Final Project

Introduction

Research Question

Does working more or less than average affect your work life balance?

For this project I will be analyzing whether work life balance is affected by the number of hours worked. The data that I will be analyzing is the 2016 American Time Use Survey (ATUS). The purpose of this analysis is to illustrate the fact that work life balance does in fact suffer or benefit from the number of hours worked relative to our definition of average (working less or more than a 9 to 5, or 8 hours, will affect other facets of life). I believe that by illustrating this fact people may begin to more effectively allocate their time.

I will provide a variety of summary statistics that illustrate work and leisure hours across age, sex, and employment status. Based on the statistics provided I will draw my final conclusion; given that this is an observational study causality will not be examined, but rather we will examine the correlation associated with the underlying variables.

“Work life balance” will be defined as the level of leisure time experienced outside of work versus the time spent at work. For both simplicity’s sake and for a direct and effective analysis I will only examine leisure against work as it provides a very insightful amount of data into time spent.

“Average work day” will be defined as 8 hours, which is a 9 to 5 job as this is the clear association with average that people generally construct.

Data: Cases, Structure, Variables, and Collection

Sampling

Given that the survey is a random sample of the population it would be prudent to create multiple strata to analyze homogenous data points without a strong skew.

Cases

Each case represents a household in the U.S. There are 10493 cases in the primary dataset (atusresp file).

Data Collection

Data collection is sponsored by the Bureau of Labor Statistics as part of the American Time Use Survey (ATUS) and is conducted by the U.S Census Bureau. “In ATUS, individuals are randomly selected from a subset of households that have completed their eighth and final month of interviews for the Current Population Survey(CPS). ATUS respondents are interviewed only one time about how they spent their time on the previous day, where they were, and whom they were with.” - Quoted from the ATUS User Guide

Type of Study

This is an observational study. This can be discerned by the method of collection which is the use of surveys and interviews. No control groups are established indicating that this is not an experiment.

Data Source

Data is collected by the U.S Census Bureau and is available online at two sources, http://asdfree.com/american-time-use-survey-atus.html and https://www.bls.gov/tus/datafiles_2016.htm. In order to collect this data in an efficient and reproducible manner I utilized the lodown library to download the files.

(June 2018). American Time Use Survey User’s Guide. https://www.bls.gov/tus/atususersguide.pdf#page=33 (2003-2017). The American Time Use Survey (ATUS) is sponsored by the Bureau of Labor Statistics and conducted by the U.S. Census Bureau. Retrieved October 30th, 2018 from https://www.bls.gov/tus/datafiles_2016.htm

Response

The response variable is time dedicated to leisure and it is numerical.

Explanatory Variable

The explanatory variable is hours worked per day and it is numerical.

Scope of Inference

The population of interest is everyone who works greater than 8 hours and those who work less than 8 hours. Given that this is a national study the findings can be generalized to the population with the scope of the definition established in this project. If a person is seeking a broader scope of “work life balance” more variables will need to be incorporated. Skews/biases may exist in the data based on age, employment status, and gender. I will attempt to remove these in this analysis to the best of my ability.

Causality cannot be determined in this study as it is observational and not an experiment with a control group.

Parsing

## Download of ATUS files using the lodown library

atus_cat = get_catalog("atus",output_dir = file.path(path.expand( "~" ) , "ATUS" ))
atus_cat = subset(atus_cat,directory == 2016)
atus_cat = lodown("atus",atus_cat)

## Load libraries

library(lodown)
library(psych)
load("inference.RData")
load("multilines.RData")
## Data load 
## Parsing to read ATUS files 

atusact <- readRDS( file.path( path.expand( "~" ) , "ATUS" , "2016/atusact.rds" ) )
atusact <- atusact[ c( 'tucaseid' , 'tutier1code' , 'tutier2code' , 'tuactdur24' ) ]

atusresp <- readRDS( file.path( path.expand( "~" ) , "ATUS" , "2016/atusresp.rds" ) )
atusresp <- atusresp[ c( 'tucaseid' , 'telfs' , 'tulineno' ) ]

atusrost <- readRDS( file.path( path.expand( "~" ) , "ATUS" , "2016/atusrost.rds" ) )
atusrost <- atusrost[ , c( 'tucaseid' , 'tulineno' , 'teage' , 'tesex' ) ]

atuswgts <- readRDS( file.path( path.expand( "~" ) , "ATUS" , "2016/atuswgts.rds" ) )
atuswgts <- atuswgts[ , c( 1 , grep( 'finlwgt' , names( atuswgts ) ) ) ]

