Introduction

The association between modifiable risk factors such as binge or episodic drinking and mental health has been a primary focus of many research studies. Similarly, increased smoking behavior has also been associated with higher levels of mental distress levels. In this brief analysis, we will focus on, two of the most common modifiable risk factors “Alcohol Consumption” and “Cigarette Smoking” and their effects on mental health.

Summary of Data

In this analysis, the data set from National Health Interview Survey will be used. The NHIS data has been extensively used throughout the Department of Health and Human Services (DHHS) to track and monitor trends in illness and disability and also to track progress toward achieving national health objectives. The data is also used by the public health research communities for epidemiology and policy analysis of issues as characterizing those with various health problems, determining barriers to accessing and using appropriate health care and evaluating Federal health programs (CDC).

Main Dependent Variable

For the mental distress levels, K6 (Nonspecific Psychological Scale) variable was formed. The K6 consist of six questions that were asked in NHIS about the frequency of experiencing the symptoms of mental distress. The K6 nonspecific distress scale has been extensively used in the epidemiological studies and its ability to discriminate DSM-IV (Diagnostic and Statistical Manual of Mental Disorders) cases from non-cases make the K6 and K10 attractive for use in general-purpose health surveys. (Kessler et al). The variables was formed by combing different scores from 6 different questions in the NHIS data set into one consolidated variable K6. The variable, K6, represent three different levels of mental distress 1- NORMAL (0-5 Score), 2- Moderate (6-12 Score), 3- Severe (13-24 Score).

Key Independent Variables

The variable “drinks” in the study will be used to predict the high alcohol consumption and risk of more severe mental distress levels. The default variable in the data sets is “alcamt” which states the average number of drinks on days drank according to the NHIS code book. The variable is recoded into a new metric/numerical variable drinks which are divided by 10 because the default variable is coded in 10s; for example, 1 drink is coded as 10, and 10 drinks are coded as 100. So the variable was divided by ‘10’ and then recoded for missing values. After this, another variable was created to categorize the average drinks per day into five categories <5, 06-10, 11-15, and 16-20 drinks. The reference category for this variable is <5 drinks per day.Similarly, the variable “cigs” will be used to investigate the effect on the dependent variable. The average cigarettes variables define the average cigarettes smoked per day of the respondents on the overall basis. The reference category for this variable is <5 cigarettes smoked. The other 3 categories are 6-10, 11-19, and 20-40.We will also use the variable “sex” to analyze gender differences as interaction variable.

Statistical Analysis

For the analysis, an ordered logistic regression model will be used. The ordered logistic is regression is precisely used when the dependent variable is categorical with multiple categories. Also, we will use an bivariate table to analyze the key independent and the dependent variables. This would help us to see the percentage of cases in each category of mental distress level and also help us understand the stregth of the effect of independent variables.

