Regressions_Final
2022-11-20
Dataframe Setup
df = read.csv("HSV_Final.csv")
df <- subset(df, select = -c(X, WAIST_SIZE, MISSING_WAIST_SIZE))
# Nonzero WWH
df_nonzero <- subset(df, df$WEEKLY_WORK_HOURS > 0)
df_any <- df
df_any$ZERO_WWH <- ifelse(df_any$WEEKLY_WORK_HOURS > 0, 1, 0)
df_any <- subset(df_any, select = -WEEKLY_WORK_HOURS)
# Log dependent
df_logged <- df_nonzero
df_logged$LOG_WWH <- log(df_logged$WEEKLY_WORK_HOURS)
df_logged <- subset(df_logged, select = -WEEKLY_WORK_HOURS)
# Transformed
df_transformed <- df
df_transformed$HOSPITAL_EXPENSES <- log(I(df_transformed$HOSPITAL_EXPENSES + 1))
df_transformed$DEBTS <- log(I(df_transformed$DEBTS + 1))
df_transformed$HOME_VALUE <- log(I(df_transformed$HOME_VALUE + 1))Histograms
fit.normal <- fitdistr(df_nonzero$WEEKLY_WORK_HOURS, "normal")
class(fit.normal)## [1] "fitdistr"para.normal <- fit.normal$estimate
hist(df_nonzero$WEEKLY_WORK_HOURS,
main="Weekly Work Hours - Normal Fit",
xlab="Weekly Work Hours",
col="darkblue",
freq=FALSE,
prob = TRUE)
curve(dnorm(x, para.normal[1], para.normal[2]), col = 2, add = TRUE)
fit.normal <- fitdistr(df_nonzero$AGE, "normal")
class(fit.normal)## [1] "fitdistr"para.normal <- fit.normal$estimate
hist(df_nonzero$AGE,
main="Age - Normal Fit",
xlab="Years of Age",
col="orange",
freq=FALSE,
prob = TRUE)
curve(dnorm(x, para.normal[1], para.normal[2]), col = 2, add = TRUE)
fit.normal <- fitdistr(df_nonzero$WEEKS_PAID_VACATION, "normal")
class(fit.normal)## [1] "fitdistr"para.normal <- fit.normal$estimate
hist(df_nonzero$WEEKS_PAID_VACATION,
main="Weeks of Paid Vacation - Normal Fit",
xlab="Weeks",
col="brown",
freq=FALSE,
prob = TRUE)
curve(dnorm(x, para.normal[1], para.normal[2]), col = 2, add = TRUE)
fit.normal <- fitdistr(df_nonzero$YEARS_EDUCATED, "normal")
class(fit.normal)## [1] "fitdistr"para.normal <- fit.normal$estimate
hist(df_nonzero$YEARS_EDUCATED,
main="Highest Education - Normal Fit",
xlab="Years of Education",
col="blue",
freq=FALSE,
prob = TRUE)
curve(dnorm(x, para.normal[1], para.normal[2]), col = 2, add = TRUE)
hist(df_nonzero$HOSPITAL_EXPENSES,
main="Hospital Expenses - Mass at 0, Skewed Right",
xlab="Total Expenses Units: thousands of dollars",
col="darkmagenta",
freq=FALSE,
prob = TRUE)
hist(df_nonzero$HOME_VALUE,
main="Primary Home Value - Mass at 0, Skewed Right",
xlab="Home Value Units: thousands of dollars",
col="darkgreen",
freq=FALSE,
prob = TRUE)
hist(df_nonzero$DEBTS,
main="Debts Owed - Mass at 0, Skewed Right",
xlab="Debts Owed Units: thousands of dollars",
col="darkred",
freq=FALSE,
prob = TRUE)
fit.normal <- fitdistr(df_transformed$HOSPITAL_EXPENSES, "normal")
class(fit.normal)## [1] "fitdistr"para.normal <- fit.normal$estimate
hist(df_transformed$HOSPITAL_EXPENSES,
main="Hospital Expenses: Transformed - More Normal",
xlab="ln(Total Expenses + 1) Units: thousands of dollars",
col="darkmagenta",
freq=FALSE,
prob = TRUE)
curve(dnorm(x, para.normal[1], para.normal[2]), col = 2, add = TRUE)
fit.normal <- fitdistr(df_transformed$HOME_VALUE, "normal")
class(fit.normal)## [1] "fitdistr"para.normal <- fit.normal$estimate
hist(df_transformed$HOME_VALUE,
main="Primary Home Value: Transformed - More Normal",
xlab="ln(Home Value + 1) Units: thousands of dollars",
col="darkgreen",
freq=FALSE,
prob = TRUE)
curve(dnorm(x, para.normal[1], para.normal[2]), col = 2, add = TRUE)
fit.normal <- fitdistr(df_transformed$DEBTS, "normal")
class(fit.normal)## [1] "fitdistr"para.normal <- fit.normal$estimate
hist(df_transformed$DEBTS,
main="Debts Owed: Transformed - More Normal",
xlab="ln(Debts Owed + 1) Units: thousands of dollars",
col="darkred",
freq=FALSE,
prob = TRUE)
curve(dnorm(x, para.normal[1], para.normal[2]), col = 2, add = TRUE)
OLS and Stepwise
OLS is just as good, very little change in AIC
lm.model <- lm(WEEKLY_WORK_HOURS ~ ., data = df)
lm.step.both <- stepAIC(lm.model, direction = "both", trace = FALSE)
summary(lm.model)##
## Call:
## lm(formula = WEEKLY_WORK_HOURS ~ ., data = df)
##
## Residuals:
## Min 1Q Median 3Q Max
## -39.476 -5.405 -0.553 4.866 68.511
##
## Coefficients: (1 not defined because of singularities)
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.447e+01 2.287e+00 19.440 < 2e-16 ***
## AGE -1.955e-01 3.388e-02 -5.770 8.36e-09 ***
## HEART_CONDITION 1.583e-01 4.736e-01 0.334 0.738231
## ANY_DEPENDENTS 8.008e-02 3.990e-01 0.201 0.840940
## WORKING_SPOUSE -1.334e-01 3.542e-01 -0.377 0.706379
## WEEKS_PAID_VACATION 8.100e-01 8.127e-02 9.967 < 2e-16 ***
## REDUCE_PAID_WORK_HOURS -9.218e-01 3.664e-01 -2.516 0.011900 *
## MEDICARE -9.265e-01 5.626e-01 -1.647 0.099660 .
## MEDICAID -3.827e+00 5.526e-01 -6.925 4.86e-12 ***
## HOSPITAL_EXPENSES 3.241e-01 1.278e-01 2.535 0.011262 *
## RETIRED -1.313e+01 4.930e-01 -26.641 < 2e-16 ***
## VOLUNTEER -4.267e-01 3.208e-01 -1.330 0.183557
## HOME_VALUE -2.004e-04 2.433e-04 -0.824 0.410151
## MALE 5.127e+00 3.115e-01 16.460 < 2e-16 ***
## YEARS_EDUCATED 1.626e-02 5.679e-02 0.286 0.774622
## WORK_LIMITING_CONDITION -1.303e+00 4.881e-01 -2.670 0.007600 **
## DEBTS 8.539e-04 7.340e-04 1.163 0.244747
## MISSING_WEEKS_PAID_VACATION 2.152e+00 8.972e-01 2.399 0.016464 *
## MISSING_REDUCE_PAID_WORK_HOURS -1.036e+00 9.129e-01 -1.135 0.256466
## MISSING_HOME_VALUE -2.012e-01 3.470e-01 -0.580 0.561962
## MISSING_DEBTS NA NA NA NA
## ACTIVE_ONCE_WEEKLY 7.725e-01 4.275e-01 1.807 0.070812 .
## ACTIVE_DAILY 1.115e+00 4.848e-01 2.299 0.021533 *
## NOT_ACTIVE 1.311e+00 3.939e-01 3.328 0.000881 ***
## EXCELLENT_HEALTH 4.779e-01 5.319e-01 0.899 0.368898
## VERY_GOOD_HEALTH -5.240e-01 3.650e-01 -1.436 0.151184
## FAIR_HEALTH -1.177e+00 4.796e-01 -2.454 0.014173 *
## POOR_HEALTH -4.982e+00 1.116e+00 -4.465 8.19e-06 ***
## WORKING 4.244e+00 6.186e-01 6.861 7.61e-12 ***
## REALLY_LIKE_WORKING 6.015e-01 3.473e-01 1.732 0.083372 .
## DISLIKE_WORKING 9.691e-03 5.567e-01 0.017 0.986111
## REALLY_DISLIKE_WORKING 5.551e-01 1.128e+00 0.492 0.622557
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 11.11 on 5415 degrees of freedom
## Multiple R-squared: 0.3597, Adjusted R-squared: 0.3562
## F-statistic: 101.4 on 30 and 5415 DF, p-value: < 2.2e-16# summary(lm.step.both)GLM - Linear Entering Variables
probit.model <- glm(ZERO_WWH ~ .,
family = binomial(link = "probit"),
data = df_any)## Warning: glm.fit: algorithm did not converge## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurredlogit.model <- glm(ZERO_WWH ~ .,
family = "binomial",
data = df_any)## Warning: glm.fit: algorithm did not converge
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurredlm.nonzero <- lm(WEEKLY_WORK_HOURS ~ .,
data = df_nonzero)
log.nonzero <- lm(log(WEEKLY_WORK_HOURS) ~ .,
data = df_nonzero)
gamma.nonzero <- glm(WEEKLY_WORK_HOURS ~ .,
family = Gamma(link = "log"),
data = df_nonzero)
poisson.model <- glm(WEEKLY_WORK_HOURS ~ .,
family = "poisson",
data = df)
poisson.nonzero <- glm(WEEKLY_WORK_HOURS ~ .,
family = "poisson",
data = df_nonzero)
nb2.model <- glm.nb(WEEKLY_WORK_HOURS ~ .,
data = df)
nb2.nonzero <- glm.nb(WEEKLY_WORK_HOURS ~ .,
data = df_nonzero)
aic.probit <- AIC(probit.model)
aic.logit <- AIC(logit.model)
aic.lm.nonzero <- AIC(lm.nonzero)
aic.log.nonzero <- AIC(log.nonzero)
aic.gamma.nonzero <- AIC(gamma.nonzero)
aic.poisson.nonzero <- AIC(poisson.nonzero)
aic.nb2.nonzero <- AIC(nb2.nonzero)
aic.lm <- AIC(lm.model)
aic.lm.step <- AIC(lm.step.both)
aic.poisson <- AIC(poisson.model)
aic.nb2 <- AIC(nb2.model)GLM - Non-Linear Entery
lm.interaction <- glm(WEEKLY_WORK_HOURS~ 1 + HOSPITAL_EXPENSES + MEDICAID + VOLUNTEER
+ MALE + WORKING_SPOUSE:ANY_DEPENDENTS + WEEKS_PAID_VACATION:AGE
+ WEEKS_PAID_VACATION:HEART_CONDITION + HOSPITAL_EXPENSES:AGE
+ REDUCE_PAID_WORK_HOURS:WEEKS_PAID_VACATION + MEDICARE:AGE
+ MEDICARE:HOSPITAL_EXPENSES + YEARS_EDUCATED+ HOME_VALUE
+ MEDICAID:AGE + MEDICAID:ANY_DEPENDENTS + MEDICAID:WORKING_SPOUSE
+ MEDICAID:HOSPITAL_EXPENSES + MEDICAID:MEDICARE
+ RETIRED:HEART_CONDITION
+ RETIRED:WEEKS_PAID_VACATION + RETIRED:MEDICAID
+ VOLUNTEER:RETIRED + MALE:HEART_CONDITION
+ MALE:WEEKS_PAID_VACATION + MALE:RETIRED
+ YEARS_EDUCATED:AGE + YEARS_EDUCATED:REDUCE_PAID_WORK_HOURS
