Wage gap at workspace

1. Introduction

According to a recent Glassdoor survey, more than two-thirds (67 percent) of U.S. employees say they would not apply for jobs at employers where they believe a gender pay gap exists.Today, the gender pay gap is more than a social or legal issue. It’s an issue that can affect the ability of employers to attract and retain talent. Gender wage gap is also directly related to women poverty.

The society now aspires to bridge the gender wage gap .It is very essential for us to look at the data to recognize gender wagegsp patterns, best practices contribute for lessening the gender wage gap and practices directly contributing to increase in gender wage gap.

This analysis aims to recognize patterns regarding gender wage gap over different industries and different occupations. It tries to gain insights on the correlation of different factors like industry, occupation, payscale, age group etc. with gender wage gap and to look at correlations and identify the causal relationship. We can also compare the hiring and human resource management structures of industries with low wagegap to identify the best practices which can be applied industries with high wagegap.

2. Packages Required

To reproduce the results of this project you will need to load the following packages

library(tidyverse) # Collection of R packages designed for data science.
library(knitr) #A general-purpose tool for dynamic report generation in R
library(ggplot2) # Asthetic depiction library

3. Data Preparation

Prior to the analysis , it is essential to properly load, examine and clean the data

Loading the data

The data for this project origniated from the following source.

jobs_gender <- readr::read_csv("https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2019/2019-03-05/jobs_gender.csv")
earnings_female <- readr::read_csv("https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2019/2019-03-05/earnings_female.csv") 
employed_gender <- readr::read_csv("https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2019/2019-03-05/employed_gender.csv")

Source data has 3 different files.

1. Earnings_Female

This table has information about the salary of females in comparision to males yearwise and age group wise. #### Variable Description

Year - Year
Group - Age group
percent - Female salary percent of male salary

The data doesn’t have any missing values.

2. Employed_gender

This table has Yearwise part time and full time employeement percentage for males as well as females. The variables are .

Variable Description

year - Year
total_full_time - Percent of total employed people usually working full time
total_part_time - Percent of total employed people usually working part time
full_time_female - Percent of employed women usually working full time
part_time_female - Percent of employed women usually working part time
full_time_male - Percent of employed men usually working full time
part_time_male - Percent of employed men usually working part time

3. Jobs_Gender

This table has salary comparision amoong males and females in different occupations over a range of major and minor category of industries.

Variable Description

occupation - Specific job/career
major_category - Broad category of occupation
minor_category - character Fine category of occupation

d.total_workers - Total estimated full-time workers > 16 years old

workers_male - Estimated MALE full-time workers > 16 years old
workers_female - Estimated FEMALE full-time workers > 16 years old
percent_female - The percent of females for specific occupation
total_earnings - Total estimated median earnings for full-time workers > 16 years old
total_earnings_male - Estimated MALE median earnings for full-time workers > 16 years old
total_earnings_female - Estimated FEMALE median earnings for full-time workers > 16 years old
wage_percent_of_male - Female wages as percent of male wages - NA for occupations with small sample size

4. Data Cleaning

Variable Types

The variable types in each data file are appropriate.

Missing Values

The missing values in the data is encoded with ‘NA’.There are no missing values in the Earnings_female and employed_gender files. Wage-percent _male column which represent female wage as percent of male wages has 40 % of its entries missing. Since the column is reproducible from the other columns in the data and since more than 40 % of its entries being missing, the column has been deleted.

Duplicate entries

There are no duplicate entries in any of the 3 data files.

Variable addition

Wage gap is calculated by the following formula

Wage Gap = (Average male Pay - Average Female Pay )/(Average Male Pay)


Sum_Earnings_male = Workers_male * Total_earnings_male - Total earnings by male for the occupation

 
Sum_earnings_female = Workers_female * Total_earnings_female - Total Earnings female for the occupation

Outlier Analysis

Outlier analysis is not appropriate for this data since definite range of salaries cannot be identified since there is data from variety of categories and industries.