## Filter into two separate dfs

atusactF = atusact[atusact$tutier1code==5|atusact$tutier1code==12,]
atusact[atusact$tutier1code==5|atusact$tutier1code==12,'tutier1code']=
  atusact[atusact$tutier1code==5|atusact$tutier1code==12,'tutier2code']

agg_table = aggregate(tuactdur24~tucaseid+tutier1code,data = atusactF,sum)

wide_table = reshape(agg_table,idvar = 'tucaseid' , 
                     timevar = 'tutier1code' , direction = 'wide' )

wide_table[is.na(wide_table)] = 0
wide_table[,-1] = wide_table[,-1]/60

#####

merge_table_one_resp = merge(atusresp,wide_table)
merge_table_one_rost = merge(merge_table_one_resp,atusrost)

names(merge_table_one_rost) = gsub("\\.","_",names(merge_table_one_rost))

merge_table_one_rost$Gender = ifelse(merge_table_one_rost$tesex == 1,"male","female")
merge_table_one_rost$Employment_Status = 
  ifelse((merge_table_one_rost$telfs==1)|(merge_table_one_rost$telfs==2),"Employed",
         ifelse((merge_table_one_rost$telfs==3)|(merge_table_one_rost$telfs==4),"Unemployed","Out of Work Force"))

General Summary Statistics

describe(merge_table_one_rost$tuactdur24_5)

##    vars     n mean   sd median trimmed mad min  max range skew kurtosis
## X1    1 10315 2.67 4.04      0    1.97   0   0 22.5  22.5 1.17     0.06
##      se
## X1 0.04

describe(merge_table_one_rost$tuactdur24_12)

##    vars     n mean   sd median trimmed  mad min  max range skew kurtosis
## X1    1 10315 5.06 3.53   4.42    4.74 3.58   0 23.2  23.2 0.82     0.41
##      se
## X1 0.03

describe(merge_table_one_rost$teage)

##    vars     n  mean   sd median trimmed   mad min max range skew kurtosis
## X1    1 10315 49.56 18.1     49   49.49 20.76  15  85    70 0.04    -0.92
##      se
## X1 0.18

table(merge_table_one_rost$Gender,useNA = "ifany")

## 
## female   male 
##   5701   4614

table(merge_table_one_rost$Employment_Status,useNA = "ifany")

## 
##          Employed Out of Work Force        Unemployed 
##              6282              3699               334

prop.table(table(merge_table_one_rost$Employment_Status,useNA = "ifany"))*100

## 
##          Employed Out of Work Force        Unemployed 
##         60.901600         35.860397          3.238003

## Description of general work hours by employment status
## Tier One Code of 5 composes the general working hours according to the Lexicon

describeBy(merge_table_one_rost$tuactdur24_5,group = 
             merge_table_one_rost$Employment_Status, mat = TRUE)

##     item            group1 vars    n      mean        sd   median
## X11    1          Employed    1 6282 4.3426350 4.4145744 3.508333
## X12    2 Out of Work Force    1 3699 0.0446427 0.4892487 0.000000
## X13    3        Unemployed    1  334 0.3816866 1.2749171 0.000000
##        trimmed      mad min       max     range       skew   kurtosis
## X11 3.94284719 5.201455   0 22.500000 22.500000  0.4270567  -1.080162
## X12 0.00000000 0.000000   0  9.516667  9.516667 13.5751986 201.951041
## X13 0.04266169 0.000000   0  8.333333  8.333333  4.2446138  18.942907
##              se
## X11 0.055698035
## X12 0.008044286
## X13 0.069760358