+ YEARS_EDUCATED:MEDICARE + YEARS_EDUCATED:RETIRED
+ AGE + HEART_CONDITION + ANY_DEPENDENTS + WORK_LIMITING_CONDITION
+ DEBTS + MISSING_WEEKS_PAID_VACATION
+ MISSING_REDUCE_PAID_WORK_HOURS
+ MISSING_HOME_VALUE + ACTIVE_ONCE_WEEKLY
+ ACTIVE_DAILY + NOT_ACTIVE + WORKING + REALLY_DISLIKE_WORKING
+ DISLIKE_WORKING + REALLY_DISLIKE_WORKING + EXCELLENT_HEALTH
+ VERY_GOOD_HEALTH + FAIR_HEALTH + POOR_HEALTH + MISSING_DEBTS,
family = gaussian(link = "identity"),
data = df_nonzero)
poisson.interaction <- glm(WEEKLY_WORK_HOURS~ 1 + HOSPITAL_EXPENSES + MEDICAID + VOLUNTEER
+ MALE + WORKING_SPOUSE:ANY_DEPENDENTS + WEEKS_PAID_VACATION:AGE
+ WEEKS_PAID_VACATION:HEART_CONDITION + HOSPITAL_EXPENSES:AGE
+ REDUCE_PAID_WORK_HOURS:WEEKS_PAID_VACATION + MEDICARE:AGE
+ MEDICARE:HOSPITAL_EXPENSES + YEARS_EDUCATED+ HOME_VALUE
+ MEDICAID:AGE + MEDICAID:ANY_DEPENDENTS + MEDICAID:WORKING_SPOUSE
+ MEDICAID:HOSPITAL_EXPENSES + MEDICAID:MEDICARE
+ RETIRED:HEART_CONDITION
+ RETIRED:WEEKS_PAID_VACATION + RETIRED:MEDICAID
+ VOLUNTEER:RETIRED + MALE:HEART_CONDITION
+ MALE:WEEKS_PAID_VACATION + MALE:RETIRED
+ YEARS_EDUCATED:AGE + YEARS_EDUCATED:REDUCE_PAID_WORK_HOURS
+ YEARS_EDUCATED:MEDICARE + YEARS_EDUCATED:RETIRED
+ AGE + HEART_CONDITION + ANY_DEPENDENTS + WORK_LIMITING_CONDITION
+ DEBTS + MISSING_WEEKS_PAID_VACATION
+ MISSING_REDUCE_PAID_WORK_HOURS
+ MISSING_HOME_VALUE + ACTIVE_ONCE_WEEKLY
+ ACTIVE_DAILY + NOT_ACTIVE + WORKING + REALLY_DISLIKE_WORKING
+ DISLIKE_WORKING + REALLY_DISLIKE_WORKING + EXCELLENT_HEALTH
+ VERY_GOOD_HEALTH + FAIR_HEALTH + POOR_HEALTH + MISSING_DEBTS,
family = "poisson",
data = df_nonzero)
gamma.interaction <- glm(WEEKLY_WORK_HOURS~ 1 + HOSPITAL_EXPENSES + MEDICAID + VOLUNTEER
+ MALE + WORKING_SPOUSE:ANY_DEPENDENTS + WEEKS_PAID_VACATION:AGE
+ WEEKS_PAID_VACATION:HEART_CONDITION + HOSPITAL_EXPENSES:AGE
+ REDUCE_PAID_WORK_HOURS:WEEKS_PAID_VACATION + MEDICARE:AGE
+ MEDICARE:HOSPITAL_EXPENSES + YEARS_EDUCATED + HOME_VALUE
+ MEDICAID:AGE + MEDICAID:ANY_DEPENDENTS + MEDICAID:WORKING_SPOUSE
+ MEDICAID:HOSPITAL_EXPENSES + MEDICAID:MEDICARE
+ RETIRED:HEART_CONDITION
+ RETIRED:WEEKS_PAID_VACATION + RETIRED:MEDICAID
+ VOLUNTEER:RETIRED + MALE:HEART_CONDITION
+ MALE:WEEKS_PAID_VACATION
+ YEARS_EDUCATED:AGE + YEARS_EDUCATED:REDUCE_PAID_WORK_HOURS
+ YEARS_EDUCATED:MEDICARE + YEARS_EDUCATED:RETIRED
+ AGE + HEART_CONDITION + ANY_DEPENDENTS + WORK_LIMITING_CONDITION
+ DEBTS + MISSING_WEEKS_PAID_VACATION
+ MISSING_REDUCE_PAID_WORK_HOURS
+ MISSING_HOME_VALUE + ACTIVE_ONCE_WEEKLY
+ ACTIVE_DAILY + NOT_ACTIVE + WORKING + REALLY_DISLIKE_WORKING
+ DISLIKE_WORKING + REALLY_DISLIKE_WORKING + EXCELLENT_HEALTH
+ VERY_GOOD_HEALTH + FAIR_HEALTH + POOR_HEALTH + MISSING_DEBTS,
family = Gamma(link = "log"),
data = df_nonzero)
nb2.interaction <- glm.nb(WEEKLY_WORK_HOURS~ 1 + HOSPITAL_EXPENSES + MEDICAID + VOLUNTEER
+ MALE + WORKING_SPOUSE:ANY_DEPENDENTS + WEEKS_PAID_VACATION:AGE
+ WEEKS_PAID_VACATION:HEART_CONDITION + HOSPITAL_EXPENSES:AGE
+ REDUCE_PAID_WORK_HOURS:WEEKS_PAID_VACATION + MEDICARE:AGE
+ MEDICARE:HOSPITAL_EXPENSES + YEARS_EDUCATED + HOME_VALUE
+ MEDICAID:AGE + MEDICAID:ANY_DEPENDENTS + MEDICAID:WORKING_SPOUSE
+ MEDICAID:HOSPITAL_EXPENSES + MEDICAID:MEDICARE
+ RETIRED:HEART_CONDITION
+ RETIRED:WEEKS_PAID_VACATION + RETIRED:MEDICAID
+ VOLUNTEER:RETIRED + MALE:HEART_CONDITION
+ MALE:WEEKS_PAID_VACATION + MALE:RETIRED
+ YEARS_EDUCATED:AGE + YEARS_EDUCATED:REDUCE_PAID_WORK_HOURS
+ YEARS_EDUCATED:MEDICARE + YEARS_EDUCATED:RETIRED
+ AGE + HEART_CONDITION + ANY_DEPENDENTS + WORK_LIMITING_CONDITION
+ DEBTS + MISSING_WEEKS_PAID_VACATION
+ MISSING_REDUCE_PAID_WORK_HOURS
+ MISSING_HOME_VALUE + ACTIVE_ONCE_WEEKLY
+ ACTIVE_DAILY + NOT_ACTIVE + WORKING + REALLY_DISLIKE_WORKING
+ DISLIKE_WORKING + REALLY_DISLIKE_WORKING + EXCELLENT_HEALTH
+ VERY_GOOD_HEALTH + FAIR_HEALTH + POOR_HEALTH + MISSING_DEBTS,
data = df_nonzero)
plot(lm.interaction)
plot(poisson.interaction)
plot(gamma.interaction)
plot(nb2.interaction)
m_list_all_no_interaction <- list(OLS = lm.nonzero,
Stepwise_OLS = lm.step.both,
Poisson = poisson.nonzero,
NB2 = nb2.nonzero,
Gamma_Nonzero = gamma.nonzero,
Log_Nonzero = log.nonzero)
cat("\nSummary of All Regression Non-interacted Results: ")##
## Summary of All Regression Non-interacted Results:msummary(m_list_all_no_interaction)## Warning in predict.lm(object, newdata, se.fit, scale = 1, type = if (type == :
## prediction from a rank-deficient fit may be misleading
## Warning in predict.lm(object, newdata, se.fit, scale = 1, type = if (type == :
## prediction from a rank-deficient fit may be misleading
## Warning in predict.lm(object, newdata, se.fit, scale = 1, type = if (type == :
## prediction from a rank-deficient fit may be misleading| OLS | Stepwise_OLS | Poisson | NB2 | Gamma_Nonzero | Log_Nonzero | |
|---|---|---|---|---|---|---|
| (Intercept) | 44.502 | 44.321 | 3.775 | 3.833 | 3.859 | 3.809 |
| (2.284) | (2.042) | (0.035) | (0.072) | (0.074) | (0.093) | |
| AGE | −0.196 | −0.193 | −0.006 | −0.006 | −0.007 | −0.009 |
| (0.034) | (0.033) | (0.0005) | (0.001) | (0.001) | (0.001) | |
| HEART_CONDITION | 0.142 | 0.005 | 0.0005 | −0.001 | 0.010 | |
| (0.473) | (0.007) | (0.015) | (0.015) | (0.019) | ||
| ANY_DEPENDENTS | 0.066 | −0.0002 | 0.0001 | 0.00002 | −0.011 | |
| (0.398) | (0.006) | (0.012) | (0.013) | (0.016) | ||
| WORKING_SPOUSE | −0.158 | −0.006 | −0.005 | −0.005 | 0.00004 | |
| (0.354) | (0.005) | (0.011) | (0.011) | (0.014) | ||
| WEEKS_PAID_VACATION | 0.807 | 0.810 | 0.020 | 0.024 | 0.026 | 0.031 |
| (0.081) | (0.080) | (0.001) | (0.002) | (0.003) | (0.003) | |
| REDUCE_PAID_WORK_HOURS | −0.914 | −0.867 | −0.025 | −0.030 | −0.032 | −0.024 |
| (0.366) | (0.360) | (0.005) | (0.011) | (0.012) | (0.015) | |
| MEDICARE | −0.960 | −0.891 | −0.031 | −0.031 | −0.030 | −0.025 |
| (0.562) | (0.557) | (0.009) | (0.018) | (0.018) | (0.023) | |
| MEDICAID | −3.848 | −3.806 | −0.113 | −0.098 | −0.090 | −0.120 |
| (0.552) | (0.544) | (0.009) | (0.017) | (0.018) | (0.022) | |
| HOSPITAL_EXPENSES | 0.323 | 0.319 | 0.008 | 0.009 | 0.009 | 0.008 |
| (0.128) | (0.127) | (0.002) | (0.004) | (0.004) | (0.005) | |
| RETIRED | −13.101 | −13.183 | −0.432 | −0.425 | −0.424 | −0.520 |
| (0.492) | (0.491) | (0.008) | (0.016) | (0.016) | (0.020) | |
| VOLUNTEER | −0.443 | −0.013 | −0.017 | −0.018 | −0.022 | |
| (0.320) | (0.005) | (0.010) | (0.010) | (0.013) | ||
| HOME_VALUE | −0.0002 | −0.000005 | −0.000008 | −0.000009 | −0.00002 | |
| (0.0002) | (0.000004) | (0.000008) | (0.000008) | (0.00001) | ||
| MALE | 5.123 | 5.113 | 0.138 | 0.144 | 0.146 | 0.174 |
| (0.311) | (0.306) | (0.005) | (0.010) | (0.010) | (0.013) | |
| YEARS_EDUCATED | 0.019 | 0.0005 | −0.001 | −0.002 | −0.003 | |
| (0.057) | (0.0008) | (0.002) | (0.002) | (0.002) | ||
| WORK_LIMITING_CONDITION | −1.265 | −1.298 | −0.039 | −0.047 | −0.051 | −0.048 |
| (0.488) | (0.482) | (0.008) | (0.015) | (0.016) | (0.020) | |
| DEBTS | 0.0008 | 0.00002 | 0.00001 | 0.00001 | 0.00003 | |
| (0.0007) | (0.00001) | (0.00002) | (0.00002) | (0.00003) | ||
| MISSING_WEEKS_PAID_VACATION | 2.126 | 1.244 | 0.059 | 0.069 | 0.074 | 0.085 |
| (0.896) | (0.438) | (0.013) | (0.028) | (0.029) | (0.036) | |
| MISSING_REDUCE_PAID_WORK_HOURS | −1.039 | −0.023 | −0.028 | −0.030 | −0.096 | |
| (0.911) | (0.014) | (0.028) | (0.030) | (0.037) | ||
| MISSING_HOME_VALUE | −0.218 | −0.005 | −0.009 | −0.010 | −0.019 | |
| (0.346) | (0.005) | (0.011) | (0.011) | (0.014) | ||
| ACTIVE_ONCE_WEEKLY | 0.768 | 0.755 | 0.021 | 0.022 | 0.022 | 0.018 |
| (0.427) | (0.424) | (0.006) | (0.013) | (0.014) | (0.017) | |
| ACTIVE_DAILY | 1.086 | 1.156 | 0.030 | 0.031 | 0.031 | 0.017 |
| (0.484) | (0.483) | (0.007) | (0.015) | (0.016) | (0.020) | |
| NOT_ACTIVE | 1.290 | 1.335 | 0.036 | 0.034 | 0.033 | 0.027 |
| (0.393) | (0.388) | (0.006) | (0.012) | (0.013) | (0.016) | |
| EXCELLENT_HEALTH | 0.522 | 0.013 | 0.014 | 0.014 | 0.009 | |
| (0.531) | (0.008) | (0.016) | (0.017) | (0.022) | ||
| VERY_GOOD_HEALTH | −0.494 | −0.674 | −0.014 | −0.016 | −0.017 | −0.017 |
| (0.365) | (0.336) | (0.005) | (0.011) | (0.012) | (0.015) | |