glimpse(nhis)
## Observations: 795,645
## Variables: 29
## $ YEAR        <int> 2010, 2010, 2010, 2010, 2010, 2010, 2010, 2010, 20...
## $ SERIAL      <int> 1, 1, 1, 2, 2, 2, 3, 3, 3, 3, 5, 6, 6, 6, 6, 9, 9,...
## $ STRATA      <int> 6001, 6001, 6001, 6249, 6249, 6249, 6045, 6045, 60...
## $ PSU         <int> 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1, 2, 2, 2, 2, 2, 2,...
## $ NHISHID     <int> 2010000001, 2010000001, 2010000001, 2010000002, 20...
## $ HHWEIGHT    <int> 3042, 3042, 3042, 1124, 1124, 1124, 3713, 3713, 37...
## $ NHISPID     <dbl> 2.01e+13, 2.01e+13, 2.01e+13, 2.01e+13, 2.01e+13, ...
## $ HHX         <int> 1, 1, 1, 2, 2, 2, 6, 6, 6, 6, 8, 9, 9, 9, 9, 13, 1...
## $ FMX         <int> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,...
## $ PX          <int> 1, 2, 3, 1, 2, 3, 1, 2, 3, 4, 1, 1, 2, 3, 4, 1, 2,...
## $ PERNUM      <int> 1, 2, 3, 1, 2, 3, 1, 2, 3, 4, 1, 1, 2, 3, 4, 1, 2,...
## $ PERWEIGHT   <int> 3248, 3981, 3181, 1454, 1370, 1607, 4661, 4396, 42...
## $ SAMPWEIGHT  <int> 6011, 0, 2895, 0, 0, 8310, 11865, 0, 8702, 0, 3501...
## $ FWEIGHT     <int> 3181, 3181, 3181, 1370, 1370, 1370, 4228, 4228, 42...
## $ ASTATFLG    <int> 1, 3, 0, 3, 3, 1, 1, 3, 0, 0, 1, 3, 2, 0, 0, 3, 3,...
## $ CSTATFLG    <int> 0, 0, 1, 0, 0, 0, 0, 0, 1, 3, 0, 0, 0, 1, 3, 0, 0,...
## $ SEX         <int> 2, 1, 2, 2, 2, 1, 1, 2, 2, 1, 2, 2, 1, 1, 1, 1, 2,...
## $ ALCAMT      <int> 20, 0, 0, 0, 0, 40, 120, 0, 0, 0, 0, 0, 0, 0, 0, 0...
## $ CIGDAYMO    <int> 96, 96, 96, 96, 96, 96, 96, 96, 96, 96, 96, 96, 96...
## $ CIGSDAY     <int> 96, 96, 96, 96, 96, 96, 60, 96, 96, 96, 96, 96, 96...
## $ CIGSDAY1    <int> 0, 0, 0, 0, 0, 0, 60, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0...
## $ CIGSDAY2    <int> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,...
## $ AEFFORT     <int> 0, 6, 6, 6, 6, 0, 3, 6, 6, 6, 0, 6, 6, 6, 6, 6, 6,...
## $ AFEELINT1MO <int> 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0,...
## $ AHOPELESS   <int> 0, 6, 6, 6, 6, 0, 1, 6, 6, 6, 0, 6, 6, 6, 6, 6, 6,...
## $ ANERVOUS    <int> 0, 6, 6, 6, 6, 1, 3, 6, 6, 6, 2, 6, 6, 6, 6, 6, 6,...
## $ ARESTLESS   <int> 0, 6, 6, 6, 6, 0, 4, 6, 6, 6, 2, 6, 6, 6, 6, 6, 6,...
## $ ASAD        <int> 0, 6, 6, 6, 6, 1, 1, 6, 6, 6, 2, 6, 6, 6, 6, 6, 6,...
## $ AWORTHLESS  <int> 0, 6, 6, 6, 6, 0, 1, 6, 6, 6, 2, 6, 6, 6, 6, 6, 6,...

Applying Data Cleaning Procedures

Recoded the Sex as a new labeled variable.

nhis <- mutate(nhis, sex=SEX)
nhis$sex <- car::recode(nhis$sex, "2=0; 1=1")
nhis$sex <-factor(nhis$sex, labels = c("Female", "Male"))

Created Female as another gender variable, where 1=Female and 0=Male.

nhis <- mutate(nhis, female=SEX)
nhis$female <- car::recode(nhis$female, "2=1; 1=0")

Formation of K6 Composite Score Variable

nhis <- mutate(nhis, sad=ASAD, effort=AEFFORT, nervous=ANERVOUS, hopeless=AHOPELESS,restless=ARESTLESS,
               worthless=AWORTHLESS)
nhis$effort <- car::recode(nhis$effort,"6:9 = NA")

nhis$hopeless <- car::recode(nhis$hopeless,"6:9 = NA")

nhis$nervous <- car::recode(nhis$nervous,"6:9 = NA")

nhis$restless <- car::recode(nhis$restless,"6:9 = NA")

nhis$sad <- car::recode(nhis$sad,"6:9 = NA")

nhis$worthless <- car::recode(nhis$worthless,"6:9 = NA")
nhis <- mutate(nhis, K6= effort+hopeless+nervous+restless+sad+worthless)