## Description of general work hours by employment status 
## Tier One Code of 12 composes the general leisure hours according to the Lexicon

describeBy(merge_table_one_rost$tuactdur24_12,group = 
             merge_table_one_rost$Employment_Status, mat = TRUE)

##     item            group1 vars    n     mean       sd   median  trimmed
## X11    1          Employed    1 6282 3.983660 2.974093 3.333333 3.669999
## X12    2 Out of Work Force    1 3699 6.798919 3.674028 6.500000 6.609580
## X13    3        Unemployed    1  334 6.035828 3.582504 5.666667 5.798570
##         mad min      max    range      skew   kurtosis         se
## X11 2.71810   0 20.43333 20.43333 0.9906968 0.86421358 0.03752369
## X12 3.87947   0 23.20000 23.20000 0.5046794 0.03012316 0.06040882
## X13 3.95360   0 20.45000 20.45000 0.6825584 0.38503967 0.19602587

## Description of general work hours by age

plot(merge_table_one_rost$tuactdur24_5~merge_table_one_rost$teage)

## Description of general leisure hours by age

plot(merge_table_one_rost$tuactdur24_12~merge_table_one_rost$teage)

## Description of leisure against work hours

plot(merge_table_one_rost$tuactdur24_12~merge_table_one_rost$tuactdur24_5)

## Description of general work hours by gender

describeBy(merge_table_one_rost$tuactdur24_5,group = merge_table_one_rost$Gender, mat = TRUE)

##     item group1 vars    n     mean       sd median  trimmed mad min
## X11    1 female    1 5701 2.138388 3.632258      0 1.420997   0   0
## X12    2   male    1 4614 3.333788 4.412284      0 2.696167   0   0
##          max    range      skew   kurtosis         se
## X11 21.31667 21.31667 1.4202368  0.6790776 0.04810621
## X12 22.50000 22.50000 0.8937269 -0.5361100 0.06495681

## Description of general leisure hours by gender

describeBy(merge_table_one_rost$tuactdur24_12,group = merge_table_one_rost$Gender, mat = TRUE)

##     item group1 vars    n     mean       sd   median  trimmed     mad min
## X11    1 female    1 5701 4.847708 3.343635 4.250000 4.554769 3.33585   0
## X12    2   male    1 4614 5.321572 3.731478 4.583333 4.987681 3.83005   0
##          max    range      skew  kurtosis         se
## X11 22.41667 22.41667 0.8198401 0.5305863 0.04428364
## X12 23.20000 23.20000 0.7828516 0.1962094 0.05493411

## Histograms

hist(merge_table_one_rost$tuactdur24_5)

hist(merge_table_one_rost$tuactdur24_12)

## Linear Model: Age and Gender against Hours worked
age_gender_lm = lm(merge_table_one_rost$tuactdur24_5~merge_table_one_rost$teage+as.factor(merge_table_one_rost$Gender))
age_gender_lm

## 
## Call:
## lm(formula = merge_table_one_rost$tuactdur24_5 ~ merge_table_one_rost$teage + 
##     as.factor(merge_table_one_rost$Gender))
## 
## Coefficients:
##                                (Intercept)  
##                                    4.35254  
##                 merge_table_one_rost$teage  
##                                   -0.04391  
## as.factor(merge_table_one_rost$Gender)male  
##                                    1.11036

multiLines(age_gender_lm)

## Linear Model: Age and Gender against Hours of Leisure
age_gender_leisure_lm = lm(merge_table_one_rost$tuactdur24_12~merge_table_one_rost$teage+as.factor(merge_table_one_rost$Gender))
age_gender_leisure_lm

## 
## Call:
## lm(formula = merge_table_one_rost$tuactdur24_12 ~ merge_table_one_rost$teage + 
##     as.factor(merge_table_one_rost$Gender))
## 
## Coefficients:
##                                (Intercept)  
##                                    1.79800  
##                 merge_table_one_rost$teage  
##                                    0.06048  
## as.factor(merge_table_one_rost$Gender)male  
##                                    0.59100

multiLines(age_gender_leisure_lm)