Code

jobs_gender <- read.csv('women_in_the_workplace_data/jobs_gender.csv')
knitr::kable(head(jobs_gender))
jobs_gender <- read.csv('women_in_the_workplace_data/jobs_gender.csv')
# checking variable types
str(jobs_gender) # variable types need not be changed

unique(jobs_gender$occupation) # All are unique
dim(jobs_gender) # dimensions of the dataframe
# Checking missing values
colSums(is.na(jobs_gender))/2088 #  remove wage_percent_of_male since there are many missing values
?complete.cases
jobs_gender1 <- jobs_gender[!duplicated(jobs_gender, nmax = 1), ] # Removing duplicate values if there are any
# REmoving redundant columns
jobs_gender2 <- jobs_gender[, 1:11] # remove wage_percent_male column since there are lot of missing values and if necessary can be calculated from other columns
head(jobs_gender2)
#Variable addition

jobs_gender2$wage_gap <- (jobs_gender2$total_earnings_male - jobs_gender2$total_earnings_female)/jobs_gender2$total_earnings_male*100 #adding wage gap column

head(jobs_gender2)
colSums(is.na(jobs_gender2))
jobs_gender3 <- jobs_gender2[complete.cases(jobs_gender2), ] # remove those values since wage gap cant be calculated for those values
colSums(is.na(jobs_gender3)) 

jobs_gender3$sum_earnings_male <-  round(jobs_gender3$workers_male)*jobs_gender3$total_earnings_male #calculated sum of male earnings and sum of female earnings  and added the column so that wage gap by major cateogry can be calculated
jobs_gender3$sum_earnings_female <- round(jobs_gender3$workers_female)*jobs_gender3$total_earnings_female

#checking the data after the changes
unique(jobs_gender3$year)
unique(jobs_gender3$minor_category)
unique(jobs_gender3$occupation)
head(jobs_gender3)
jobs_bymajor <- filter(jobs_gender3, major_category == "Management, Business, and Financial")
typeof(jobs_bymajor)
head(jobs_bymajor)
jobs_bymajor <-  data.frame(jobs_bymajor)
typeof(jobs_bymajor)
head(jobs_gender3)
str(jobs_gender3)
unique(jobs_gender3$major_category)

5. Final data sets

Female_Earnings

Year	group	percent
1979	Total, 16 years and older	62.3
1980	Total, 16 years and older	64.2
1981	Total, 16 years and older	64.4
1982	Total, 16 years and older	65.7
1983	Total, 16 years and older	66.5
1984	Total, 16 years and older	67.6

Variable Description :

Year - Year
Group - Age group
percent - Female salary percent of male salary

The data doesn’t have any missing values.

Employed_Gender

year	total_full_time	total_part_time	full_time_female	part_time_female	full_time_male	part_time_male
1968	86.0	14.0	75.1	24.9	92.2	7.8
1969	85.5	14.5	74.9	25.1	91.8	8.2
1970	84.8	15.2	73.9	26.1	91.5	8.5
1971	84.4	15.6	73.2	26.8	91.2	8.8
1972	84.3	15.7	73.1	26.9	91.1	8.9
1973	84.4	15.6	73.2	26.8	91.4	8.6

Variable Description

year - Year
total_full_time - Percent of total employed people usually working full time
total_part_time - Percent of total employed people usually working part time
full_time_female - Percent of employed women usually working full time
part_time_female Percent of employed women usually working part time
full_time_male - Percent of employed men usually working full time
part_time_male - Percent of employed men usually working part time

