#df_ibm_hr_cat['BusinessTravel'] = np.where(df_ibm_hr_cat['BusinessTravel'] == 'Travel_Rarely', 1) #Categorize BusinessTravel
#df_ibm_hr_cat['BusinessTravel'] = np.where(df_ibm_hr_cat['BusinessTravel'] == 'Travel_Frequently', 2) #Categorize BusinessTravel

print("Our original dataset contains:")
display(df_ibm_hr_filtered)
print('\n')
print("And our codified dataset contains:")
display(df_ibm_hr_cat)

Our original dataset contains:

	Age	Attrition	BusinessTravel	DistanceFromHome	Education	EnvironmentSatisfaction	Gender	HourlyRate	JobLevel	JobSatisfaction	MonthlyIncome	NumCompaniesWorked	OverTime	PerformanceRating	RelationshipSatisfaction	StockOptionLevel	TotalWorkingYears	TrainingTimesLastYear	WorkLifeBalance	YearsAtCompany	YearsInCurrentRole	YearsSinceLastPromotion	YearsWithCurrManager
0	41	1	1	1	2	2	0	94	2	4	5993	8	1	3	1	0	8	0	1	6	4	0	5
1	49	0	2	8	1	3	1	61	2	2	5130	1	0	4	4	1	10	3	3	10	7	1	7
2	37	1	1	2	2	4	1	92	1	3	2090	6	1	3	2	0	7	3	3	0	0	0	0
3	33	0	2	3	4	4	0	56	1	3	2909	1	1	3	3	0	8	3	3	8	7	3	0
4	27	0	1	2	1	1	1	40	1	2	3468	9	0	3	4	1	6	3	3	2	2	2	2
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
1465	36	0	2	23	2	3	1	41	2	4	2571	4	0	3	3	1	17	3	3	5	2	0	3
1466	39	0	1	6	1	4	1	42	3	1	9991	4	0	3	1	1	9	5	3	7	7	1	7
1467	27	0	1	4	3	2	1	87	2	2	6142	1	1	4	2	1	6	0	3	6	2	0	3
1468	49	0	2	2	3	4	1	63	2	2	5390	2	0	3	4	0	17	3	2	9	6	0	8
1469	34	0	1	8	3	2	1	82	2	3	4404	2	0	3	1	0	6	3	4	4	3	1	2

1470 rows × 23 columns

And our codified dataset contains:

	Age	Attrition	BusinessTravel	DistanceFromHome	Education	EnvironmentSatisfaction	Gender	HourlyRate	JobLevel	JobSatisfaction	MonthlyIncome	NumCompaniesWorked	OverTime	PerformanceRating	RelationshipSatisfaction	StockOptionLevel	TotalWorkingYears	TrainingTimesLastYear	WorkLifeBalance	YearsAtCompany	YearsInCurrentRole	YearsSinceLastPromotion	YearsWithCurrManager
0	41	1	1	1	2	2	0	94	2	4	5993	8	1	3	1	0	8	0	1	6	4	0	5
1	49	0	2	8	1	3	1	61	2	2	5130	1	0	4	4	1	10	3	3	10	7	1	7
2	37	1	1	2	2	4	1	92	1	3	2090	6	1	3	2	0	7	3	3	0	0	0	0
3	33	0	2	3	4	4	0	56	1	3	2909	1	1	3	3	0	8	3	3	8	7	3	0
4	27	0	1	2	1	1	1	40	1	2	3468	9	0	3	4	1	6	3	3	2	2	2	2
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
1465	36	0	2	23	2	3	1	41	2	4	2571	4	0	3	3	1	17	3	3	5	2	0	3
1466	39	0	1	6	1	4	1	42	3	1	9991	4	0	3	1	1	9	5	3	7	7	1	7
1467	27	0	1	4	3	2	1	87	2	2	6142	1	1	4	2	1	6	0	3	6	2	0	3
1468	49	0	2	2	3	4	1	63	2	2	5390	2	0	3	4	0	17	3	2	9	6	0	8
1469	34	0	1	8	3	2	1	82	2	3	4404	2	0	3	1	0	6	3	4	4	3	1	2

1470 rows × 23 columns

Let's see if any data is correlated to another

I like to start with this step because we can gain some great insight into our data, and in turn our organization. For instance, maybe we hadn't thought about it before but Employee's with higher education levels might be correlated with one particular gender, meaning they are more likely to be one gender than the other.