| FAIR_HEALTH | −1.175 | −1.229 | −0.032 | −0.033 | −0.034 | −0.035 |
| (0.479) | (0.458) | (0.007) | (0.015) | (0.016) | (0.019) | |
| POOR_HEALTH | −4.995 | −5.039 | −0.151 | −0.156 | −0.158 | −0.205 |
| (1.114) | (1.107) | (0.018) | (0.036) | (0.036) | (0.045) | |
| WORKING | 4.233 | 4.279 | 0.170 | 0.169 | 0.168 | 0.302 |
| (0.618) | (0.617) | (0.011) | (0.020) | (0.020) | (0.025) | |
| REALLY_LIKE_WORKING | 0.612 | 0.573 | 0.017 | 0.016 | 0.015 | 0.019 |
| (0.347) | (0.338) | (0.005) | (0.011) | (0.011) | (0.014) | |
| DISLIKE_WORKING | −0.004 | 0.0002 | −0.007 | −0.010 | −0.00009 | |
| (0.556) | (0.008) | (0.017) | (0.018) | (0.023) | ||
| REALLY_DISLIKE_WORKING | 0.541 | 0.014 | 0.018 | 0.020 | 0.042 | |
| (1.126) | (0.016) | (0.035) | (0.037) | (0.046) | ||
| Num.Obs. | 5443 | 5446 | 5443 | 5443 | 5443 | 5443 |
| R2 | 0.360 | 0.359 | 0.374 | |||
| R2 Adj. | 0.356 | 0.357 | 0.371 | |||
| AIC | 41675.1 | 41698.5 | 50384.3 | 43104.3 | 43850.3 | 44936.1 |
| BIC | 41886.4 | 41830.5 | 50588.9 | 43315.6 | 44061.6 | 45147.4 |
| Log.Lik. | −20805.559 | −20829.235 | −25161.139 | −21520.145 | −21893.173 | −3378.778 |
| F | 168.791 | |||||
| RMSE | 11.06 | 11.09 | 11.11 | 11.14 | 11.16 | 0.45 |
Transformations
lm.interaction.t <- glm(WEEKLY_WORK_HOURS~ 1 + HOSPITAL_EXPENSES + MEDICAID + VOLUNTEER
+ MALE + WORKING_SPOUSE:ANY_DEPENDENTS + WEEKS_PAID_VACATION:AGE
+ WEEKS_PAID_VACATION:HEART_CONDITION + HOSPITAL_EXPENSES:AGE
+ REDUCE_PAID_WORK_HOURS:WEEKS_PAID_VACATION + MEDICARE:AGE
+ MEDICARE:HOSPITAL_EXPENSES + YEARS_EDUCATED + HOME_VALUE
+ MEDICAID:AGE + MEDICAID:ANY_DEPENDENTS + MEDICAID:WORKING_SPOUSE
+ MEDICAID:HOSPITAL_EXPENSES + MEDICAID:MEDICARE
+ RETIRED:HEART_CONDITION
+ RETIRED:WEEKS_PAID_VACATION + RETIRED:MEDICAID
+ VOLUNTEER:RETIRED + MALE:HEART_CONDITION
+ MALE:WEEKS_PAID_VACATION + MALE:RETIRED
+ YEARS_EDUCATED:AGE + YEARS_EDUCATED:REDUCE_PAID_WORK_HOURS
+ YEARS_EDUCATED:MEDICARE + YEARS_EDUCATED:RETIRED
+ AGE + HEART_CONDITION + ANY_DEPENDENTS + WORK_LIMITING_CONDITION
+ DEBTS + MISSING_WEEKS_PAID_VACATION
+ MISSING_REDUCE_PAID_WORK_HOURS
+ MISSING_HOME_VALUE + ACTIVE_ONCE_WEEKLY
+ ACTIVE_DAILY + NOT_ACTIVE + WORKING + REALLY_DISLIKE_WORKING
+ DISLIKE_WORKING + REALLY_DISLIKE_WORKING + EXCELLENT_HEALTH
+ VERY_GOOD_HEALTH + FAIR_HEALTH + POOR_HEALTH + MISSING_DEBTS,
family = gaussian(link = "identity"),
data = df_transformed)
summary(lm.interaction.t)##
## Call:
## glm(formula = WEEKLY_WORK_HOURS ~ 1 + HOSPITAL_EXPENSES + MEDICAID +
## VOLUNTEER + MALE + WORKING_SPOUSE:ANY_DEPENDENTS + WEEKS_PAID_VACATION:AGE +
## WEEKS_PAID_VACATION:HEART_CONDITION + HOSPITAL_EXPENSES:AGE +
## REDUCE_PAID_WORK_HOURS:WEEKS_PAID_VACATION + MEDICARE:AGE +
## MEDICARE:HOSPITAL_EXPENSES + YEARS_EDUCATED + HOME_VALUE +
## MEDICAID:AGE + MEDICAID:ANY_DEPENDENTS + MEDICAID:WORKING_SPOUSE +
## MEDICAID:HOSPITAL_EXPENSES + MEDICAID:MEDICARE + RETIRED:HEART_CONDITION +
## RETIRED:WEEKS_PAID_VACATION + RETIRED:MEDICAID + VOLUNTEER:RETIRED +
## MALE:HEART_CONDITION + MALE:WEEKS_PAID_VACATION + MALE:RETIRED +
## YEARS_EDUCATED:AGE + YEARS_EDUCATED:REDUCE_PAID_WORK_HOURS +
## YEARS_EDUCATED:MEDICARE + YEARS_EDUCATED:RETIRED + AGE +
## HEART_CONDITION + ANY_DEPENDENTS + WORK_LIMITING_CONDITION +
## DEBTS + MISSING_WEEKS_PAID_VACATION + MISSING_REDUCE_PAID_WORK_HOURS +
## MISSING_HOME_VALUE + ACTIVE_ONCE_WEEKLY + ACTIVE_DAILY +
## NOT_ACTIVE + WORKING + REALLY_DISLIKE_WORKING + DISLIKE_WORKING +
## REALLY_DISLIKE_WORKING + EXCELLENT_HEALTH + VERY_GOOD_HEALTH +
## FAIR_HEALTH + POOR_HEALTH + MISSING_DEBTS, family = gaussian(link = "identity"),
## data = df_transformed)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -41.717 -5.114 -0.603 4.915 62.525
##
## Coefficients: (1 not defined because of singularities)
## Estimate Std. Error t value
## (Intercept) 29.541803 8.837878 3.343
## HOSPITAL_EXPENSES -3.560136 5.338833 -0.667
## MEDICAID -30.730646 6.289749 -4.886
## VOLUNTEER -0.090220 0.353599 -0.255
## MALE 6.543001 0.457769 14.293
## YEARS_EDUCATED 1.239555 0.596655 2.078
## HOME_VALUE 0.027321 0.073393 0.372
## AGE 0.037914 0.149223 0.254
## HEART_CONDITION 1.416079 0.869506 1.629
## ANY_DEPENDENTS 0.824483 0.586033 1.407
## WORK_LIMITING_CONDITION -1.410264 0.481040 -2.932
## DEBTS 0.167709 0.119039 1.409
## MISSING_WEEKS_PAID_VACATION 2.023545 0.882594 2.293
## MISSING_REDUCE_PAID_WORK_HOURS -1.185904 0.900860 -1.316
## MISSING_HOME_VALUE 0.156131 0.430461 0.363
## ACTIVE_ONCE_WEEKLY 0.744454 0.420497 1.770
## ACTIVE_DAILY 1.207348 0.475273 2.540
## NOT_ACTIVE 1.231421 0.387438 3.178
## WORKING 3.388880 0.614446 5.515
## REALLY_DISLIKE_WORKING 0.134360 1.103609 0.122
## DISLIKE_WORKING -0.228408 0.538312 -0.424
## EXCELLENT_HEALTH 0.455243 0.521739 0.873
## VERY_GOOD_HEALTH -0.546674 0.358626 -1.524
## FAIR_HEALTH -1.345974 0.471171 -2.857
## POOR_HEALTH -5.153936 1.099733 -4.687
## MISSING_DEBTS NA NA NA
## WORKING_SPOUSE:ANY_DEPENDENTS -2.058374 0.727436 -2.830
## WEEKS_PAID_VACATION:AGE 0.015418 0.001882 8.193
## WEEKS_PAID_VACATION:HEART_CONDITION -0.578665 0.231265 -2.502
## HOSPITAL_EXPENSES:AGE 0.075943 0.091353 0.831
## WEEKS_PAID_VACATION:REDUCE_PAID_WORK_HOURS 0.420521 0.174370 2.412
## AGE:MEDICARE -0.105619 0.033175 -3.184
## HOSPITAL_EXPENSES:MEDICARE -2.248795 1.464522 -1.536
## MEDICAID:AGE 0.414165 0.107781 3.843
## MEDICAID:ANY_DEPENDENTS 3.295654 1.344084 2.452
## MEDICAID:WORKING_SPOUSE 2.746723 1.149801 2.389
## HOSPITAL_EXPENSES:MEDICAID 5.230765 1.746514 2.995
## MEDICAID:MEDICARE -3.541750 1.623118 -2.182
## HEART_CONDITION:RETIRED -2.225802 1.043469 -2.133
## WEEKS_PAID_VACATION:RETIRED 1.560236 0.276503 5.643
## MEDICAID:RETIRED 5.287868 1.359699 3.889
## VOLUNTEER:RETIRED -1.392509 0.758679 -1.835
## MALE:HEART_CONDITION 1.103158 0.907525 1.216
## MALE:WEEKS_PAID_VACATION -0.791534 0.149180 -5.306
## MALE:RETIRED -2.195403 0.770872 -2.848
## AGE:YEARS_EDUCATED -0.018590 0.010113 -1.838
## REDUCE_PAID_WORK_HOURS:YEARS_EDUCATED -0.117330 0.034758 -3.376
## MEDICARE:YEARS_EDUCATED 0.488794 0.155469 3.144
## YEARS_EDUCATED:RETIRED -0.936117 0.053601 -17.464
## Pr(>|t|)
## (Intercept) 0.000836 ***
## HOSPITAL_EXPENSES 0.504904
## MEDICAID 1.06e-06 ***
## VOLUNTEER 0.798618
## MALE < 2e-16 ***
## YEARS_EDUCATED 0.037802 *
## HOME_VALUE 0.709716
## AGE 0.799445
## HEART_CONDITION 0.103456
## ANY_DEPENDENTS 0.159517
## WORK_LIMITING_CONDITION 0.003385 **
## DEBTS 0.158936
## MISSING_WEEKS_PAID_VACATION 0.021902 *
## MISSING_REDUCE_PAID_WORK_HOURS 0.188091
## MISSING_HOME_VALUE 0.716839
## ACTIVE_ONCE_WEEKLY 0.076715 .
## ACTIVE_DAILY 0.011103 *
## NOT_ACTIVE 0.001489 **
## WORKING 3.64e-08 ***
## REALLY_DISLIKE_WORKING 0.903105
## DISLIKE_WORKING 0.671361
## EXCELLENT_HEALTH 0.382948
## VERY_GOOD_HEALTH 0.127478
## FAIR_HEALTH 0.004298 **
## POOR_HEALTH 2.85e-06 ***
## MISSING_DEBTS NA
## WORKING_SPOUSE:ANY_DEPENDENTS 0.004677 **
## WEEKS_PAID_VACATION:AGE 3.17e-16 ***
## WEEKS_PAID_VACATION:HEART_CONDITION 0.012373 *
## HOSPITAL_EXPENSES:AGE 0.405834
## WEEKS_PAID_VACATION:REDUCE_PAID_WORK_HOURS 0.015913 *
## AGE:MEDICARE 0.001462 **
## HOSPITAL_EXPENSES:MEDICARE 0.124716
## MEDICAID:AGE 0.000123 ***
## MEDICAID:ANY_DEPENDENTS 0.014239 *
## MEDICAID:WORKING_SPOUSE 0.016934 *
## HOSPITAL_EXPENSES:MEDICAID 0.002757 **
## MEDICAID:MEDICARE 0.029148 *
## HEART_CONDITION:RETIRED 0.032963 *
## WEEKS_PAID_VACATION:RETIRED 1.76e-08 ***
## MEDICAID:RETIRED 0.000102 ***
## VOLUNTEER:RETIRED 0.066496 .
## MALE:HEART_CONDITION 0.224203
## MALE:WEEKS_PAID_VACATION 1.17e-07 ***
## MALE:RETIRED 0.004417 **
## AGE:YEARS_EDUCATED 0.066086 .
## REDUCE_PAID_WORK_HOURS:YEARS_EDUCATED 0.000742 ***
## MEDICARE:YEARS_EDUCATED 0.001676 **
## YEARS_EDUCATED:RETIRED < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 118.8153)
##
## Null deviance: 1044148 on 5445 degrees of freedom
## Residual deviance: 641365 on 5398 degrees of freedom
## AIC: 41524
##
## Number of Fisher Scoring iterations: 2plot(lm.interaction.t)
Endogenous Variables
library(lares)##