Exploratory Analysis

## Df of average hours worked (assuming 9 to 5 is average hours worked)

avg_data_frame = merge_table_one_rost[merge_table_one_rost$tuactdur24_5==8,]

## Df of greater than 8 hours (assuming 9 to 5 is average hours worked)
max_data_frame = merge_table_one_rost[merge_table_one_rost$tuactdur24_5>8,]

## Df of less than 8 hours (assuming 9 to 5 is average hours worked)
min_data_frame = merge_table_one_rost[merge_table_one_rost$tuactdur24_5<8,]

## Average Hours: Summary

describe(avg_data_frame$tuactdur24_12)

##    vars   n mean   sd median trimmed  mad min  max range skew kurtosis
## X1    1 117 2.58 1.94   2.33    2.39 1.98   0 9.15  9.15 0.82     0.41
##      se
## X1 0.18

describeBy(avg_data_frame$tuactdur24_12,group = 
             avg_data_frame$Employment_Status, mat = TRUE)

##     item     group1 vars   n     mean       sd   median  trimmed     mad
## X11    1   Employed    1 115 2.584638 1.953425 2.333333 2.394444 1.97680
## X12    2 Unemployed    1   2 2.250000 1.060660 2.250000 2.250000 1.11195
##     min  max range      skew   kurtosis        se
## X11 0.0 9.15  9.15 0.8137692  0.3659598 0.1821578
## X12 1.5 3.00  1.50 0.0000000 -2.7500000 0.7500000

plot(avg_data_frame$tuactdur24_12~avg_data_frame$teage)

describeBy(avg_data_frame$tuactdur24_12,group = avg_data_frame$Gender, mat = TRUE)

##     item group1 vars  n     mean       sd   median  trimmed     mad min
## X11    1 female    1 50 2.117333 1.631962 1.916667 2.011250 2.02622   0
## X12    2   male    1 67 2.923383 2.086208 2.750000 2.723333 2.10035   0
##      max range      skew   kurtosis        se
## X11 6.00  6.00 0.3983394 -0.9248016 0.2307942
## X12 9.15  9.15 0.8273920  0.1592530 0.2548709

hist(avg_data_frame$tuactdur24_12)

age_gender_leisure_avg_lm = lm(avg_data_frame$tuactdur24_12~avg_data_frame$teage+as.factor(avg_data_frame$Gender))
age_gender_leisure_avg_lm

## 
## Call:
## lm(formula = avg_data_frame$tuactdur24_12 ~ avg_data_frame$teage + 
##     as.factor(avg_data_frame$Gender))
## 
## Coefficients:
##                          (Intercept)  
##                              1.10887  
##                 avg_data_frame$teage  
##                              0.02308  
## as.factor(avg_data_frame$Gender)male  
##                              0.82358

multiLines(age_gender_leisure_avg_lm)

## Max Hours: Summary

describe(max_data_frame$tuactdur24_5)

##    vars    n mean  sd median trimmed  mad  min  max range skew kurtosis
## X1    1 1699 9.97 1.9   9.42    9.66 1.43 8.02 22.5 14.48 1.82     4.92
##      se
## X1 0.05

describe(max_data_frame$tuactdur24_12)

##    vars    n mean   sd median trimmed  mad min   max range skew kurtosis
## X1    1 1699 2.05 1.56   1.87    1.92 1.66   0 10.42 10.42  0.8     0.79
##      se
## X1 0.04

describeBy(max_data_frame$tuactdur24_5,group = 
             max_data_frame$Employment_Status, mat = TRUE)

##     item            group1 vars    n     mean        sd   median  trimmed
## X11    1          Employed    1 1693 9.972672 1.9024081 9.416667 9.667528
## X12    2 Out of Work Force    1    5 8.666667 0.5478493 8.500000 8.666667
## X13    3        Unemployed    1    1 8.333333        NA 8.333333 8.333333
##         mad      min       max    range      skew  kurtosis         se
## X11 1.43318 8.016667 22.500000 14.48333 1.8193553  4.904075 0.04623546
## X12 0.49420 8.066667  9.516667  1.45000 0.4652411 -1.541227 0.24500567
## X13 0.00000 8.333333  8.333333  0.00000        NA        NA         NA

describeBy(max_data_frame$tuactdur24_12,group = 
             max_data_frame$Employment_Status, mat = TRUE)