What is correlation?

Correlation, in statistics, is a representation of the relationship between two things, or variables - they are represented on a scale from negative 1 to positive 1 based on the strength of the relationship in something called a "correlation coefficient". The closer the correlation coefficient is to positive or negative 1 the stronger that relationship is. The positivity or negativity depends on the context of the data itself.

Credit to Boston University School of Public Health for the following 2 images

from IPython.display import Image, display
display(Image(filename='Images/Correlation_Coefficients.png'))
display(Image(filename='Images/Correlation_Coefficient_Examples.png'))

# Compute the correlation matrix
corr = df_ibm_hr_cat.corr()

# Generate a mask for the upper triangle
mask = np.triu(np.ones_like(corr, dtype=bool))

# Set up the matplotlib figure
f, ax = plt.subplots(figsize=(20, 18))

# Generate a custom diverging colormap
cmap = sns.diverging_palette(20, 230, sep=7, s=95, as_cmap=True)

# Draw the heatmap with the mask and correct aspect ratio
heatmap_tri = sns.heatmap(corr, mask=mask, cmap=cmap, vmin=-1, vmax=1, center=0, annot=True, fmt='.3f',
            square=True, linewidths=.5, cbar_kws={"shrink": .5})

heatmap_tri.set_title('Triangle Correlation Heatmap - IBM HR Data', fontdict={'fontsize':18}, pad=16);

Does anything interesting come out here?

It sure does! Here are some items of note:

All the general correlations between age and total working years, and other representations of total working years, as well as income

Employees who are authorized to work overtime rated higher EnvironmentSatisfaction scores

Employees who worked at more companies before this rated lower JobSatisfaction scores

Employees who had more trainings last year worked at fewer companies before this

Employees who had more trainings last year also were more likely to be authorized to work overtime

Employees who had attained higher education levels worked at more companies, but also worked here longer

Let's look right at our attrition data

The moment we've all been waiting for. Understanding our employees more completely is great, but right now we are very interested in understanding what factors lead to employee attrition.

df_ibm_hr_cat.corr()[['Attrition']].sort_values(by='Attrition', ascending=False)

plt.figure(figsize=(1, 6))
heatmap_attrition = sns.heatmap(df_ibm_hr_cat.corr()[['Attrition']].sort_values(by='Attrition', ascending=False), vmin=-1, vmax=1, annot=True, fmt='.3f', cmap=cmap)
heatmap_attrition.set_title('Features Correlating with Attrition', fontdict={'fontsize':18}, pad=16);

Very interesting data here

My first note is that nothing is outwardly strong in correlation with Attrition. This might discourage some but what I have found through other research is that if we know this, and we know our attrition rates for different groups of employees, say attrition rates of men and women, that if they show a significant gap to each other that we can still find these otherwise "weak" correlations as somewhat telling. What a weak correlation tends to mean in this case is that, those attrition rates may be close, maybe within a few percentage points, but we could still want to pay attention to those factors. All we have to do is reframe what we call "weak" and "strong" correlations to our results - so if our strongest correlation is 0.30, then that is now our "strong" correlation. I have found previously that my cutoff tends to be 0.05 for a relationship worthy of noting.

My second note is that it seems odd that OverTime is positively correlated with Attrition. When I think of something positive, I usually think of something being "good". That is until I remember that our context for Attrition is that "Yes", leaving the company, is represented by 1. In this case a 1 for an employee being authorized to work overtime correlate to 1 for "Yes". So what we are seeing is that employee's who are authorized to work overtime are, at some level, leaving the company more often. This is why it important to codify our data in a way that makes sense to us when we are thinking about the data and always ensure that when we are looking at and referencing analysis results that we keep those contexts in mind.