Jobs_Gender

X	year	occupation	major_category	minor_category	total_workers	workers_male	workers_female	percent_female	total_earnings	total_earnings_male	total_earnings_female	wage_gap	sum_earnings_male	sum_earnings_female
1	2013	Chief executives	Management, Business, and Financial	Management	1024259	782400	241859	23.6	120254	126142	95921	23.957920	98693500800	23199357139
2	2013	General and operations managers	Management, Business, and Financial	Management	977284	681627	295657	30.3	73557	81041	60759	25.026838	55239733707	17963823663
3	2013	Legislators	Management, Business, and Financial	Management	14815	8375	6440	43.5	67155	71530	65325	8.674682	599063750	420693000
4	2013	Advertising and promotions managers	Management, Business, and Financial	Management	43015	17775	25240	58.7	61371	75190	55860	25.708206	1336502250	1409906400
5	2013	Marketing and sales managers	Management, Business, and Financial	Management	754514	440078	314436	41.7	78455	91998	65040	29.302811	40486295844	20450917440
6	2013	Public relations and fundraising managers	Management, Business, and Financial	Management	44198	16141	28057	63.5	74114	90071	66052	26.666741	1453836011	1853220964

Variable Description

This table has salary comparision amoong males and females in different occupations over a ranje of major and minor category of industries.

occupation - Specific job/career
major_category - Broad category of occupation
minor_category - character Fine category of occupation

d.total_workers - Total estimated full-time workers > 16 years old

workers_male - Estimated MALE full-time workers > 16 years old
workers_female - Estimated FEMALE full-time workers > 16 years old
percent_female - The percent of females for specific occupation
total_earnings - Total estimated median earnings for full-time workers > 16 years old
total_earnings_male - Estimated MALE median earnings for full-time workers > 16 years old
total_earnings_female - Estimated FEMALE median earnings for full-time workers > 16 years old
wage_percent_of_male - Female wages as percent of male wages - NA for occupations with small sample size

6. Exploratory Data Analysis

New Variables Created

Total earnings by male for the occupation

                Sum_Earnings_male =  Workers_male * Total_earnings_male

Total Earnings by female for the occupation

                Sum_earnings_female  =  Workers_female * Total_earnings_female

The above variables are added to carry on exploratory data analysis.

Analysis

Wage Gap Decreasing over the years

The graph indicates that there is a steady increase in the female salary percent of male salary, which means the wagegap has been on a decline.

There is a visible trend of increasing wagegap in some of the age groups from 1990 to 2000.Such a sudden trend should have a specific combination socio political reasons.There is value in investigating this trend

Can Existence of wagegap attributed to higher part time female work force?

From the data, we observe that part time female work force is higher than part time male work force. one can argue that wagegap is more because of higher part time workforce in female, but the percent of female part time work force has been almost non increasing over the years, even though the wage gap has been on the decline.

Regression Analysis(ANOVA) to show Wagegap is not dependent upon the percent of part time female work force

## Analysis of Variance Table
## 
## Model 1: female_earn5$`Total, 16 years and older` ~ female_earn5$Year
## Model 2: female_earn5$`Total, 16 years and older` ~ female_earn5$Year + 
##     (female_earn5$part_time_female)
##   Res.Df    RSS Df Sum of Sq      F Pr(>F)
## 1     31 74.887                           
## 2     30 72.379  1    2.5079 1.0395 0.3161

The Low F statistic confirms that gender wagegap is not due to the higher percentage of part time female work force.

Comparision of gender wagegap over major categories of businesses

major_category	avg_earnings_male	avg_earnings_female	total_wage_gap
Computer, Engineering, and Science	81046.05	67619.32	16.56680
Education, Legal, Community Service, Arts, and Media	65979.61	47696.79	27.70980
Healthcare Practitioners and Technical	112765.64	62566.19	44.51662
Management, Business, and Financial	81058.16	59784.28	26.24521
Natural Resources, Construction, and Maintenance	42087.10	33805.00	19.67848
Production, Transportation, and Material Moving	40154.60	28335.56	29.43384
Sales and Office	48149.99	35162.98	26.97199
Service	34961.41	24806.43	29.04626

This analysis clearly shows that the gender wagegap is disproportionately higher in “Healthcare Practitioners and Technical” major category. Since there are no minor categories in the data, further analysis on this major category is not possible with the given data.For example, Factors such as greater part of female workforce working in less skillful jobs compared the male workforce can contribute to the higher gender wagegap.