Giving appropriate categorical scores.

nhis$K6 <- car::recode(nhis$K6, "0:5=1; 6:12=2; 13:24=3")

Validating the new variable, as a factor variable

nhis$K6.f <- factor(nhis$K6)
is.factor(nhis$K6.f)
## [1] TRUE

Recoding for the odered logistic regression.

nhis$K6.f <-factor(nhis$K6.f, labels = c("Normal", "Moderate", "Severe"))

nhis$K6.f2 <- factor(nhis$K6.f, ordered = TRUE,
                         levels = c("Normal", "Moderate", "Severe"))

table.K6.f <- table(nhis$K6.f)

Checking new factor variable K6

View (table.K6.f)

Creating Variable of “Alchohal Drinks”

nhis <- mutate(nhis, drinks=ALCAMT/10)

nhis$drinks <- car::recode(nhis$drinks,"0 = NA; 21:99.9=NA")

Recoding the variable drinks into categories

nhis$drinks <- car::recode(nhis$drinks,"1:5=1; 6:10=2; 11:15=3; 16:20=4")

nhis$drinksf <- factor(nhis$drinks)

nhis$drinksf <-factor(nhis$drinksf, labels= c("<5", "Between 6-10", "Between 11-15", "Between 16-20"))

table.drinksf <- table(nhis$drinksf)

View (table.drinksf)

Creating Variable of “cigs”

nhis <- mutate(nhis, cigs=CIGSDAY)

nhis$cigs <- car::recode(nhis$cigs,"0=NA; 41:99=NA")

Recoding the variable cigs into categories

nhis$cigs <- car::recode(nhis$cigs,"1:5=1; 6:10=2; 11:19=3; 20:40=4")

nhis$cigsf <- factor(nhis$cigs)

nhis$cigsf <-factor(nhis$cigsf, labels = c("<5","Between 6-10","Between 11-19","Between 20-40"))

table.cigsf <- table(nhis$cigsf)

View (table.cigsf)

Descriptives of Cleaned Data

Using the compareGroups package we created a descriptives table to see the overall percentages of our independent variables in regard to mental distress levels. This will help us to view an overall condition of our data.

From the table, it can been seen that, when it comes to gender, females have higher reported cases of mental distress levels as compared to males. 63% of females reported to have severe mental distress levels as compared 36% of men.

For Drinks consumed, people who consumed between 16-20 drinks, there is very small percentage of of people in each of the mental distress level category. This indicates that there might be a weak effect of episodic or binge drinking on mental health.

Interestingly, for cigarettes smoked, the effect seems to be strong. This is because, in the between 20-40 cigarettes smoked category there is number almost 38% cases reported to be of severe mental distress, and 30% to be reported to be moderate mental distress level, and 26% to be normal.

These are some of important findings from the descriptives table which help us understand basic question of this assingment.

export2md(check2, header.labels = c(p.overall = "p-value"))
Summary descriptives table by groups of `K6.f’
Normal N=209416 Moderate N=33584 Severe N=9530 p-value
sex: <0.001
    Female 112914 (53.9%) 20517 (61.1%) 6093 (63.9%)
    Male 96502 (46.1%) 13067 (38.9%) 3437 (36.1%)
drinksf: <0.001
    <5 125157 (94.2%) 19002 (91.7%) 4564 (89.0%)
    Between 6-10 6719 (5.06%) 1452 (7.01%) 446 (8.70%)
    Between 11-15 914 (0.69%) 206 (0.99%) 96 (1.87%)
    Between 16-20 122 (0.09%) 52 (0.25%) 22 (0.43%)
cigsf: <0.001
    <5 9571 (31.1%) 2448 (28.0%) 856 (23.5%)
    Between 6-10 9638 (31.3%) 2625 (30.1%) 1037 (28.5%)
    Between 11-19 3329 (10.8%) 974 (11.2%) 344 (9.45%)
    Between 20-40 8259 (26.8%) 2681 (30.7%) 1405 (38.6%)