##     item            group1 vars    n     mean       sd   median  trimmed
## X11    1          Employed    1 1693 2.050158 1.561476 1.866667 1.921956
## X12    2 Out of Work Force    1    5 2.863333 2.061465 3.000000 2.863333
## X13    3        Unemployed    1    1 0.000000       NA 0.000000 0.000000
##         mad min      max    range       skew   kurtosis         se
## X11 1.65557   0 10.41667 10.41667 0.80360613  0.8024708 0.03794957
## X12 1.01311   0  5.75000  5.75000 0.01085767 -1.4853717 0.92191528
## X13 0.00000   0  0.00000  0.00000         NA         NA         NA

plot(max_data_frame$tuactdur24_5~max_data_frame$teage)

plot(max_data_frame$tuactdur24_12~max_data_frame$teage)

plot(max_data_frame$tuactdur24_12~max_data_frame$tuactdur24_5)

hist(max_data_frame$tuactdur24_12)

age_gender_leisure_max_lm = lm(max_data_frame$tuactdur24_12~max_data_frame$teage+as.factor(max_data_frame$Gender))
age_gender_leisure_max_lm

## 
## Call:
## lm(formula = max_data_frame$tuactdur24_12 ~ max_data_frame$teage + 
##     as.factor(max_data_frame$Gender))
## 
## Coefficients:
##                          (Intercept)  
##                              1.26796  
##                 max_data_frame$teage  
##                              0.01503  
## as.factor(max_data_frame$Gender)male  
##                              0.20209

multiLines(age_gender_leisure_max_lm)

age_gender_work_max_lm = lm(max_data_frame$tuactdur24_5~max_data_frame$teage+as.factor(max_data_frame$Gender))
age_gender_work_max_lm

## 
## Call:
## lm(formula = max_data_frame$tuactdur24_5 ~ max_data_frame$teage + 
##     as.factor(max_data_frame$Gender))
## 
## Coefficients:
##                          (Intercept)  
##                             9.550398  
##                 max_data_frame$teage  
##                             0.003167  
## as.factor(max_data_frame$Gender)male  
##                             0.468375

multiLines(age_gender_work_max_lm)

## Min Hours: Summary

describe(min_data_frame$tuactdur24_5)

##    vars    n mean  sd median trimmed mad min  max range skew kurtosis   se
## X1    1 8499 1.14 2.4      0    0.52   0   0 7.98  7.98  1.9     1.98 0.03

describe(min_data_frame$tuactdur24_12)

##    vars    n mean   sd median trimmed  mad min  max range skew kurtosis
## X1    1 8499  5.7 3.51   5.17    5.43 3.58   0 23.2  23.2 0.69     0.28
##      se
## X1 0.04

describeBy(min_data_frame$tuactdur24_5,group = 
             min_data_frame$Employment_Status, mat = TRUE)

##     item            group1 vars    n       mean        sd median   trimmed
## X11    1          Employed    1 4474 2.11817166 2.9506412      0 1.6861825
## X12    2 Out of Work Force    1 3694 0.03297239 0.3722608      0 0.0000000
## X13    3        Unemployed    1  331 0.31163142 1.0449477      0 0.0318239
##     mad min      max    range       skew    kurtosis          se
## X11   0   0 7.983333 7.983333  0.9357372  -0.8394584 0.044113184
## X12   0   0 7.500000 7.500000 13.9706090 213.7941289 0.006124898
## X13   0   0 7.800000 7.800000  4.3069768  19.9906945 0.057435519

describeBy(min_data_frame$tuactdur24_12,group = 
             min_data_frame$Employment_Status, mat = TRUE)