Gender wagegap with income group

earnings_range	avg_earnings_male	avg_earnings_female	total_wage_gap
Greater than 150000	210107.86	149503.92	28.84420
Between 50000 and 150000	76777.87	60835.89	20.76377
Less than 50000	36725.40	31346.12	14.64730

From the above table we can infer that gender wagegap is higher for people in the higher income band, but (from the regression analysis between wage_gap and avg income) wagegap doesn’t necessarily depend on the income of the occupation.

Wage Gap within major categories vs Income range

major_category	earnings_range	avg_earnings_male	avg_earnings_female	total_wage_gap
Computer, Engineering, and Science	Between 50000 and 150000	81721.19	68882.02	15.71094
Computer, Engineering, and Science	Less than 50000	47114.04	40979.37	13.02088
Education, Legal, Community Service, Arts, and Media	Between 50000 and 150000	79274.72	57593.45	27.34953
Education, Legal, Community Service, Arts, and Media	Less than 50000	46891.33	40341.33	13.96848
Healthcare Practitioners and Technical	Between 50000 and 150000	82297.56	65917.84	19.90304
Healthcare Practitioners and Technical	Greater than 150000	210107.86	149503.92	28.84420
Healthcare Practitioners and Technical	Less than 50000	43898.51	37966.68	13.51260
Management, Business, and Financial	Between 50000 and 150000	85155.77	61857.30	27.35983
Management, Business, and Financial	Less than 50000	43263.02	35802.90	17.24365
Natural Resources, Construction, and Maintenance	Between 50000 and 150000	56113.52	48607.68	13.37616
Natural Resources, Construction, and Maintenance	Less than 50000	37460.71	28955.85	22.70340
Production, Transportation, and Material Moving	Between 50000 and 150000	61056.27	44019.22	27.90385
Production, Transportation, and Material Moving	Less than 50000	36726.50	26713.49	27.26372
Sales and Office	Between 50000 and 150000	66313.38	50103.75	24.44398
Sales and Office	Less than 50000	38907.55	32775.06	15.76169
Service	Between 50000 and 150000	66170.20	54894.34	17.04069
Service	Less than 50000	29036.44	23930.70	17.58391

From the above table, it can be observed that even among major categories, the above seen global trend of higher wagegap for a higher income bucket is being observed in all the major categories except for Natural Resources , Construction and Maintenance and Service industry.

Code for Analysis

# Wage Gap Decreasing over years

female_earn <-  read.csv("women_in_the_workplace_data/earnings_female.csv")
female_earn3 <- female_earn %>% 
  filter(group == c("20-24 years","25-34 years","35-44 years","45-54 years","55-64 years","Total, 16 years and older"))

ggplot(data = female_earn3, aes(x = Year, y = percent, color = group)) +
  geom_line()+  #wagegap vs year graph perfect

  scale_y_continuous(name = "Female salary percent of male salary")+
  ggtitle("Increase of Female salary percent of male salary over the years")

female_earn <-  read.csv("women_in_the_workplace_data/earnings_female.csv")
employed_gender <- read.csv('women_in_the_workplace_data/employed_gender.csv')
female_earn2 <- female_earn %>% spread(group,percent)

 # Regression Analysis(ANOVA) to show Wagegap is not dependent upon the percent of part time female work force
female_earn5 <- female_earn2 %>% left_join(employed_gender, by = c("Year" = "year"))

lmearn8  = lm( female_earn5$`Total, 16 years and older`~female_earn5$Year + (female_earn5$part_time_female))

lmearn9  = lm( female_earn5$`Total, 16 years and older`~female_earn5$Year)

ano = anova(lmearn9,lmearn8)
ano

# Gender wagegap with income group


jobs_gender3 <- read.csv('jobs_gender3.csv')