Barplots to give more clear view of the data.

plot(check1[1:3], bivar=T, type=3)

Model-1 “Drinks Consumed” + “Cigs Smoked” vs “K6 (Mental Distress Level”

The model predicts the effect of binge drinking and cigs smoked on mental distres levels. To have better understanding of the model, we exponentiated the coefficients into odds ratios just to interpret the main effects.

Key Interpretations:

1- People who drink 6-10 drinks have increased odds of moving to a more severe mental distress category by factor of 1.02

2- People who drink 16-20 drinks have increased odds of moving to a more severe mental distress category by factor of 1.92

3- People who smoke 11-19 cigarettes have increased odds of moving to a more severe mental distress category by factor of 1.14

4- People who smoke 20-40 cigarettes have increased odds of moving to a more severe mental distress category by factor of 1.45

m1 <- zelig(K6.f ~ drinksf + cigsf, model = "ologit", data = nhis, cite = F)
summary(m1)
## Model: 
## Call:
## z5$zelig(formula = K6.f ~ drinksf + cigsf, data = nhis)
## 
## Coefficients:
##                          Value Std. Error  t value
## drinksfBetween 6-10  0.0009579    0.03947  0.02427
## drinksfBetween 11-15 0.1448629    0.08886  1.63025
## drinksfBetween 16-20 0.7417602    0.18930  3.91848
## cigsfBetween 6-10    0.0855528    0.03276  2.61166
## cigsfBetween 11-19   0.1283129    0.04418  2.90434
## cigsfBetween 20-40   0.3642636    0.03278 11.11093
## 
## Intercepts:
##                 Value   Std. Error t value
## Normal|Moderate  1.1348  0.0236    48.0509
## Moderate|Severe  2.6939  0.0296    90.8686
## 
## Residual Deviance: 46535.80 
## AIC: 46551.80 
## (764228 observations deleted due to missingness)
## Next step: Use 'setx' method
exp(coef(m1))
##  drinksfBetween 6-10 drinksfBetween 11-15 drinksfBetween 16-20 
##             1.000958             1.155881             2.099628 
##    cigsfBetween 6-10   cigsfBetween 11-19   cigsfBetween 20-40 
##             1.089319             1.136909             1.439454

Setting Counterfactuals For Alcohal Drinks

x.low <- setx(m1, drinksf = "<5") 
x.high <- setx(m1, drinksf = "Between 16-20")

Running Simulation

We used the simulation for the effects of episodic drinking on mental health. Where two counter factual situations were given 1- <5 drinks and 2- Between 16-20 Drinks. From the summary, we can see the proability of developing is severe mental distress level is low as 0.06 when a person drinks less than 5 drinks, however, if a person drinks betweem 16-20 drinks the probability of developing mental distress levels is high as 11%. The first differences tells predicted probabilty of developing mental distress level is 0.05 between <5 drinks and between 16-20. This simply indicates that people who drink more than 5 drinks are 5% more likely to develop mental distress level as compared to people who drink less than 5 drinks.