##     item            group1 vars    n     mean       sd   median  trimmed
## X11    1          Employed    1 4474 4.751274 3.057170 4.250000 4.504912
## X12    2 Out of Work Force    1 3694 6.804246 3.673031 6.500000 6.615212
## X13    3        Unemployed    1  331 6.076939 3.570524 5.666667 5.837987
##         mad min      max    range      skew  kurtosis         se
## X11 3.08875   0 20.43333 20.43333 0.7654666 0.3973029 0.04570583
## X12 3.83005   0 23.20000 23.20000 0.5041833 0.0302574 0.06043328
## X13 3.95360   0 20.45000 20.45000 0.6839070 0.3953828 0.19625375

plot(min_data_frame$tuactdur24_5~min_data_frame$teage)

plot(min_data_frame$tuactdur24_12~min_data_frame$teage)

plot(min_data_frame$tuactdur24_12~min_data_frame$tuactdur24_5)

hist(min_data_frame$tuactdur24_12)

age_gender_leisure_min_lm = lm(min_data_frame$tuactdur24_12~min_data_frame$teage+as.factor(min_data_frame$Gender))
age_gender_leisure_min_lm

## 
## Call:
## lm(formula = min_data_frame$tuactdur24_12 ~ min_data_frame$teage + 
##     as.factor(min_data_frame$Gender))
## 
## Coefficients:
##                          (Intercept)  
##                               2.4889  
##                 min_data_frame$teage  
##                               0.0544  
## as.factor(min_data_frame$Gender)male  
##                               1.0727

multiLines(age_gender_leisure_min_lm)

age_gender_work_min_lm = lm(min_data_frame$tuactdur24_5~min_data_frame$teage+as.factor(min_data_frame$Gender))
age_gender_work_min_lm

## 
## Call:
## lm(formula = min_data_frame$tuactdur24_5 ~ min_data_frame$teage + 
##     as.factor(min_data_frame$Gender))
## 
## Coefficients:
##                          (Intercept)  
##                              2.17758  
##                 min_data_frame$teage  
##                             -0.02244  
## as.factor(min_data_frame$Gender)male  
##                              0.24582

multiLines(age_gender_work_min_lm)

Findings of Exploratory Analysis

The general summary statistics indicate a couple of key high level facts:

Males tend to work more than their female counterparts and enjoy more leisure time. This is an odd statement if one is only looking at the scope of the variables in this project, work and leisure. However, if one were to make an assumption based on societal norms, women may spend the difference between work and leisure on other activities such as tending to their family, chores, and variety of other things.
There appears to be a downward trend when examining work against age; as age increases time spent working decreases. In addition, as one continues to age leisure time increases. This however is not true from the start as one ages from a teenager to about 45 years old work sees an increase and leisure time decreases. Thus, it can be stated that work increases then decreases for a specific range of ages; from a practical standpoint that makes sense as one is expected to work more in their adult years versus their youth and are expected to work less once they begin approaching a elderly age. This can be seen below from the coefficients of the linear model for leisure and work.

Leisure:

\(y = 1.79800 + 0.06048 * age + 0.59100 * 1\)

Work:

\(y = 4.35254 + -0.04391 * age + 1.11036 * 1\)

In what may be a clear relationship, leisure and work times vary based on employment status with leisure time being highest amoung those that are out of the work force or unemployed and work being highest among those that are employed.
The general relationship that is established is that as leisure increases hours worked decreases and vice versa.
What the general statistics make clear is that for this analysis we must break out the data into groups based on average hours worked, greater than average hours worked, and less than average hours worked. Looking at the data holistically skews the data and does not allow us to discern if additional hours worked above or below the average affect our definition of work life balance.

Findings of groupings based on hours:

The groups are divided based on those that work greater than 8 hours, less than 8 hours, and the exact 9-5 day of 8 hours.