jobs_gender4 <- jobs_gender3 %>% mutate(earnings_range = case_when(total_earnings >= 150000 ~ "Greater than 150000",
                                                                   #total_earnings >= 100000 & total_earnings < 150000 ~  ,
                                                                   total_earnings >=50000 & total_earnings < 150000 ~ "Between 50000 and 150000",
                                                                   #total_earnings >=30000 & total_earnings < 50000 ~4,
                                                                   TRUE ~ "Less than 50000"))

table2 <- jobs_gender4 %>%
 group_by(earnings_range) %>%
 dplyr::summarise( avg_earnings_male = sum(sum_earnings_male)/sum(workers_male),avg_earnings_female = sum(sum_earnings_female)/sum(workers_female), total_wage_gap =( avg_earnings_male- avg_earnings_female)/avg_earnings_male*100 ) %>%
 arrange(desc(total_wage_gap))

dplyr:: kable(table2)

# Wage Gap within major categories vs Income range

table1 <- jobs_gender4 %>% 
  group_by(major_category , earnings_range) %>% 
  dplyr::summarise( avg_earnings_male = sum(sum_earnings_male)/sum(workers_male),avg_earnings_female = sum(sum_earnings_female)/sum(workers_female), total_wage_gap =( avg_earnings_male- avg_earnings_female)/avg_earnings_male*100 ) # %>% 
 # arrange(desc(total_wage_gap))
dplyr:: kable(table1)
#

Future Scope

The analysis goals listed below requires additional data :

Gender wage gap with respect to Education qualification
Gender Wage gap w.r.t Race of the individual
Gender wage gap w.r.t region of the individual
Comparision of Hiring and HR practices Low wagegap and high wage gap industries

7. Summary and Recommendations

The Analysis indicates that there is a steady increase in the female salary percent of male salary, which means the wagegap has been on a decline. This gives us a positive hope the coroporate society is aware of the problem and is self correcting . Now, we have to find ways to fasten the decling trend of gender wage gap.

There is a visible trend of increasing wagegap in some of the age groups from 1990 to 2000.Such a sudden trend should have a specific combination of socio political reasons.There is value in investigating this trend by investigating major government and corporate policy changes which become effective between 1990 and 2000 across the business spectrum

one can argue that wagegap is more because of higher part time workforce in female, but the percent of female part time work force has been almost non increasing over the years, even though the wage gap has been on the decline.The Low F statistic (of analysis of variance between Wagegap and part time female work force) confirms that gender wagegap is not due to the higher percentage of part time female work force.

The anlysis pointed to a definite trend of wagegap being higher for a higher income bucket in almost all major categories, but the regression analysis (between wagegap and income) didn’t yield any significant relation, instead it solely depends on the income bucket that in which the occupation exists. This implies that gender wage gap is higher in occupations with higher income. This trend although non intuitive, is identified through data analysis.

8. References

https://www.glassdoor.com/research/app/uploads/sites/2/2019/02/Gender-Pay-Gap-2019-Research-Report.pdf

Gender Wage gap Analysis

Manoj

Wage gap at workspace

1. Introduction

2. Packages Required

3. Data Preparation

Loading the data

1. Earnings_Female

2. Employed_gender

Variable Description

3. Jobs_Gender

Variable Description

4. Data Cleaning

Variable Types

Missing Values

Duplicate entries

Variable addition

Outlier Analysis

Code

5. Final data sets

Female_Earnings

Variable Description :

Employed_Gender

Variable Description

Jobs_Gender

Variable Description

6. Exploratory Data Analysis

New Variables Created

Total earnings by male for the occupation

Total Earnings by female for the occupation

Analysis

Wage Gap Decreasing over the years

Can Existence of wagegap attributed to higher part time female work force?

Regression Analysis(ANOVA) to show Wagegap is not dependent upon the percent of part time female work force

Comparision of gender wagegap over major categories of businesses

Gender wagegap with income group

Wage Gap within major categories vs Income range

Code for Analysis

Future Scope

7. Summary and Recommendations

8. References