## Attaching package: 'lares'## The following objects are masked from 'package:performance':
##
## mae, mse, rmse## The following objects are masked from 'package:tictoc':
##
## tic, tocdf_corr <- subset(df_nonzero, select = -c(MISSING_WEEKS_PAID_VACATION,
MISSING_REDUCE_PAID_WORK_HOURS,
MISSING_HOME_VALUE,
MISSING_DEBTS,
EXCELLENT_HEALTH,
VERY_GOOD_HEALTH,
FAIR_HEALTH,
POOR_HEALTH,
WEEKLY_WORK_HOURS))
corr_cross(df_corr, # name of dataset
max_pvalue = 0.05, # display only significant correlations (at 5% level)
top = 10 # display top 10 couples of variables (by correlation coefficient)
)## Returning only the top 10. You may override with the 'top' argument## Warning in .font_global(font, quiet = FALSE): Font 'Arial Narrow' is not
## installed, has other name, or can't be found
IV Testing
# lm.age.test <- glm(WEEKLY_WORK_HOURS ~ AGE + ANY_DEPENDENTS
# + WEEKS_PAID_VACATION + REDUCE_PAID_WORK_HOURS + MEDICAID + HOSPITAL_EXPENSES
# + VOLUNTEER + HOME_VALUE + MALE + YEARS_EDUCATED + WORK_LIMITING_CONDITION
# + DEBTS + MISSING_WEEKS_PAID_VACATION + MISSING_REDUCE_PAID_WORK_HOURS
# + MISSING_HOME_VALUE
# + ACTIVE_ONCE_WEEKLY + ACTIVE_DAILY + NOT_ACTIVE + WORKING
# + REALLY_DISLIKE_WORKING + DISLIKE_WORKING + REALLY_DISLIKE_WORKING
# + EXCELLENT_HEALTH
# + VERY_GOOD_HEALTH + FAIR_HEALTH + POOR_HEALTH,
# family = gaussian(link = "identity"),
# data = df_nonzero)#Removed medicare, retired, working_spouse, heart_condition
# summary(lm.age.test) #Age significant, cannot be IV
#
# lm.medicare.test <- glm(WEEKLY_WORK_HOURS ~ MEDICARE +
# + WEEKS_PAID_VACATION + REDUCE_PAID_WORK_HOURS + MEDICAID + HOSPITAL_EXPENSES
# + VOLUNTEER + HOME_VALUE + MALE + YEARS_EDUCATED + WORK_LIMITING_CONDITION
# + DEBTS + MISSING_WEEKS_PAID_VACATION + MISSING_REDUCE_PAID_WORK_HOURS
# + MISSING_HOME_VALUE
# + ACTIVE_ONCE_WEEKLY + ACTIVE_DAILY + NOT_ACTIVE + WORKING + ANY_DEPENDENTS
# + REALLY_DISLIKE_WORKING + DISLIKE_WORKING + REALLY_DISLIKE_WORKING
# + EXCELLENT_HEALTH
# + VERY_GOOD_HEALTH + FAIR_HEALTH + POOR_HEALTH,
# family = gaussian(link = "identity"),
# data = df_nonzero)#Removed age, retired, working_spouse, heart_condition
# summary(lm.medicare.test) #Medicare significant
#
# lm.retired.test <- glm(WEEKLY_WORK_HOURS ~ RETIRED +
# + WEEKS_PAID_VACATION + REDUCE_PAID_WORK_HOURS + MEDICAID + HOSPITAL_EXPENSES
# + VOLUNTEER + HOME_VALUE + MALE + YEARS_EDUCATED + WORK_LIMITING_CONDITION
# + DEBTS + MISSING_WEEKS_PAID_VACATION + MISSING_REDUCE_PAID_WORK_HOURS
# + MISSING_HOME_VALUE
# + ACTIVE_ONCE_WEEKLY + ACTIVE_DAILY + NOT_ACTIVE + WORKING + ANY_DEPENDENTS
# + REALLY_DISLIKE_WORKING + DISLIKE_WORKING + REALLY_DISLIKE_WORKING
# + EXCELLENT_HEALTH
# + VERY_GOOD_HEALTH + FAIR_HEALTH + POOR_HEALTH,
# family = gaussian(link = "identity"),
# data = df_nonzero)#Removed age, medicare, working_spouse, heart_condition
# summary(lm.retired.test) #Retired significant
#
# lm.ws.test <- glm(WEEKLY_WORK_HOURS ~ WORKING_SPOUSE +
# + WEEKS_PAID_VACATION + REDUCE_PAID_WORK_HOURS + MEDICAID + HOSPITAL_EXPENSES
# + VOLUNTEER + HOME_VALUE + MALE + YEARS_EDUCATED + WORK_LIMITING_CONDITION
# + DEBTS + MISSING_WEEKS_PAID_VACATION + MISSING_REDUCE_PAID_WORK_HOURS
# + MISSING_HOME_VALUE
# + ACTIVE_ONCE_WEEKLY + ACTIVE_DAILY + NOT_ACTIVE + WORKING + ANY_DEPENDENTS
# + REALLY_DISLIKE_WORKING + DISLIKE_WORKING + REALLY_DISLIKE_WORKING
# + EXCELLENT_HEALTH
# + VERY_GOOD_HEALTH + FAIR_HEALTH + POOR_HEALTH,
# family = gaussian(link = "identity"),
# data = df_nonzero)#Removed age, medicare, retired, heart_condition
# summary(lm.ws.test) #Working spouse significant
#
# lm.hc.test <- glm(WEEKLY_WORK_HOURS ~ HEART_CONDITION +
# + WEEKS_PAID_VACATION + REDUCE_PAID_WORK_HOURS + MEDICAID + HOSPITAL_EXPENSES
# + VOLUNTEER + HOME_VALUE + MALE + YEARS_EDUCATED + WORK_LIMITING_CONDITION
# + DEBTS + MISSING_WEEKS_PAID_VACATION + MISSING_REDUCE_PAID_WORK_HOURS
# + MISSING_HOME_VALUE + EXCELLENT_HEALTH
# + ACTIVE_ONCE_WEEKLY + ACTIVE_DAILY + NOT_ACTIVE + WORKING + ANY_DEPENDENTS
# + REALLY_DISLIKE_WORKING + DISLIKE_WORKING + REALLY_DISLIKE_WORKING
# + VERY_GOOD_HEALTH + FAIR_HEALTH + POOR_HEALTH,
# family = gaussian(link = "identity"),
# data = df_nonzero)#Removed age, medicare, retired, heart_condition
# summary(lm.hc.test) #Heart condition significantCreating Endogenous Model
age.hat <- glm(AGE ~ MEDICARE + HEART_CONDITION + WORKING_SPOUSE
+ WEEKS_PAID_VACATION + REDUCE_PAID_WORK_HOURS + MEDICAID + HOSPITAL_EXPENSES
+ VOLUNTEER + HOME_VALUE + MALE + YEARS_EDUCATED + WORK_LIMITING_CONDITION
+ DEBTS + MISSING_WEEKS_PAID_VACATION + MISSING_REDUCE_PAID_WORK_HOURS
+ MISSING_HOME_VALUE
+ ACTIVE_ONCE_WEEKLY + ACTIVE_DAILY + NOT_ACTIVE + WORKING + ANY_DEPENDENTS
+ REALLY_DISLIKE_WORKING + DISLIKE_WORKING + REALLY_DISLIKE_WORKING
+ EXCELLENT_HEALTH
+ VERY_GOOD_HEALTH + FAIR_HEALTH + POOR_HEALTH,
family = gaussian(link = "identity"),
data = df_nonzero)
df_hat <- df_nonzero
df_hat$AGE <- df_hat$AGE + age.hat$residuals
df_hat <- subset(df_hat, select = -c(MEDICARE,
HEART_CONDITION,
WORKING_SPOUSE))
lm.endog <- glm(WEEKLY_WORK_HOURS ~ .,
family = gaussian(link = "identity"),
data = df_hat)
summary(lm.endog)##
## Call:
## glm(formula = WEEKLY_WORK_HOURS ~ ., family = gaussian(link = "identity"),
## data = df_hat)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -38.319 -5.409 -0.563 4.902 67.300
##
## Coefficients: (1 not defined because of singularities)
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.981e+01 1.424e+00 27.952 < 2e-16 ***
## AGE -1.251e-01 1.572e-02 -7.961 2.06e-15 ***
## ANY_DEPENDENTS 3.285e-01 3.949e-01 0.832 0.40561
## WEEKS_PAID_VACATION 8.254e-01 8.117e-02 10.170 < 2e-16 ***
## REDUCE_PAID_WORK_HOURS -1.026e+00 3.658e-01 -2.805 0.00505 **
## MEDICAID -3.693e+00 5.497e-01 -6.720 2.01e-11 ***
## HOSPITAL_EXPENSES 3.330e-01 1.275e-01 2.613 0.00900 **
## RETIRED -1.374e+01 4.713e-01 -29.141 < 2e-16 ***
## VOLUNTEER -4.587e-01 3.205e-01 -1.431 0.15253
## HOME_VALUE -2.378e-04 2.428e-04 -0.980 0.32728
## MALE 5.061e+00 3.074e-01 16.463 < 2e-16 ***
## YEARS_EDUCATED 1.313e-02 5.657e-02 0.232 0.81649
## WORK_LIMITING_CONDITION -1.359e+00 4.851e-01 -2.801 0.00512 **
## DEBTS 8.386e-04 7.331e-04 1.144 0.25269
## MISSING_WEEKS_PAID_VACATION 2.090e+00 8.969e-01 2.330 0.01982 *
## MISSING_REDUCE_PAID_WORK_HOURS -1.155e+00 9.124e-01 -1.266 0.20548
## MISSING_HOME_VALUE -1.155e-01 3.244e-01 -0.356 0.72170
## MISSING_DEBTS NA NA NA NA
## ACTIVE_ONCE_WEEKLY 7.393e-01 4.272e-01 1.730 0.08360 .
## ACTIVE_DAILY 1.073e+00 4.844e-01 2.215 0.02677 *
## NOT_ACTIVE 1.141e+00 3.928e-01 2.904 0.00370 **
## EXCELLENT_HEALTH 5.242e-01 5.306e-01 0.988 0.32317
## VERY_GOOD_HEALTH -5.061e-01 3.641e-01 -1.390 0.16460
## FAIR_HEALTH -1.109e+00 4.776e-01 -2.322 0.02025 *
## POOR_HEALTH -4.771e+00 1.114e+00 -4.281 1.89e-05 ***
## WORKING 4.617e+00 6.165e-01 7.490 8.01e-14 ***
## REALLY_LIKE_WORKING 5.666e-01 3.471e-01 1.632 0.10267
## DISLIKE_WORKING 1.111e-01 5.561e-01 0.200 0.84172
## REALLY_DISLIKE_WORKING 6.832e-01 1.127e+00 0.606 0.54446
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 123.4437)
##
## Null deviance: 1040019 on 5442 degrees of freedom
## Residual deviance: 668448 on 5415 degrees of freedom
## AIC: 41689
##
## Number of Fisher Scoring iterations: 2lm.endog.interactions <- glm(WEEKLY_WORK_HOURS ~ 1 + HOSPITAL_EXPENSES + MEDICAID + VOLUNTEER
+ MALE + WEEKS_PAID_VACATION:AGE + HOSPITAL_EXPENSES:AGE
+ REDUCE_PAID_WORK_HOURS:WEEKS_PAID_VACATION + YEARS_EDUCATED
+ HOME_VALUE
+ MEDICAID:AGE + MEDICAID:ANY_DEPENDENTS
+ MEDICAID:HOSPITAL_EXPENSES + RETIRED:WEEKS_PAID_VACATION
+ RETIRED:MEDICAID + VOLUNTEER:RETIRED + MALE:WEEKS_PAID_VACATION
+ MALE:RETIRED+ YEARS_EDUCATED:AGE
+ YEARS_EDUCATED:REDUCE_PAID_WORK_HOURS
+ YEARS_EDUCATED:RETIRED + AGE + ANY_DEPENDENTS
+ WORK_LIMITING_CONDITION
+ DEBTS + MISSING_WEEKS_PAID_VACATION
+ MISSING_REDUCE_PAID_WORK_HOURS
+ MISSING_HOME_VALUE + ACTIVE_ONCE_WEEKLY
+ ACTIVE_DAILY + NOT_ACTIVE + WORKING + REALLY_DISLIKE_WORKING
+ DISLIKE_WORKING + REALLY_DISLIKE_WORKING + EXCELLENT_HEALTH
+ VERY_GOOD_HEALTH + FAIR_HEALTH + POOR_HEALTH + MISSING_DEBTS,
family = gaussian(link = "identity"),
data = df_hat)