s.drinks_m1<- sim(m1, x = x.low, x1 = x.high)
summary(s.drinks_m1)
## 
##  sim x :
##  -----
## ev
##                mean          sd        50%       2.5%      97.5%
## Normal   0.75666867 0.004350249 0.75657456 0.74838795 0.76500964
## Moderate 0.17993224 0.001994205 0.17990057 0.17620592 0.18392857
## Severe   0.06339908 0.003893040 0.06333277 0.05621198 0.07097624
## pv
##       mean        sd 50% 2.5% 97.5%
## [1,] 1.287 0.5682904   1    1     3
## 
##  sim x1 :
##  -----
## ev
##               mean         sd       50%       2.5%     97.5%
## Normal   0.5976138 0.04508028 0.5980924 0.50742387 0.6854678
## Moderate 0.2772370 0.02474130 0.2783962 0.22625548 0.3250617
## Severe   0.1251492 0.02156441 0.1240729 0.08775881 0.1720230
## pv
##      mean        sd 50% 2.5% 97.5%
## [1,] 1.54 0.7133785   1    1     3
## fd
##                 mean         sd         50%        2.5%      97.5%
## Normal   -0.15905490 0.04490566 -0.15925892 -0.25170153 -0.0714073
## Moderate  0.09730478 0.02462923  0.09838101  0.04661683  0.1447511
## Severe    0.06175012 0.02074528  0.06048290  0.02440759  0.1067845

Plots for Alcohal Drinks and Predicted Probility of Mental Distress Levels

plot(s.drinks_m1)

Setting Counterfactuals For Cigarettes Smoked

x.low <- setx(m1, cigsf = "<5")
x.high <- setx(m1, cigsf = "Between 20-40")

Running Simulation for “cigs” smoked and mental health.

Similarly, for cigarette smoking the it was found that the smokers who smoked <5 cigarettes are less likely to develop severe mental distress level (0.06). However, smoker who smoked 20-40 cigarettes are more likely to develop mental distress levels (0.08). The first differences tells us that the people who smoke less than 5 cigarettes are 7% less likely to develop mental distress levels.

s.cigs_m1 <- sim(m1, x = x.low, x1 = x.high)
summary(s.cigs_m1)
## 
##  sim x :
##  -----
## ev
##                mean          sd       50%       2.5%      97.5%
## Normal   0.75632891 0.004234028 0.7565981 0.74772786 0.76467173
## Moderate 0.17996155 0.001981824 0.1800185 0.17607060 0.18376952
## Severe   0.06370954 0.003896524 0.0635669 0.05647336 0.07127411
## pv
##       mean        sd 50% 2.5% 97.5%
## [1,] 1.298 0.5757843   1    1     3
## 
##  sim x1 :
##  -----
## ev
##                mean          sd        50%       2.5%      97.5%
## Normal   0.68372618 0.005120622 0.68374320 0.67369977 0.69403084
## Moderate 0.22733231 0.004999989 0.22732143 0.21786808 0.23727089
## Severe   0.08894151 0.004151857 0.08896061 0.08098005 0.09682775
## pv
##       mean        sd 50% 2.5% 97.5%
## [1,] 1.401 0.6577474   1    1     3
## fd
##                 mean          sd         50%        2.5%       97.5%
## Normal   -0.07260273 0.006387744 -0.07265003 -0.08478372 -0.05970240
## Moderate  0.04737075 0.004733811  0.04757514  0.03822193  0.05651808
## Severe    0.02523197 0.002002351  0.02520585  0.02132065  0.02915438

Plots for Cigs Smoked and Predicted Probility of Mental Distress Levels

plot(s.cigs_m1)

Model-2: ‘Drinks Consumed * Female’ + ‘Cigs Smoked* Female’ vs “K6 (Mental Distress Level”

The model#2 takes account of the interaction variable female. From the model following key findings have been obtained. Key Interpretations:

1- Being female increases your odds of moving to a more severe mental distress category by a factor of 1.68

2- Interestingly, being female, who drink 16-20 drinks have decresed odds of moving to a more severe mental distress category by factor of 0.69