While analyzing the summary statistics of each group for leisure and work, I noted that the means for those who work 8 hours on average have about .53 hours more of leisure while they work ~2 hours less than the group that works above 8 hours on average. The 8 hour group also experiences 3.12 hours less of leisure than those that work less than 8 hours on average, while also working 6.86 hours more than that group. We can already see a high level difference in leisure time and work time between the three groups.
Examining work and leisure across employment levels also supports the high level analysis and provides more in depth insight into the data. For those that work 8 hours or more there are hardly any people categorized as unemployed which could be data anomalies or exceptions based on the scope of the definition. It is clear that the highest amount of employed workers is among those two groups and that they both experience, on average, ~3 hours of leisure with leisure decreasing as hours worked increases and vice versa.
Gender findings from the higher level statistics are confirmed by the relationships existing within these three groups of observations. Men work and experience leisure at higher rates than their female counterparts. However the fact that men can both work and experience more leisure than females indicates that females spend their time on an additional elements outside of these two categories; this additional research is beyond the scope of the project.
Age findings from the higher level statistics are also confirmed by the relationships existing within these groups, as the youth and elderly work less and experience more leisure than those who are in the prime of their lives.

Inference

Hypothesis

\({ H }_{ 0 } = Leisure(work > 8 hrs) = Leisure(work=8hrs) = Leisure(work<8hrs\) \({ H }_{ A } = Leisure(work > 8 hrs) \neq Leisure(work=8hrs) \neq Leisure(work<8hrs)\)

Checking Conditions

## Leisure Comparison

boxplot(min_data_frame$tuactdur24_12,avg_data_frame$tuactdur24_12,max_data_frame$tuactdur24_12,names=c('< 8 Hours work','8 hours work','> 8 hours work'), main = 'Work Life Balance', ylab = 'Leisure')

tempdf = merge_table_one_rost
tempdf$workflag = ifelse(tempdf$tuactdur24_5>=8,'Max Work','Min Work')

## Reject null hypothesis as 0 is not in CI range and p value less than .05

inference(y = tempdf$tuactdur24_12, x = tempdf$workflag, est = "mean", type = "ci", null = 0, 
          alternative = "twosided", method = "theoretical")

## Response variable: numerical, Explanatory variable: categorical
## Difference between two means
## Summary statistics:
## n_Max Work = 1816, mean_Max Work = 2.0853, sd_Max Work = 1.5949
## n_Min Work = 8499, mean_Min Work = 5.6952, sd_Min Work = 3.5053

## Observed difference between means (Max Work-Min Work) = -3.6099
## 
## Standard error = 0.0534 
## 95 % Confidence interval = ( -3.7144 , -3.5053 )

inference(y = tempdf$tuactdur24_12, x = tempdf$workflag, est = "mean", type = "ht", null = 0, 
          alternative = "twosided", method = "theoretical")

## Response variable: numerical, Explanatory variable: categorical
## Difference between two means
## Summary statistics:
## n_Max Work = 1816, mean_Max Work = 2.0853, sd_Max Work = 1.5949
## n_Min Work = 8499, mean_Min Work = 5.6952, sd_Min Work = 3.5053

## Observed difference between means (Max Work-Min Work) = -3.6099
## 
## H0: mu_Max Work - mu_Min Work = 0 
## HA: mu_Max Work - mu_Min Work != 0 
## Standard error = 0.053 
## Test statistic: Z =  -67.662 
## p-value =  0

Given that the CI theoretical inference model does not include the null hypothesis value in the CI interval and the hypothesis test theoretical inference model has a p-value less than .05, we reject the null hypothesis in favor of the alternative concluding that work life balance as we have defined it (balance between leisure and work) is affected by the number of hours worked below or above the defined average hours (9 to 5 job or 8 hours).

Conclusion

This research project has given me great insight into how leisure and work is divided among employment, gender, and age groups. From a statistically significant stance there is a clear difference in “work life balance” based on hours worked among the groups and at a general level, but from a practical sense the concept has evolved over time and has been challenged in the modern day. It very well could be that we reach a point where the current definition is obsolete and this research would need to be redone. For future analysis it would be interesting to bring in other variables to further examine the breakout of people’s spent time outside of these two categories (i.e sports, religion, chores etc.) and possibly discern if a field could be added to indicate whether people consider these activites as additions to their labor or their leisure time. With the additional field a broader scope can be applied to the concept of ‘work/life balance’ and more interesting insights could be produced.

Appendix

Condensed estimates for activity by average hours per day from the 2016 survey can be found at https://www.bls.gov/tus/a1_2016.pdf

https://smallbusiness.chron.com/constitutes-work-day-salaried-employee-18816.html