summary(lm.endog.interactions)##
## Call:
## glm(formula = WEEKLY_WORK_HOURS ~ 1 + HOSPITAL_EXPENSES + MEDICAID +
## VOLUNTEER + MALE + WEEKS_PAID_VACATION:AGE + HOSPITAL_EXPENSES:AGE +
## REDUCE_PAID_WORK_HOURS:WEEKS_PAID_VACATION + YEARS_EDUCATED +
## HOME_VALUE + MEDICAID:AGE + MEDICAID:ANY_DEPENDENTS + MEDICAID:HOSPITAL_EXPENSES +
## RETIRED:WEEKS_PAID_VACATION + RETIRED:MEDICAID + VOLUNTEER:RETIRED +
## MALE:WEEKS_PAID_VACATION + MALE:RETIRED + YEARS_EDUCATED:AGE +
## YEARS_EDUCATED:REDUCE_PAID_WORK_HOURS + YEARS_EDUCATED:RETIRED +
## AGE + ANY_DEPENDENTS + WORK_LIMITING_CONDITION + DEBTS +
## MISSING_WEEKS_PAID_VACATION + MISSING_REDUCE_PAID_WORK_HOURS +
## MISSING_HOME_VALUE + ACTIVE_ONCE_WEEKLY + ACTIVE_DAILY +
## NOT_ACTIVE + WORKING + REALLY_DISLIKE_WORKING + DISLIKE_WORKING +
## REALLY_DISLIKE_WORKING + EXCELLENT_HEALTH + VERY_GOOD_HEALTH +
## FAIR_HEALTH + POOR_HEALTH + MISSING_DEBTS, family = gaussian(link = "identity"),
## data = df_hat)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -41.645 -5.201 -0.654 4.938 61.784
##
## Coefficients: (1 not defined because of singularities)
## Estimate Std. Error t value
## (Intercept) 3.846e+01 4.411e+00 8.719
## HOSPITAL_EXPENSES -1.101e+00 7.817e-01 -1.409
## MEDICAID -1.904e+01 3.245e+00 -5.868
## VOLUNTEER -1.199e-01 3.541e-01 -0.339
## MALE 6.743e+00 4.458e-01 15.123
## YEARS_EDUCATED 3.439e-01 3.110e-01 1.106
## HOME_VALUE -1.639e-04 2.388e-04 -0.686
## AGE -1.379e-01 6.977e-02 -1.976
## ANY_DEPENDENTS -1.407e-01 4.088e-01 -0.344
## WORK_LIMITING_CONDITION -1.388e+00 4.774e-01 -2.908
## DEBTS 5.681e-04 7.235e-04 0.785
## MISSING_WEEKS_PAID_VACATION 1.761e+00 8.833e-01 1.993
## MISSING_REDUCE_PAID_WORK_HOURS -1.047e+00 9.004e-01 -1.163
## MISSING_HOME_VALUE -1.220e-01 3.194e-01 -0.382
## ACTIVE_ONCE_WEEKLY 7.849e-01 4.205e-01 1.867
## ACTIVE_DAILY 1.212e+00 4.764e-01 2.545
## NOT_ACTIVE 1.079e+00 3.866e-01 2.790
## WORKING 3.964e+00 6.114e-01 6.484
## REALLY_DISLIKE_WORKING 2.530e-01 1.105e+00 0.229
## DISLIKE_WORKING -1.015e-01 5.383e-01 -0.189
## EXCELLENT_HEALTH 4.661e-01 5.201e-01 0.896
## VERY_GOOD_HEALTH -5.506e-01 3.577e-01 -1.539
## FAIR_HEALTH -1.169e+00 4.700e-01 -2.487
## POOR_HEALTH -4.853e+00 1.099e+00 -4.417
## MISSING_DEBTS NA NA NA
## WEEKS_PAID_VACATION:AGE 1.487e-02 1.803e-03 8.243
## HOSPITAL_EXPENSES:AGE 2.233e-02 1.302e-02 1.715
## WEEKS_PAID_VACATION:REDUCE_PAID_WORK_HOURS 4.107e-01 1.737e-01 2.365
## MEDICAID:AGE 2.216e-01 5.377e-02 4.121
## MEDICAID:ANY_DEPENDENTS 3.729e+00 1.314e+00 2.838
## HOSPITAL_EXPENSES:MEDICAID 9.380e-01 3.931e-01 2.386
## WEEKS_PAID_VACATION:RETIRED 1.526e+00 2.768e-01 5.512
## MEDICAID:RETIRED 4.409e+00 1.269e+00 3.475
## VOLUNTEER:RETIRED -1.381e+00 7.606e-01 -1.816
## MALE:WEEKS_PAID_VACATION -8.403e-01 1.488e-01 -5.646
## MALE:RETIRED -2.438e+00 7.581e-01 -3.217
## AGE:YEARS_EDUCATED -1.970e-03 5.036e-03 -0.391
## REDUCE_PAID_WORK_HOURS:YEARS_EDUCATED -1.217e-01 3.469e-02 -3.507
## YEARS_EDUCATED:RETIRED -9.697e-01 5.110e-02 -18.978
## Pr(>|t|)
## (Intercept) < 2e-16 ***
## HOSPITAL_EXPENSES 0.159001
## MEDICAID 4.68e-09 ***
## VOLUNTEER 0.734990
## MALE < 2e-16 ***
## YEARS_EDUCATED 0.268932
## HOME_VALUE 0.492606
## AGE 0.048177 *
## ANY_DEPENDENTS 0.730685
## WORK_LIMITING_CONDITION 0.003649 **
## DEBTS 0.432355
## MISSING_WEEKS_PAID_VACATION 0.046288 *
## MISSING_REDUCE_PAID_WORK_HOURS 0.244876
## MISSING_HOME_VALUE 0.702491
## ACTIVE_ONCE_WEEKLY 0.062008 .
## ACTIVE_DAILY 0.010958 *
## NOT_ACTIVE 0.005289 **
## WORKING 9.74e-11 ***
## REALLY_DISLIKE_WORKING 0.818843
## DISLIKE_WORKING 0.850398
## EXCELLENT_HEALTH 0.370174
## VERY_GOOD_HEALTH 0.123871
## FAIR_HEALTH 0.012905 *
## POOR_HEALTH 1.02e-05 ***
## MISSING_DEBTS NA
## WEEKS_PAID_VACATION:AGE < 2e-16 ***
## HOSPITAL_EXPENSES:AGE 0.086467 .
## WEEKS_PAID_VACATION:REDUCE_PAID_WORK_HOURS 0.018070 *
## MEDICAID:AGE 3.82e-05 ***
## MEDICAID:ANY_DEPENDENTS 0.004555 **
## HOSPITAL_EXPENSES:MEDICAID 0.017062 *
## WEEKS_PAID_VACATION:RETIRED 3.72e-08 ***
## MEDICAID:RETIRED 0.000516 ***
## VOLUNTEER:RETIRED 0.069375 .
## MALE:WEEKS_PAID_VACATION 1.73e-08 ***
## MALE:RETIRED 0.001305 **
## AGE:YEARS_EDUCATED 0.695645
## REDUCE_PAID_WORK_HOURS:YEARS_EDUCATED 0.000456 ***
## YEARS_EDUCATED:RETIRED < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 119.3353)
##
## Null deviance: 1040019 on 5442 degrees of freedom
## Residual deviance: 645007 on 5405 degrees of freedom
## AIC: 41515
##
## Number of Fisher Scoring iterations: 2df_hat_transformed <- df_hat
df_hat_transformed$HOSPITAL_EXPENSES <- log(I(df_hat_transformed$HOSPITAL_EXPENSES + 1))
df_hat_transformed$DEBTS <- log(I(df_hat_transformed$DEBTS + 1))
df_hat_transformed$HOME_VALUE <- log(I(df_hat_transformed$HOME_VALUE + 1))
lm.endog.interactions.t <- glm(WEEKLY_WORK_HOURS ~ 1 + HOSPITAL_EXPENSES + MEDICAID + VOLUNTEER
+ MALE + WEEKS_PAID_VACATION:AGE + HOSPITAL_EXPENSES:AGE
+ REDUCE_PAID_WORK_HOURS:WEEKS_PAID_VACATION + YEARS_EDUCATED
+ HOME_VALUE
+ MEDICAID:AGE + MEDICAID:ANY_DEPENDENTS
+ MEDICAID:HOSPITAL_EXPENSES + RETIRED:WEEKS_PAID_VACATION
+ RETIRED:MEDICAID + VOLUNTEER:RETIRED + MALE:WEEKS_PAID_VACATION
+ MALE:RETIRED+ YEARS_EDUCATED:AGE
+ YEARS_EDUCATED:REDUCE_PAID_WORK_HOURS
+ YEARS_EDUCATED:RETIRED + AGE + ANY_DEPENDENTS
+ WORK_LIMITING_CONDITION
+ DEBTS + MISSING_WEEKS_PAID_VACATION
+ MISSING_REDUCE_PAID_WORK_HOURS
+ MISSING_HOME_VALUE + ACTIVE_ONCE_WEEKLY
+ ACTIVE_DAILY + NOT_ACTIVE + WORKING + REALLY_DISLIKE_WORKING
+ DISLIKE_WORKING + REALLY_DISLIKE_WORKING + EXCELLENT_HEALTH
+ VERY_GOOD_HEALTH + FAIR_HEALTH + POOR_HEALTH + MISSING_DEBTS,
family = gaussian(link = "identity"),
data = df_hat_transformed)
summary(lm.endog.interactions.t)##
## Call:
## glm(formula = WEEKLY_WORK_HOURS ~ 1 + HOSPITAL_EXPENSES + MEDICAID +
## VOLUNTEER + MALE + WEEKS_PAID_VACATION:AGE + HOSPITAL_EXPENSES:AGE +
## REDUCE_PAID_WORK_HOURS:WEEKS_PAID_VACATION + YEARS_EDUCATED +
## HOME_VALUE + MEDICAID:AGE + MEDICAID:ANY_DEPENDENTS + MEDICAID:HOSPITAL_EXPENSES +
## RETIRED:WEEKS_PAID_VACATION + RETIRED:MEDICAID + VOLUNTEER:RETIRED +
## MALE:WEEKS_PAID_VACATION + MALE:RETIRED + YEARS_EDUCATED:AGE +
## YEARS_EDUCATED:REDUCE_PAID_WORK_HOURS + YEARS_EDUCATED:RETIRED +
## AGE + ANY_DEPENDENTS + WORK_LIMITING_CONDITION + DEBTS +
## MISSING_WEEKS_PAID_VACATION + MISSING_REDUCE_PAID_WORK_HOURS +
## MISSING_HOME_VALUE + ACTIVE_ONCE_WEEKLY + ACTIVE_DAILY +
## NOT_ACTIVE + WORKING + REALLY_DISLIKE_WORKING + DISLIKE_WORKING +
## REALLY_DISLIKE_WORKING + EXCELLENT_HEALTH + VERY_GOOD_HEALTH +
## FAIR_HEALTH + POOR_HEALTH + MISSING_DEBTS, family = gaussian(link = "identity"),
## data = df_hat_transformed)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -41.700 -5.161 -0.652 4.946 61.596
##
## Coefficients: (1 not defined because of singularities)
## Estimate Std. Error t value
## (Intercept) 38.284832 4.414084 8.673
## HOSPITAL_EXPENSES -1.384162 2.751124 -0.503
## MEDICAID -18.654471 3.249404 -5.741
## VOLUNTEER -0.136441 0.354288 -0.385
## MALE 6.723348 0.446214 15.068
## YEARS_EDUCATED 0.325866 0.310782 1.049
## HOME_VALUE 0.021631 0.073188 0.296
## AGE -0.137825 0.069793 -1.975
## ANY_DEPENDENTS -0.151732 0.409185 -0.371
## WORK_LIMITING_CONDITION -1.430111 0.478175 -2.991
## DEBTS 0.171217 0.119116 1.437
## MISSING_WEEKS_PAID_VACATION 1.719272 0.882618 1.948
## MISSING_REDUCE_PAID_WORK_HOURS -1.031903 0.900712 -1.146
## MISSING_HOME_VALUE 0.168668 0.423649 0.398
## ACTIVE_ONCE_WEEKLY 0.802120 0.420860 1.906
## ACTIVE_DAILY 1.210597 0.476196 2.542
## NOT_ACTIVE 1.090748 0.386913 2.819
## WORKING 3.961806 0.611303 6.481
## REALLY_DISLIKE_WORKING 0.226865 1.105300 0.205
## DISLIKE_WORKING -0.133391 0.538553 -0.248
## EXCELLENT_HEALTH 0.504548 0.521392 0.968
## VERY_GOOD_HEALTH -0.520374 0.358520 -1.451
## FAIR_HEALTH -1.144212 0.470197 -2.433
## POOR_HEALTH -4.798499 1.098964 -4.366
## MISSING_DEBTS NA NA NA
## WEEKS_PAID_VACATION:AGE 0.014781 0.001804 8.191
## HOSPITAL_EXPENSES:AGE 0.033831 0.045341 0.746
## WEEKS_PAID_VACATION:REDUCE_PAID_WORK_HOURS 0.415226 0.173793 2.389
## MEDICAID:AGE 0.214975 0.053874 3.990
## MEDICAID:ANY_DEPENDENTS 3.670541 1.313923 2.794
## HOSPITAL_EXPENSES:MEDICAID 4.775081 1.730420 2.759
## WEEKS_PAID_VACATION:RETIRED 1.536728 0.276813 5.551
## MEDICAID:RETIRED 4.485591 1.268588 3.536
## VOLUNTEER:RETIRED -1.362506 0.760655 -1.791
## MALE:WEEKS_PAID_VACATION -0.833576 0.148855 -5.600
## MALE:RETIRED -2.418379 0.757621 -3.192
## AGE:YEARS_EDUCATED -0.001808 0.005030 -0.359
## REDUCE_PAID_WORK_HOURS:YEARS_EDUCATED -0.123118 0.034698 -3.548
## YEARS_EDUCATED:RETIRED -0.970823 0.051117 -18.992
## Pr(>|t|)
## (Intercept) < 2e-16 ***
## HOSPITAL_EXPENSES 0.614896
## MEDICAID 9.93e-09 ***
## VOLUNTEER 0.700168
## MALE < 2e-16 ***
## YEARS_EDUCATED 0.294440
## HOME_VALUE 0.767579
## AGE 0.048344 *
## ANY_DEPENDENTS 0.710789
## WORK_LIMITING_CONDITION 0.002795 **
## DEBTS 0.150661
## MISSING_WEEKS_PAID_VACATION 0.051476 .
## MISSING_REDUCE_PAID_WORK_HOURS 0.251989
## MISSING_HOME_VALUE 0.690549
## ACTIVE_ONCE_WEEKLY 0.056715 .
## ACTIVE_DAILY 0.011043 *
## NOT_ACTIVE 0.004833 **
## WORKING 9.93e-11 ***
## REALLY_DISLIKE_WORKING 0.837383
## DISLIKE_WORKING 0.804388
## EXCELLENT_HEALTH 0.333241
## VERY_GOOD_HEALTH 0.146712
## FAIR_HEALTH 0.014987 *
## POOR_HEALTH 1.29e-05 ***
## MISSING_DEBTS NA
## WEEKS_PAID_VACATION:AGE 3.20e-16 ***
## HOSPITAL_EXPENSES:AGE 0.455615
## WEEKS_PAID_VACATION:REDUCE_PAID_WORK_HOURS 0.016919 *
## MEDICAID:AGE 6.69e-05 ***
## MEDICAID:ANY_DEPENDENTS 0.005231 **
## HOSPITAL_EXPENSES:MEDICAID 0.005809 **
## WEEKS_PAID_VACATION:RETIRED 2.97e-08 ***
## MEDICAID:RETIRED 0.000410 ***
## VOLUNTEER:RETIRED 0.073313 .
## MALE:WEEKS_PAID_VACATION 2.25e-08 ***
## MALE:RETIRED 0.001421 **
## AGE:YEARS_EDUCATED 0.719332
## REDUCE_PAID_WORK_HOURS:YEARS_EDUCATED 0.000391 ***
## YEARS_EDUCATED:RETIRED < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 119.3468)
##
## Null deviance: 1040019 on 5442 degrees of freedom
## Residual deviance: 645070 on 5405 degrees of freedom
## AIC: 41515
##
## Number of Fisher Scoring iterations: 2Renewed Interactions
df_intr <- df_hat
names(df_intr)[1] = "WWH" # Weekly Work Hours
names(df_intr)[3] = "AD" # Any Dependents
names(df_intr)[4] = "WV" # Weeks of Paid Vacation
names(df_intr)[5] = "RH" # Reduce Paid Work Hours
names(df_intr)[6] = "Md" # Medicaid
names(df_intr)[7] = "HE" # Hospital Expenses
names(df_intr)[8] = "R" # Retired
names(df_intr)[9] = "V" # Volunteer
names(df_intr)[10] = "HV" # Home Value
names(df_intr)[11] = "Ma" # Male
names(df_intr)[12] = "YE" # Years of education
names(df_intr)[13] = "WL" # Work Limiting Condition
names(df_intr)[14] = "Db" # Debts
names(df_intr)[19] = "AW" # Active Once Weekly
names(df_intr)[20] = "AcD" # Active Daily
names(df_intr)[21] = "N" # Not Active
names(df_intr)[22] = "EH" # Excellent Health
names(df_intr)[23] = "VH" # Very Good Health
names(df_intr)[24] = "FH" # Fair Health
names(df_intr)[25] = "PH" # Poor Health
names(df_intr)[26] = "W" # Working
names(df_intr)[27] = "LW" # Really Likes Working
names(df_intr)[28] = "DW" # Dislike Working
names(df_intr)[29] = "RW" # Really Dislike Working
lm.all.interactions.hat <- glmulti(WWH ~ AGE + AD
+ WV + HE
+ Db + R + Ma
+ HV + YE
+ EH + VH + FH + PH,
data = df_intr,
crit = aicc,
level = 2,
method = "g", # otherwise choose "d", "g" for algorithmic, or "h" for brute force
family = gaussian,
fit = glm,
confsetsize = 100)## Initialization...