3- Being female, who smoke 11-19 cigarettes have decreased odds of moving to a more severe mental distress category by factor of 0.94

m2 <- zelig(K6.f ~ drinksf*female + cigsf*female, model = "ologit", data = nhis, cite = F)
summary (m2)
## Model: 
## Call:
## z5$zelig(formula = K6.f ~ drinksf * female + cigsf * female, 
##     data = nhis)
## 
## Coefficients:
##                                Value Std. Error t value
## drinksfBetween 6-10          0.10370    0.04924  2.1061
## drinksfBetween 11-15         0.22176    0.10249  2.1638
## drinksfBetween 16-20         0.97894    0.21039  4.6530
## female                       0.54580    0.04774 11.4327
## cigsfBetween 6-10            0.10280    0.05033  2.0425
## cigsfBetween 11-19           0.17369    0.06399  2.7143
## cigsfBetween 20-40           0.40127    0.04648  8.6341
## drinksfBetween 6-10:female   0.08373    0.08540  0.9805
## drinksfBetween 11-15:female  0.50077    0.21675  2.3104
## drinksfBetween 16-20:female -0.30139    0.48823 -0.6173
## female:cigsfBetween 6-10    -0.08692    0.06661 -1.3049
## female:cigsfBetween 11-19   -0.06091    0.08893 -0.6849
## female:cigsfBetween 20-40    0.03629    0.06636  0.5469
## 
## Intercepts:
##                 Value   Std. Error t value
## Normal|Moderate  1.4285  0.0357    40.0639
## Moderate|Severe  3.0024  0.0402    74.7324
## 
## Residual Deviance: 46094.02 
## AIC: 46124.02 
## (764228 observations deleted due to missingness)
## Next step: Use 'setx' method
exp(coef(m2))
##         drinksfBetween 6-10        drinksfBetween 11-15 
##                   1.1092685                   1.2482709 
##        drinksfBetween 16-20                      female 
##                   2.6616442                   1.7259854 
##           cigsfBetween 6-10          cigsfBetween 11-19 
##                   1.1082644                   1.1896860 
##          cigsfBetween 20-40  drinksfBetween 6-10:female 
##                   1.4937268                   1.0873361 
## drinksfBetween 11-15:female drinksfBetween 16-20:female 
##                   1.6499914                   0.7397869 
##    female:cigsfBetween 6-10   female:cigsfBetween 11-19 
##                   0.9167522                   0.9409104 
##   female:cigsfBetween 20-40 
##                   1.0369552

Counterfactuals for Alcohal Drinks (Model#2)

x.low <- setx(m2, drinksf = "<5") 
x.high <- setx(m2, drinksf = "Between 16-20")

Running Simulations

Interestingly, model-2 presents more texured results. The first differences from simulation based on model tells us, that the different predicited probability of developing severe mental distress is 0.05 between two counter factual situations. So it can be said, people who drink between 1-20 drinks are 5% likely to develop severe mental distress levels.

s.drinksf_m2<- sim(m2, x = x.low, x1 = x.high)
summary(s.drinksf_m2)
## 
##  sim x :
##  -----
## ev
##                mean          sd       50%       2.5%      97.5%
## Normal   0.76402744 0.004552034 0.7639713 0.75492116 0.77317113
## Moderate 0.17586727 0.002697612 0.1758135 0.17078030 0.18118351
## Severe   0.06010529 0.004826622 0.0599673 0.05115888 0.06970437
## pv
##       mean        sd 50% 2.5% 97.5%
## [1,] 1.299 0.5743944   1    1     3
## 
##  sim x1 :
##  -----
## ev
##               mean         sd       50%       2.5%     97.5%
## Normal   0.5809438 0.05681169 0.5823045 0.46753340 0.6850015
## Moderate 0.2874402 0.03085673 0.2874365 0.22562425 0.3469616
## Severe   0.1316160 0.02805548 0.1295232 0.08573101 0.1925056
## pv
##       mean        sd 50% 2.5% 97.5%
## [1,] 1.523 0.7028264   1    1     3
## fd
##                mean         sd         50%        2.5%       97.5%
## Normal   -0.1830836 0.05672413 -0.18191267 -0.29862571 -0.07805681
## Moderate  0.1115729 0.03047272  0.11192514  0.05228560  0.17000970
## Severe    0.0715107 0.02709907  0.06903953  0.02603809  0.12954914

Counterfactuals - Gender differences.

x.low <- setx(m2, female = 0) 
x.high <- setx(m2, female = 1)

Running Simulations for Gender Differences.