Looking at the rest, it seems like everything makes common sense:

Employees who have to travel for work more often leave more often, to some degree

Employees working further from work leave more often, to some degree

Employees that report high satisfaction in any way stay more often

Employees who attained higher position ranks stay more often

Employees who attained a higher StockOptionLevel stay more often

Employees who have worked at this company longer, have worked longer in general, have worked with their current manager longer, etc. stay more often

Employees who have worked at more companies before this are weakly associated with leaving. We might want to look into that further, maybe combining that data with some other categories to see if there are any patterns that emerge.

Interestingly, there was no appreciable change in attrition for employees who had been rated higher or lower in terms of their performance.

Let's dig into that OverTime data a little further

We can visualize our data in another way that might make it more clear what we are talking about with correlations.

I will switch to aggregating the data and creating charts in Microsoft Excel as that is a familiar platform for me where I have done this type of research before.

display(Image(filename='Images/employee_attrition_by_overtime_auth.png'))

We can see here a large disparity in both the total number of employees who have been authorized for overtime, as well as a disparity in the proportion of employees authorized or not authorized for overtime who left the company. Let's look just at the attrition rates themselves, which would be the number of employees who left out of the total number of employees - in this case we are separating that by Overtime Authorization.

display(Image(filename='Images/employee_attrition_by_overtime_auth.png'))

Wow! It sure seems as if employees who are authorized to work overtime leave the company at a higher rate, 3 times higher, in fact, than employees who are not. This is a visual representation of our correlation coefficients - and keep in mind that the correlation factor between OverTime and Attrition was only 0.246, a correlation coefficient that is considered "weak" according to statistics literature.

So we can use correlation matrixes/heatmaps to quickly identify groups of employees that we want to investigate further, that we can see some connection between, and then dive into the data from there, using Pivot Tables to quickly manipulate the data into any many different frames of reference.

Let's dive in further, then.

We can also see a very weak correlation between OverTime and Gender so let's investigate that further.

display(Image(filename='Images/overtime_auth_by_gender.png'))

We don's see much of a difference here between the ratio of men and women who are authorized to work overtime, which is why the correlation coefficient is so weak, only -0.042. But let's see if there is a difference in attrition rates for men and women who have been authorized for overtime.

display(Image(filename='Images/attrition_rate_by_overtime_auth_and_gender.png'))

Now that is quite interesting, for employees who were not authorized for overtime there is a negligible different in attrition rates. But our men who are authorized for overtime leave at a considerably higher rate than our women who are authorized to work overtime, by 8 percentage points!

Conclusions

This came to be an interesting analysis. We found that there is a disparity in attrition rates between our company's employee's who are authorized to work overtime, and especilly a disparity in those employees based on their gender. This would warrant some further research into the experiences of men and women who are authorized to work overtime to find any reasons for that.

That is one area that this research lacks is that it only sees relationships, but not necessarily causal relationships. Someone may need to add more questions to an employee exit interview to investigate this further to determine a reason for this disparity we found.

In the future it is possible to look into every variable and every combination of variables. Of course we will strategically choose which to look further into to save time for ourselves based on some of our preliminary work along the way.

Appendix

Codified Data

Education

1 'Below College'

2 'College'

3 'Bachelor'

4 'Master'

5 'Doctor'

EnvironmentSatisfaction

1 'Low'

2 'Medium'

3 'High'

4 'Very High'

JobInvolvement

1 'Low'

2 'Medium'

3 'High'

4 'Very High'

JobSatisfaction

1 'Low'

2 'Medium'

3 'High'

4 'Very High'

PerformanceRating

1 'Low'

2 'Good'

3 'Excellent'

4 'Outstanding'

RelationshipSatisfaction

1 'Low'

2 'Medium'

3 'High'

4 'Very High'

WorkLifeBalance

1 'Bad'

2 'Good'

3 'Better'

4 'Best'