## TASK: Genetic algorithm in the candidate set.
## Initialization...
## Algorithm started...
##
## After 10 generations:
## Best model: WWH~1+WV+Db+R+Ma+YE+VH+FH+WV:AGE+WV:AD+HE:WV+Db:AD+R:AGE+R:WV+R:HE+Ma:AD+Ma:WV+Ma:HE+Ma:R+HV:AD+HV:WV+HV:HE+HV:Db+HV:Ma+YE:AGE+YE:AD+YE:WV+YE:Db+YE:R+YE:HV+EH:AGE+EH:AD+EH:WV+EH:Ma+EH:HV+EH:YE+VH:Db+VH:R+VH:Ma+VH:YE+VH:EH+FH:WV+FH:HE+FH:Db+FH:R+FH:Ma+FH:YE+PH:AGE+PH:AD+PH:WV+PH:R+PH:Ma+PH:EH+PH:FH
## Crit= 41708.3202570512
## Mean crit= 41950.2761145703
## Change in best IC: 31708.3202570512 / Change in mean IC: 31950.2761145703
##
## After 20 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:WV+Db:WV+R:AGE+R:WV+R:HE+Ma:WV+Ma:HE+Ma:Db+Ma:R+HV:WV+YE:AGE+YE:AD+YE:WV+YE:Db+YE:R+EH:AGE+EH:AD+EH:HE+VH:AGE+VH:HE+VH:Db+VH:Ma+FH:AGE+FH:AD+FH:WV+FH:HE+FH:Db+FH:Ma+PH:AGE+PH:AD+PH:HE+PH:Db+PH:EH+PH:VH+PH:FH
## Crit= 41688.8222312507
## Mean crit= 41793.9252757946
## Change in best IC: -19.4980258005016 / Change in mean IC: -156.350838775732
##
## After 30 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+AD:AGE+WV:AGE+HE:WV+Db:WV+R:AGE+R:WV+Ma:WV+Ma:HE+Ma:Db+Ma:R+HV:WV+YE:AGE+YE:AD+YE:WV+YE:HE+YE:Db+YE:R+EH:AGE+EH:AD+EH:HE+VH:AGE+VH:HE+VH:Db+VH:Ma+FH:AGE+FH:AD+FH:WV+FH:HE+FH:Db+FH:Ma+PH:AGE+PH:HE+PH:Db+PH:EH+PH:VH+PH:FH
## Crit= 41681.4208977273
## Mean crit= 41765.8475837466
## Change in best IC: -7.40133352341945 / Change in mean IC: -28.0776920480421
##
## After 40 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+AD:AGE+WV:AGE+HE:AGE+HE:WV+Db:WV+R:AD+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:AGE+YE:AD+YE:WV+YE:Db+YE:R+EH:AGE+EH:AD+EH:HE+VH:AGE+VH:HE+VH:Db+VH:Ma+FH:AGE+FH:AD+FH:WV+FH:HE+FH:Db+FH:Ma+PH:AGE+PH:HE+PH:Db+PH:EH+PH:VH+PH:FH
## Crit= 41673.8572087935
## Mean crit= 41742.0297617133
## Change in best IC: -7.56368893371109 / Change in mean IC: -23.8178220332484
##
## After 50 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:AD+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:AGE+YE:AD+YE:WV+YE:Db+YE:R+EH:AGE+EH:AD+EH:HE+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:WV+FH:HE+FH:Ma+PH:AGE+PH:HE+PH:EH+PH:VH+PH:FH
## Crit= 41667.488947751
## Mean crit= 41724.8879385601
## Change in best IC: -6.36826104256033 / Change in mean IC: -17.1418231532225
##
## After 60 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:AD+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:AD+YE:WV+YE:Db+YE:R+EH:AGE+EH:AD+EH:HE+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:WV+FH:HE+PH:AGE+PH:HE+PH:EH+PH:VH+PH:FH
## Crit= 41664.4533933129
## Mean crit= 41707.4082359489
## Change in best IC: -3.03555443807272 / Change in mean IC: -17.4797026112137
##
## After 70 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:AD+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:AD+YE:WV+YE:Db+YE:R+EH:AGE+EH:AD+EH:HE+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:WV+PH:AGE+PH:HE+PH:EH+PH:VH+PH:FH
## Crit= 41662.4266729449
## Mean crit= 41696.7513272161
## Change in best IC: -2.02672036802687 / Change in mean IC: -10.6569087328025
##
## After 80 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:AD+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:AD+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:WV+PH:AGE+PH:HE+PH:EH+PH:VH+PH:FH
## Crit= 41662.1425470045
## Mean crit= 41686.1019053612
## Change in best IC: -0.284125940343074 / Change in mean IC: -10.6494218548542
##
## After 90 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:AD+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:AD+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:WV+PH:AGE+PH:HE+PH:EH+PH:VH+PH:FH
## Crit= 41659.9185834411
## Mean crit= 41675.5625784293
## Change in best IC: -2.22396356342506 / Change in mean IC: -10.5393269319175
##
## After 100 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:AD+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:AD+EH:HE+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:WV+FH:YE+PH:AGE+PH:EH+PH:VH+PH:FH
## Crit= 41657.8755075183
## Mean crit= 41670.0842086293
## Change in best IC: -2.04307592284749 / Change in mean IC: -5.47836980005377
##
## After 110 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:AD+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:AGE+PH:EH+PH:VH+PH:FH
## Crit= 41656.5250906619
## Mean crit= 41664.8535524076
## Change in best IC: -1.35041685638134 / Change in mean IC: -5.23065622167633
##
## After 120 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:AD+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:WV+FH:YE+PH:AGE+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41654.4664629805
## Mean crit= 41663.4879208671
## Change in best IC: -2.05862768135557 / Change in mean IC: -1.36563154047326
##
## After 130 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:AD+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:AGE+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41653.3412950177
## Mean crit= 41661.524946383
## Change in best IC: -1.12516796278214 / Change in mean IC: -1.96297448405676
##
## After 140 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:AD+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:AGE+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41653.3412950177
## Mean crit= 41660.4818478414
## Change in best IC: 0 / Change in mean IC: -1.04309854167514
##
## After 150 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:AD+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:AGE+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41653.3412950177
## Mean crit= 41659.4990576886
## Change in best IC: 0 / Change in mean IC: -0.982790152775124
##
## After 160 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:AD+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:AGE+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41652.95437192
## Mean crit= 41658.6759311569
## Change in best IC: -0.386923097728868 / Change in mean IC: -0.823126531664457
##
## After 170 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:AD+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:AGE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41651.8280364797
## Mean crit= 41657.4841712529
## Change in best IC: -1.12633544034179 / Change in mean IC: -1.19175990408257
##
## After 180 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41650.6066625744
## Mean crit= 41656.4194389578
## Change in best IC: -1.22137390526768 / Change in mean IC: -1.0647322950681
##
## After 190 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41650.6066625744
## Mean crit= 41655.4258499526
## Change in best IC: 0 / Change in mean IC: -0.993589005134709
##
## After 200 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41654.9777390728
## Change in best IC: -0.83534896261699 / Change in mean IC: -0.448110879806336
##
## After 210 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41654.1578372024
## Change in best IC: 0 / Change in mean IC: -0.819901870476315
##
## After 220 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41653.8669944624
## Change in best IC: 0 / Change in mean IC: -0.290842739916116
##
## After 230 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41653.5386888924
## Change in best IC: 0 / Change in mean IC: -0.328305570059456
##
## After 240 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41653.0097061926
## Change in best IC: 0 / Change in mean IC: -0.528982699812332
##
## After 250 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41652.6803383962
## Change in best IC: 0 / Change in mean IC: -0.329367796410224
##
## After 260 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41652.315574393
## Change in best IC: 0 / Change in mean IC: -0.364764003199525
##
## After 270 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41652.1138082236
## Change in best IC: 0 / Change in mean IC: -0.20176616939716
##
## After 280 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41652.0387722502
## Change in best IC: 0 / Change in mean IC: -0.075035973379272
##
## After 290 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41651.8468107547
## Change in best IC: 0 / Change in mean IC: -0.191961495533178
##
## After 300 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41651.7925800446
## Change in best IC: 0 / Change in mean IC: -0.0542307100768085
##
## After 310 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41651.713235772
## Change in best IC: 0 / Change in mean IC: -0.0793442725553177
##
## After 320 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41651.6425419273
## Change in best IC: 0 / Change in mean IC: -0.0706938447256107
##
## After 330 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41651.6274049255
## Change in best IC: 0 / Change in mean IC: -0.0151370017701993
##
## After 340 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41651.4780031535
## Change in best IC: 0 / Change in mean IC: -0.149401771981502
##
## After 350 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41651.4174380778
## Change in best IC: 0 / Change in mean IC: -0.0605650757133844
##
## After 360 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41651.3600651809
## Change in best IC: 0 / Change in mean IC: -0.0573728969175136
##
## After 370 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41651.2028886874
## Change in best IC: 0 / Change in mean IC: -0.157176493557927
##
## After 380 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41651.1112267338
## Change in best IC: 0 / Change in mean IC: -0.0916619536001235
##
## After 390 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41651.0193304457
## Change in best IC: 0 / Change in mean IC: -0.0918962880168692
##
## After 