The Simulation for gender differences found that the probability of developing severe mental distress is 0.07 for females, and 0.04 for males. The simualated first difference between men and women of developing severe mental distress is 0.02. This indicates that females are 2% more likely to develop severe mental distress, as comapred to males.

s.sex_m2<- sim(m2, x = x.low, x1 = x.high)
summary(s.sex_m2)
## 
##  sim x :
##  -----
## ev
##                mean          sd        50%       2.5%      97.5%
## Normal   0.80648786 0.005381296 0.80641593 0.79590097 0.81675104
## Moderate 0.14598422 0.002000856 0.14591811 0.14199415 0.15002705
## Severe   0.04752791 0.004212222 0.04746313 0.03988848 0.05539142
## pv
##       mean        sd 50% 2.5% 97.5%
## [1,] 1.238 0.5173235   1    1     3
## 
##  sim x1 :
##  -----
## ev
##                mean          sd        50%      2.5%      97.5%
## Normal   0.70709260 0.006692022 0.70718694 0.6939185 0.71980557
## Moderate 0.21372469 0.006392670 0.21354319 0.2017735 0.22696437
## Severe   0.07918271 0.004996372 0.07897018 0.0699054 0.08948998
## pv
##      mean       sd 50% 2.5% 97.5%
## [1,] 1.38 0.638755   1    1     3
## fd
##                 mean          sd         50%        2.5%       97.5%
## Normal   -0.09939526 0.008542775 -0.09898266 -0.11639649 -0.08331130
## Moderate  0.06774047 0.006935330  0.06722014  0.05491137  0.08177678
## Severe    0.03165480 0.002423963  0.03162185  0.02676999  0.03651698

Better Model

Comparing two models, it was found that the model#2 is a slightly better model, as it gives us more texutred results. Also the obatied AIC values for model 1 is = 42368, where for model 2 is = 42007, which also clearly tells that model#2 is a better model.

htmlreg(list(m1, m2))
Statistical models
Model 1 Model 2
drinksfBetween 6-10 0.00 0.10*
(0.04) (0.05)
drinksfBetween 11-15 0.14 0.22*
(0.09) (0.10)
drinksfBetween 16-20 0.74*** 0.98***
(0.19) (0.21)
cigsfBetween 6-10 0.09** 0.10*
(0.03) (0.05)
cigsfBetween 11-19 0.13** 0.17**
(0.04) (0.06)
cigsfBetween 20-40 0.36*** 0.40***
(0.03) (0.05)
female 0.55***
(0.05)
drinksfBetween 6-10:female 0.08
(0.09)
drinksfBetween 11-15:female 0.50*
(0.22)
drinksfBetween 16-20:female -0.30
(0.49)
female:cigsfBetween 6-10 -0.09
(0.07)
female:cigsfBetween 11-19 -0.06
(0.09)
female:cigsfBetween 20-40 0.04
(0.07)
AIC 46551.80 46124.02
BIC 46618.64 46249.35
Log Likelihood -23267.90 -23047.01
Deviance 46535.80 46094.02
Num. obs. 31417 31417
p < 0.001, p < 0.01, p < 0.05

Conclusion

Results show that high alcohol consumption and smoking were bothindependently associated more severe mental distress levels. Among moderate drinkers (10-15 drinks) showed significantly higher odds ratios for psychological distress than light drinkers (<5 drinks). Interestingly, it found for, females, who are moderate drinkers (10-15 drinks) have lower odds of developing mental distress. The findings of this study needs to be futher analyzed with more variables. As modifiable risk factors may have different effects on different types of people, this study can be futher investigated by adding important factors such as race, BMI, poverty threshold and other immutable charatecistics factors.