400 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41651.0022129356
## Change in best IC: 0 / Change in mean IC: -0.0171175101859262
##
## After 410 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41650.9502694325
## Change in best IC: 0 / Change in mean IC: -0.0519435030582827
##
## After 420 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41650.8986059369
## Change in best IC: 0 / Change in mean IC: -0.0516634955565678
##
## After 430 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41650.866729556
## Change in best IC: 0 / Change in mean IC: -0.0318763809555094
##
## After 440 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41650.8048932424
## Change in best IC: 0 / Change in mean IC: -0.0618363135654363
##
## After 450 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41650.7205089414
## Change in best IC: 0 / Change in mean IC: -0.0843843010443379
##
## After 460 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41650.698275538
## Change in best IC: 0 / Change in mean IC: -0.0222334033605875
##
## After 470 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41650.697600539
## Change in best IC: 0 / Change in mean IC: -0.000674998984322883
##
## After 480 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41650.6657004825
## Change in best IC: 0 / Change in mean IC: -0.0319000565432361
##
## After 490 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41650.6406904809
## Change in best IC: 0 / Change in mean IC: -0.0250100015691714
##
## After 500 generations:
## Best model: WWH~1+AGE+AD+WV+Db+Ma+YE+EH+VH+FH+PH+AD:AGE+WV:AGE+HE:AGE+HE:WV+R:WV+Ma:WV+Ma:Db+Ma:R+HV:HE+YE:WV+YE:Db+YE:R+EH:AGE+EH:HE+EH:Db+VH:AGE+VH:HE+FH:AGE+FH:AD+FH:YE+PH:WV+PH:Ma+PH:EH+PH:VH+PH:FH
## Crit= 41649.7713136118
## Mean crit= 41650.6330259536
## Improvements in best and average IC have bebingo en below the specified goals.
## Algorithm is declared to have converged.
## Completed.plot(lm.all.interactions.hat, type = "s")
##Commented to save runtime
# lm.all.interactions.hat <- glmulti(WWH ~ AGE + AD
# + WV + HE + WL
# + RH + R + Ma
# + HV + YE + Db
# + AW + AcD + N,
# data = df_intr,
# crit = aicc,
# level = 2,
# method = "g", # otherwise choose "d", "g" for algorithmic, or "h" for brute force
# family = gaussian,
# fit = glm,
# confsetsize = 100)
# plot(lm.all.interactions.hat, type = "s")
#
# lm.all.interactions.hat <- glmulti(WWH ~ AGE + AD
# + WV + HE
# + RH + R + Ma
# + HV + YE + Db
# + W + LW + DW + RW,
# data = df_intr,
# crit = aicc,
# level = 2,
# method = "g", # otherwise choose "d", "g" for algorithmic, or "h" for brute force
# family = gaussian,
# fit = glm,
# confsetsize = 100)
# plot(lm.all.interactions.hat, type = "s")Final Interacted and Endogenous Model
lm.all.interactions.hat <- glm(WEEKLY_WORK_HOURS ~ 1 + EXCELLENT_HEALTH + VERY_GOOD_HEALTH
+ FAIR_HEALTH
+ POOR_HEALTH + AGE + ANY_DEPENDENTS + MEDICAID + RETIRED
+ VOLUNTEER
+ MALE + WORK_LIMITING_CONDITION + WORKING
+ WEEKS_PAID_VACATION
+ REDUCE_PAID_WORK_HOURS + HOSPITAL_EXPENSES + HOME_VALUE
+ DEBTS
+ YEARS_EDUCATED + ACTIVE_ONCE_WEEKLY
+ ACTIVE_DAILY + NOT_ACTIVE + REALLY_LIKE_WORKING
+ DISLIKE_WORKING
+ REALLY_DISLIKE_WORKING + MISSING_WEEKS_PAID_VACATION
+ MISSING_REDUCE_PAID_WORK_HOURS + MISSING_HOME_VALUE
+ MISSING_DEBTS
+ MALE:POOR_HEALTH + ANY_DEPENDENTS:EXCELLENT_HEALTH
+ MALE:RETIRED
+ AGE:HOSPITAL_EXPENSES + AGE:FAIR_HEALTH
+ DEBTS:EXCELLENT_HEALTH
+ WEEKS_PAID_VACATION:YEARS_EDUCATED + AGE:EXCELLENT_HEALTH
+ HOME_VALUE:RETIRED + MALE:NOT_ACTIVE
+ ACTIVE_ONCE_WEEKLY:RETIRED
+ MALE:WEEKS_PAID_VACATION
+ ANY_DEPENDENTS:WORK_LIMITING_CONDITION
+ ACTIVE_ONCE_WEEKLY:DEBTS
+ HOME_VALUE:WORK_LIMITING_CONDITION
+ DISLIKE_WORKING:WORKING,
family = gaussian(link = "identity"),
data = df_hat)
lm.endog.interactions.t.so <- glm(WEEKLY_WORK_HOURS ~ 1 +
+ MEDICAID:AGE + MEDICAID:ANY_DEPENDENTS
+ MEDICAID:HOSPITAL_EXPENSES + RETIRED:WEEKS_PAID_VACATION
+ RETIRED:MEDICAID + MALE:WEEKS_PAID_VACATION
+ MALE:RETIRED + YEARS_EDUCATED:AGE + YEARS_EDUCATED:RETIRED
+ EXCELLENT_HEALTH + VERY_GOOD_HEALTH + FAIR_HEALTH
+ POOR_HEALTH + AGE + ANY_DEPENDENTS + MEDICAID + RETIRED
+ VOLUNTEER
+ MALE + WORK_LIMITING_CONDITION + WORKING + WEEKS_PAID_VACATION
+ REDUCE_PAID_WORK_HOURS + HOSPITAL_EXPENSES + HOME_VALUE
+ DEBTS
+ YEARS_EDUCATED + ACTIVE_ONCE_WEEKLY
+ ACTIVE_DAILY + NOT_ACTIVE + REALLY_LIKE_WORKING
+ DISLIKE_WORKING
+ REALLY_DISLIKE_WORKING + MISSING_WEEKS_PAID_VACATION
+ MISSING_REDUCE_PAID_WORK_HOURS + MISSING_HOME_VALUE
+ MISSING_DEBTS
+ I(AGE^2) + I(YEARS_EDUCATED^2)
+ I(YEARS_EDUCATED^2) * I(AGE^2)
+ I(AGE^2) * YEARS_EDUCATED + AGE * I(YEARS_EDUCATED^2)
+ I(WEEKS_PAID_VACATION^2),
family = gaussian(link = "identity"),
data = df_hat_transformed)
summary(lm.all.interactions.hat)##
## Call:
## glm(formula = WEEKLY_WORK_HOURS ~ 1 + EXCELLENT_HEALTH + VERY_GOOD_HEALTH +
## FAIR_HEALTH + POOR_HEALTH + AGE + ANY_DEPENDENTS + MEDICAID +
## RETIRED + VOLUNTEER + MALE + WORK_LIMITING_CONDITION + WORKING +
## WEEKS_PAID_VACATION + REDUCE_PAID_WORK_HOURS + HOSPITAL_EXPENSES +
## HOME_VALUE + DEBTS + YEARS_EDUCATED + ACTIVE_ONCE_WEEKLY +
## ACTIVE_DAILY + NOT_ACTIVE + REALLY_LIKE_WORKING + DISLIKE_WORKING +
## REALLY_DISLIKE_WORKING + MISSING_WEEKS_PAID_VACATION + MISSING_REDUCE_PAID_WORK_HOURS +
## MISSING_HOME_VALUE + MISSING_DEBTS + MALE:POOR_HEALTH + ANY_DEPENDENTS:EXCELLENT_HEALTH +
## MALE:RETIRED + AGE:HOSPITAL_EXPENSES + AGE:FAIR_HEALTH +
## DEBTS:EXCELLENT_HEALTH + WEEKS_PAID_VACATION:YEARS_EDUCATED +
## AGE:EXCELLENT_HEALTH + HOME_VALUE:RETIRED + MALE:NOT_ACTIVE +
## ACTIVE_ONCE_WEEKLY:RETIRED + MALE:WEEKS_PAID_VACATION + ANY_DEPENDENTS:WORK_LIMITING_CONDITION +
## ACTIVE_ONCE_WEEKLY:DEBTS + HOME_VALUE:WORK_LIMITING_CONDITION +
## DISLIKE_WORKING:WORKING, family = gaussian(link = "identity"),
## data = df_hat)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -39.371 -5.377 -0.646 5.068 67.847
##
## Coefficients: (1 not defined because of singularities)
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.061e+01 1.625e+00 24.988 < 2e-16
## EXCELLENT_HEALTH -3.399e+00 2.971e+00 -1.144 0.252621
## VERY_GOOD_HEALTH -4.982e-01 3.625e-01 -1.374 0.169400
## FAIR_HEALTH -7.697e+00 2.599e+00 -2.961 0.003081
## POOR_HEALTH -5.497e+00 1.322e+00 -4.157 3.28e-05
## AGE -1.521e-01 1.784e-02 -8.530 < 2e-16
## ANY_DEPENDENTS 3.601e-02 4.456e-01 0.081 0.935593
## MEDICAID -3.607e+00 5.503e-01 -6.555 6.10e-11
## RETIRED -1.285e+01 6.200e-01 -20.727 < 2e-16
## VOLUNTEER -4.321e-01 3.192e-01 -1.353 0.175967
## MALE 6.196e+00 4.865e-01 12.735 < 2e-16
## WORK_LIMITING_CONDITION -2.084e+00 5.479e-01 -3.803 0.000144
## WORKING 4.359e+00 6.308e-01 6.910 5.40e-12
## WEEKS_PAID_VACATION 1.644e+00 4.194e-01 3.919 9.00e-05
## REDUCE_PAID_WORK_HOURS -1.016e+00 3.647e-01 -2.785 0.005373
## HOSPITAL_EXPENSES -1.099e+00 7.899e-01 -1.392 0.163988
## HOME_VALUE 6.029e-05 2.712e-04 0.222 0.824041
## DEBTS 1.282e-03 8.420e-04 1.523 0.127815
## YEARS_EDUCATED 5.818e-02 6.739e-02 0.863 0.387998
## ACTIVE_ONCE_WEEKLY 4.757e-01 4.683e-01 1.016 0.309774
## ACTIVE_DAILY 1.000e+00 4.828e-01 2.072 0.038356
## NOT_ACTIVE 4.746e-01 4.914e-01 0.966 0.334172
## REALLY_LIKE_WORKING 6.112e-01 3.457e-01 1.768 0.077126
## DISLIKE_WORKING -1.866e+00 2.246e+00 -0.831 0.406228
## REALLY_DISLIKE_WORKING 7.384e-01 1.123e+00 0.657 0.510971
## MISSING_WEEKS_PAID_VACATION 2.081e+00 8.955e-01 2.324 0.020156
## MISSING_REDUCE_PAID_WORK_HOURS -1.204e+00 9.106e-01 -1.322 0.186300
## MISSING_HOME_VALUE -1.550e-01 3.234e-01 -0.479 0.631745
## MISSING_DEBTS NA NA NA NA
## POOR_HEALTH:MALE 3.059e+00 2.346e+00 1.304 0.192362
## EXCELLENT_HEALTH:ANY_DEPENDENTS -7.729e-01 1.306e+00 -0.592 0.553843
## RETIRED:MALE -1.633e+00 7.798e-01 -2.094 0.036351
## AGE:HOSPITAL_EXPENSES 2.431e-02 1.320e-02 1.842 0.065529
## FAIR_HEALTH:AGE 1.094e-01 4.247e-02 2.577 0.009999
## EXCELLENT_HEALTH:DEBTS -5.288e-03 2.874e-03 -1.840 0.065867
## WEEKS_PAID_VACATION:YEARS_EDUCATED -3.570e-02 2.765e-02 -1.291 0.196742
## EXCELLENT_HEALTH:AGE 6.864e-02 4.766e-02 1.440 0.149864
## RETIRED:HOME_VALUE -2.334e-03 6.547e-04 -3.565 0.000367
## MALE:NOT_ACTIVE 1.606e+00 7.182e-01 2.237 0.025350
## RETIRED:ACTIVE_ONCE_WEEKLY 1.851e+00 1.012e+00 1.828 0.067552
## MALE:WEEKS_PAID_VACATION -7.042e-01 1.498e-01 -4.701 2.65e-06
## ANY_DEPENDENTS:WORK_LIMITING_CONDITION 3.276e+00 1.179e+00 2.779 0.005469
## DEBTS:ACTIVE_ONCE_WEEKLY -2.407e-03 1.807e-03 -1.331 0.183102
## WORK_LIMITING_CONDITION:HOME_VALUE 1.477e-03 8.918e-04 1.656 0.097741
## WORKING:DISLIKE_WORKING 2.049e+00 2.310e+00 0.887 0.375116
##
## (Intercept) ***
## EXCELLENT_HEALTH
## VERY_GOOD_HEALTH
## FAIR_HEALTH **
## POOR_HEALTH ***
## AGE ***
## ANY_DEPENDENTS
## MEDICAID ***
## RETIRED ***
## VOLUNTEER
## MALE ***
## WORK_LIMITING_CONDITION ***
## WORKING ***
## WEEKS_PAID_VACATION ***
## REDUCE_PAID_WORK_HOURS **
## HOSPITAL_EXPENSES
## HOME_VALUE
## DEBTS
## YEARS_EDUCATED
## ACTIVE_ONCE_WEEKLY
## ACTIVE_DAILY *
## NOT_ACTIVE
## REALLY_LIKE_WORKING .
## DISLIKE_WORKING
## REALLY_DISLIKE_WORKING
## MISSING_WEEKS_PAID_VACATION *
## MISSING_REDUCE_PAID_WORK_HOURS
## MISSING_HOME_VALUE
## MISSING_DEBTS
## POOR_HEALTH:MALE
## EXCELLENT_HEALTH:ANY_DEPENDENTS
## RETIRED:MALE *
## AGE:HOSPITAL_EXPENSES .
## FAIR_HEALTH:AGE **
## EXCELLENT_HEALTH:DEBTS .
## WEEKS_PAID_VACATION:YEARS_EDUCATED
## EXCELLENT_HEALTH:AGE
## RETIRED:HOME_VALUE ***
## MALE:NOT_ACTIVE *
## RETIRED:ACTIVE_ONCE_WEEKLY .
## MALE:WEEKS_PAID_VACATION ***
## ANY_DEPENDENTS:WORK_LIMITING_CONDITION **
## DEBTS:ACTIVE_ONCE_WEEKLY
## WORK_LIMITING_CONDITION:HOME_VALUE .
## WORKING:DISLIKE_WORKING
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 122.1369)
##
## Null deviance: 1040019 on 5442 degrees of freedom
## Residual deviance: 659417 on 5399 degrees of freedom
## AIC: 41647
##
## Number of Fisher Scoring iterations: 2summary(lm.endog.interactions.t)##
## Call:
## glm(formula = WEEKLY_WORK_HOURS ~ 1 + HOSPITAL_EXPENSES + MEDICAID +
## VOLUNTEER + MALE + WEEKS_PAID_VACATION:AGE + HOSPITAL_EXPENSES:AGE +
## REDUCE_PAID_WORK_HOURS:WEEKS_PAID_VACATION + YEARS_EDUCATED +
## HOME_VALUE + MEDICAID:AGE + MEDICAID:ANY_DEPENDENTS + MEDICAID:HOSPITAL_EXPENSES +
## RETIRED:WEEKS_PAID_VACATION + RETIRED:MEDICAID + VOLUNTEER:RETIRED +
## MALE:WEEKS_PAID_VACATION + MALE:RETIRED + YEARS_EDUCATED:AGE +
## YEARS_EDUCATED:REDUCE_PAID_WORK_HOURS + YEARS_EDUCATED:RETIRED +
## AGE + ANY_DEPENDENTS + WORK_LIMITING_CONDITION + DEBTS +
## MISSING_WEEKS_PAID_VACATION + MISSING_REDUCE_PAID_WORK_HOURS +
## MISSING_HOME_VALUE + ACTIVE_ONCE_WEEKLY + ACTIVE_DAILY +
## NOT_ACTIVE + WORKING + REALLY_DISLIKE_WORKING + DISLIKE_WORKING +
## REALLY_DISLIKE_WORKING + EXCELLENT_HEALTH + VERY_GOOD_HEALTH +
## FAIR_HEALTH + POOR_HEALTH + MISSING_DEBTS, family = gaussian(link = "identity"),
## data = df_hat_transformed)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -41.700 -5.161 -0.652 4.946 61.596
##
## Coefficients: (1 not defined because of singularities)
## Estimate Std. Error t value
## (Intercept) 38.284832 4.414084 8.673
## HOSPITAL_EXPENSES -1.384162 2.751124 -0.503
## MEDICAID -18.654471 3.249404 -5.741
## VOLUNTEER -0.136441 0.354288 -0.385
## MALE 6.723348 0.446214 15.068
## YEARS_EDUCATED 0.325866 0.310782 1.049
## HOME_VALUE 0.021631 0.073188 0.296
## AGE -0.137825 0.069793 -1.975
## ANY_DEPENDENTS -0.151732 0.409185 -0.371
## WORK_LIMITING_CONDITION -1.430111 0.478175 -2.991
## DEBTS 0.171217 0.119116 1.437
## MISSING_WEEKS_PAID_VACATION 1.719272 0.882618 1.948
## MISSING_REDUCE_PAID_WORK_HOURS -1.031903 0.900712 -1.146
## MISSING_HOME_VALUE 0.168668 0.423649 0.398
## ACTIVE_ONCE_WEEKLY 0.802120 0.420860 1.906
## ACTIVE_DAILY 1.210597 0.476196 2.542
## NOT_ACTIVE 1.090748 0.386913 2.819
## WORKING 3.961806 0.611303 6.481
## REALLY_DISLIKE_WORKING 0.226865 1.105300 0.205
## DISLIKE_WORKING -0.133391 0.538553 -0.248
## EXCELLENT_HEALTH 0.504548 0.521392 0.968
## VERY_GOOD_HEALTH -0.520374 0.358520 -1.451
## FAIR_HEALTH -1.144212 0.470197 -2.433
## POOR_HEALTH -4.798499 1.098964 -4.366
## MISSING_DEBTS NA NA NA
## WEEKS_PAID_VACATION:AGE 0.014781 0.001804 8.191
## HOSPITAL_EXPENSES:AGE 0.033831 0.045341 0.746
## WEEKS_PAID_VACATION:REDUCE_PAID_WORK_HOURS 0.415226 0.173793 2.389
## MEDICAID:AGE 0.214975 0.053874 3.990
## MEDICAID:ANY_DEPENDENTS 3.670541 1.313923 2.794
## HOSPITAL_EXPENSES:MEDICAID 4.775081 1.730420 2.759
## WEEKS_PAID_VACATION:RETIRED 1.536728 0.276813 5.551
## MEDICAID:RETIRED 4.485591 1.268588 3.536
## VOLUNTEER:RETIRED -1.362506 0.760655 -1.791
## MALE:WEEKS_PAID_VACATION -0.833576 0.148855 -5.600
## MALE:RETIRED -2.418379 0.757621 -3.192
## AGE:YEARS_EDUCATED -0.001808 0.005030 -0.359
## REDUCE_PAID_WORK_HOURS:YEARS_EDUCATED -0.123118 0.034698 -3.548
## YEARS_EDUCATED:RETIRED -0.970823 0.051117 -18.992
## Pr(>|t|)
## (Intercept) < 2e-16 ***
## HOSPITAL_EXPENSES 0.614896
## MEDICAID 9.93e-09 ***
## VOLUNTEER 0.700168
## MALE < 2e-16 ***
## YEARS_EDUCATED 0.294440
## HOME_VALUE 0.767579
## AGE 0.048344 *
## ANY_DEPENDENTS 0.710789
## WORK_LIMITING_CONDITION 0.002795 **
## DEBTS 0.150661
## MISSING_WEEKS_PAID_VACATION 0.051476 .
## MISSING_REDUCE_PAID_WORK_HOURS 0.251989
## MISSING_HOME_VALUE 0.690549
## ACTIVE_ONCE_WEEKLY 0.056715 .
## ACTIVE_DAILY 0.011043 *
## NOT_ACTIVE 0.004833 **
## WORKING 9.93e-11 ***
## REALLY_DISLIKE_WORKING 0.837383
## DISLIKE_WORKING 0.804388
## EXCELLENT_HEALTH 0.333241
## VERY_GOOD_HEALTH 0.146712
## FAIR_HEALTH 0.014987 *
## POOR_HEALTH 1.29e-05 ***
## MISSING_DEBTS NA
## WEEKS_PAID_VACATION:AGE 3.20e-16 ***
## HOSPITAL_EXPENSES:AGE 0.455615
## WEEKS_PAID_VACATION:REDUCE_PAID_WORK_HOURS 0.016919 *
## MEDICAID:AGE 6.69e-05 ***
## MEDICAID:ANY_DEPENDENTS 0.005231 **
## HOSPITAL_EXPENSES:MEDICAID 0.005809 **
## WEEKS_PAID_VACATION:RETIRED 2.97e-08 ***
## MEDICAID:RETIRED 0.000410 ***
## VOLUNTEER:RETIRED 0.073313 .
## MALE:WEEKS_PAID_VACATION 2.25e-08 ***
## MALE:RETIRED 0.001421 **
## AGE:YEARS_EDUCATED 0.719332
## REDUCE_PAID_WORK_HOURS:YEARS_EDUCATED 0.000391 ***
## YEARS_EDUCATED:RETIRED < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 119.3468)
##
## Null deviance: 1040019 on 5442 degrees of freedom
## Residual deviance: 645070 on 5405 degrees of freedom
## AIC: 41515
##
## Number of Fisher Scoring iterations: 2summary(lm.endog.interactions.t.so) # Best one##
## Call:
## glm(formula = WEEKLY_WORK_HOURS ~ 1 + +MEDICAID:AGE + MEDICAID:ANY_DEPENDENTS +
## MEDICAID:HOSPITAL_EXPENSES + RETIRED:WEEKS_PAID_VACATION +
## RETIRED:MEDICAID + MALE:WEEKS_PAID_VACATION + MALE:RETIRED +
## YEARS_EDUCATED:AGE + YEARS_EDUCATED:RETIRED + EXCELLENT_HEALTH +
## VERY_GOOD_HEALTH + FAIR_HEALTH + POOR_HEALTH + AGE + ANY_DEPENDENTS +
## MEDICAID + RETIRED + VOLUNTEER + MALE + WORK_LIMITING_CONDITION +
## WORKING + WEEKS_PAID_VACATION + REDUCE_PAID_WORK_HOURS +
## HOSPITAL_EXPENSES + HOME_VALUE + DEBTS + YEARS_EDUCATED +
## ACTIVE_ONCE_WEEKLY + ACTIVE_DAILY + NOT_ACTIVE + REALLY_LIKE_WORKING +
## DISLIKE_WORKING + REALLY_DISLIKE_WORKING + MISSING_WEEKS_PAID_VACATION +
## MISSING_REDUCE_PAID_WORK_HOURS + MISSING_HOME_VALUE + MISSING_DEBTS +
## I(AGE^2) + I(YEARS_EDUCATED^2) + I(YEARS_EDUCATED^2) * I(AGE^2) +
## I(AGE^2) * YEARS_EDUCATED + AGE * I(YEARS_EDUCATED^2) + I(WEEKS_PAID_VACATION^2),
## family = gaussian(link = "identity"), data = df_hat_transformed)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -40.574 -5.224 -0.949 4.782 61.013
##
## Coefficients: (1 not defined because of singularities)
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1.745e+02 4.495e+01 3.883 0.000104 ***
## EXCELLENT_HEALTH 4.566e-01 5.209e-01 0.877 0.380761
## VERY_GOOD_HEALTH -5.223e-01 3.572e-01 -1.462 0.143714
## FAIR_HEALTH -1.118e+00 4.686e-01 -2.386 0.017057 *
## POOR_HEALTH -4.673e+00 1.094e+00 -4.272 1.97e-05 ***
## AGE -5.001e+00 1.441e+00 -3.471 0.000522 ***
## ANY_DEPENDENTS -8.639e-02 4.066e-01 -0.212 0.831734
## MEDICAID -1.745e+01 3.243e+00 -5.381 7.71e-08 ***
## RETIRED 1.019e+00 2.148e+00 0.475 0.635077
## VOLUNTEER -3.422e-01 3.151e-01 -1.086 0.277407
## MALE 6.698e+00 4.449e-01 15.056 < 2e-16 ***
## WORK_LIMITING_CONDITION -1.362e+00 4.755e-01 -2.863 0.004209 **
## WORKING 3.802e+00 6.109e-01 6.223 5.24e-10 ***
## WEEKS_PAID_VACATION 2.006e+00 1.825e-01 10.993 < 2e-16 ***
## REDUCE_PAID_WORK_HOURS -8.325e-01 3.585e-01 -2.322 0.020261 *
## HOSPITAL_EXPENSES 6.224e-01 4.617e-01 1.348 0.177679
## HOME_VALUE 2.466e-02 7.316e-02 0.337 0.736086
## DEBTS 1.954e-01 1.184e-01 1.651 0.098797 .
## YEARS_EDUCATED -2.277e+01 7.188e+00 -3.167 0.001548 **
## ACTIVE_ONCE_WEEKLY 8.013e-01 4.181e-01 1.917 0.055323 .
## ACTIVE_DAILY 1.148e+00 4.736e-01 2.424 0.015393 *
## NOT_ACTIVE 1.274e+00 3.849e-01 3.311 0.000936 ***
## REALLY_LIKE_WORKING 6.001e-01 3.399e-01 1.766 0.077496 .
## DISLIKE_WORKING -8.546e-02 5.444e-01 -0.157 0.875268
## REALLY_DISLIKE_WORKING 2.311e-01 1.103e+00 0.210 0.834015
## MISSING_WEEKS_PAID_VACATION 2.578e+00 8.862e-01 2.909 0.003645 **
## MISSING_REDUCE_PAID_WORK_HOURS -7.677e-01 8.943e-01 -0.858 0.390666
## MISSING_HOME_VALUE 2.816e-01 4.219e-01 0.668 0.504460
## MISSING_DEBTS NA NA NA NA
## I(AGE^2) 4.165e-02 1.136e-02 3.668 0.000247 ***
## I(YEARS_EDUCATED^2) 8.393e-01 2.873e-01 2.921 0.003500 **
## I(WEEKS_PAID_VACATION^2) -1.260e-01 1.805e-02 -6.978 3.35e-12 ***
## MEDICAID:AGE 2.016e-01 5.385e-02 3.744 0.000183 ***
## MEDICAID:ANY_DEPENDENTS 3.466e+00 1.306e+00 2.654 0.007981 **
## MEDICAID:HOSPITAL_EXPENSES 4.607e+00 1.724e+00 2.673 0.007543 **
## RETIRED:WEEKS_PAID_VACATION 1.632e+00 2.708e-01 6.028 1.77e-09 ***
## MEDICAID:RETIRED 4.447e+00 1.302e+00 3.415 0.000642 ***
## WEEKS_PAID_VACATION:MALE -8.582e-01 1.479e-01 -5.803 6.90e-09 ***
## RETIRED:MALE -2.344e+00 7.701e-01 -3.044 0.002347 **
## AGE:YEARS_EDUCATED 8.166e-01 2.295e-01 3.558 0.000377 ***
## RETIRED:YEARS_EDUCATED -1.076e+00 1.473e-01 -7.304 3.20e-13 ***
## I(AGE^2):I(YEARS_EDUCATED^2) 2.623e-04 7.118e-05 3.685 0.000231 ***
## YEARS_EDUCATED:I(AGE^2) -7.013e-03 1.802e-03 -3.893 0.000100 ***
## AGE:I(YEARS_EDUCATED^2) -3.023e-02 9.120e-03 -3.315 0.000922 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 117.6941)
##
## Null deviance: 1040019 on 5442 degrees of freedom
## Residual deviance: 635548 on 5400 degrees of freedom
## AIC: 41444
##
## Number of Fisher Scoring iterations: 2plot(lm.endog.interactions.t.so)
Quantile Analysis
quants <- c(.1, .25, .5, .75)
quantile.model <- map(quants,
~rq(WEEKLY_WORK_HOURS ~ EXCELLENT_HEALTH + VERY_GOOD_HEALTH + FAIR_HEALTH + POOR_HEALTH + AGE + YEARS_EDUCATED + HOSPITAL_EXPENSES + HOME_VALUE + DEBTS + WEEKS_PAID_VACATION + MALE + ANY_DEPENDENTS + WORKING_SPOUSE + HEART_CONDITION + RETIRED + VOLUNTEER + WORKING + WORK_LIMITING_CONDITION + REDUCE_PAID_WORK_HOURS + MEDICARE + MEDICAID + REALLY_LIKE_WORKING + DISLIKE_WORKING + REALLY_DISLIKE_WORKING + ACTIVE_ONCE_WEEKLY + ACTIVE_DAILY + NOT_ACTIVE + MISSING_HOME_VALUE + MISSING_WEEKS_PAID_VACATION + MISSING_REDUCE_PAID_WORK_HOURS,
tau = .x,
data = df_nonzero)
)## Warning in rq.fit.br(x, y, tau = tau, ...): Solution may be nonunique
## Warning in rq.fit.br(x, y, tau = tau, ...): Solution may be nonunique# summary(quantile.model[1])
# summary(quantile.model[2])
# summary(quantile.model[3])
stargazer::stargazer(quantile.model,
rq.se = "boot",
column.labels = c(paste("tau = ", quants)),
dep.var.labels = "Weekly Work Hours | Weekly Work Hours > 0",
model.numbers = TRUE,
model.names = FALSE,
keep.stat = c('n'),
type ='text')##
## ===========================================================================
## Dependent variable:
## --------------------------------------------
## Weekly Work Hours | Weekly Work Hours > 0
## tau = 0.1 tau = 0.25 tau = 0.5 tau = 0.75
## (1) (2) (3) (4)
## ---------------------------------------------------------------------------
## EXCELLENT_HEALTH 0.936 0.003 0.190 -0.214
## (0.791) (0.327) (0.217) (0.597)
##
## VERY_GOOD_HEALTH 0.427 -0.095 -0.056 -0.309
## (0.623) (0.225) (0.154) (0.423)
##
## FAIR_HEALTH -0.974 -0.525 -0.338 -1.376**
## (0.878) (0.336) (0.239) (0.543)
##
## POOR_HEALTH -9.172*** -5.803*** -1.945 -2.815***
## (2.601) (2.249) (1.543) (1.031)
##
## AGE -0.177*** -0.141*** -0.051** -0.102**
## (0.054) (0.032) (0.024) (0.042)
##
## YEARS_EDUCATED -0.070 0.005 0.022 0.114*
## (0.107) (0.038) (0.024) (0.059)
##
## HOSPITAL_EXPENSES 0.223 0.125 0.132* 0.137
## (0.157) (0.112) (0.079) (0.114)
##
## HOME_VALUE 0.0001 -0.00004 -0.0001 -0.0002
## (0.0003) (0.0001) (0.0001) (0.0003)
##
## DEBTS 0.002 0.001 0.0004 0.0002
## (0.001) (0.001) (0.0005) (0.001)
##
## WEEKS_PAID_VACATION 0.911*** 0.455*** 0.278** 0.643***
## (0.107) (0.072) (0.123) (0.102)
##
## MALE 6.588*** 3.337*** 1.599*** 5.959***
## (0.577) (0.335) (0.616) (0.512)
##
## ANY_DEPENDENTS -0.700 -0.626** -0.037 -0.021
## (0.718) (0.275) (0.148) (0.467)
##
## WORKING_SPOUSE 0.027 0.060 0.136 0.522
## (0.582) (0.213) (0.142) (0.390)
##
## HEART_CONDITION 0.110 0.406 -0.011 0.085
## (0.834) (0.352) (0.238) (0.554)
##
## RETIRED -15.053*** -16.527*** -15.890*** -11.650***
## (0.889) (0.763) (0.889) (0.777)
##
## VOLUNTEER -0.337 -0.294 -0.116 -0.495
## (0.620) (0.212) (0.127) (0.341)
##
## WORKING 5.330*** 6.524*** 4.166*** 3.629***
## (1.015) (0.953) (1.119) (1.250)
##
## WORK_LIMITING_CONDITION -1.379 -1.119*** -0.581 -0.841
## (0.967) (0.402) (0.394) (0.587)
##
## REDUCE_PAID_WORK_HOURS -2.337*** -1.043*** -0.330* -0.047
## (0.739) (0.265) (0.184) (0.386)
##
## MEDICARE -2.663** -2.009*** -0.945* -1.489**
## (1.082) (0.726) (0.522) (0.623)
##
## MEDICAID -7.706*** -6.426*** -1.873** -1.753***
## (1.012) (0.943) (0.779) (0.579)
##
## REALLY_LIKE_WORKING 0.149 0.097 0.118 1.103***
## (0.686) (0.248) (0.150) (0.383)
##
## DISLIKE_WORKING 1.033* -0.150 -0.287 -0.446
## (0.610) (0.338) (0.249) (0.536)
##
## REALLY_DISLIKE_WORKING 2.216 0.835* -0.006 -0.389
## (1.675) (0.490) (0.313) (1.340)
##
## ACTIVE_ONCE_WEEKLY 0.711 0.384 0.284 1.147**
## (0.725) (0.273) (0.196) (0.521)
##
## ACTIVE_DAILY 0.277 -0.048 0.274 1.996***
## (0.796) (0.298) (0.260) (0.683)
##
## NOT_ACTIVE 1.145* 0.763** 0.449* 1.337***
## (0.685) (0.326) (0.232) (0.463)
##
## MISSING_HOME_VALUE -0.694 -0.156 -0.141 -0.263
## (0.597) (0.202) (0.154) (0.406)
##
## MISSING_WEEKS_PAID_VACATION 1.261 1.504* 0.862* 1.606**
## (1.995) (0.896) (0.514) (0.723)
##
## MISSING_REDUCE_PAID_WORK_HOURS -7.510*** -5.628*** -0.365 2.750***
## (2.080) (1.035) (0.515) (0.923)
##
## Constant 34.197*** 37.265*** 37.588*** 41.396***
## (3.826) (2.115) (1.576) (3.063)
##
## ---------------------------------------------------------------------------
## Observations 5,443 5,443 5,443 5,443
## ===========================================================================
## Note: *p<0.1; **p<0.05; ***p<0.01quantile.model.sig <- map(quants,
~rq(WEEKLY_WORK_HOURS ~ FAIR_HEALTH + POOR_HEALTH + AGE + WEEKS_PAID_VACATION + MALE + RETIRED + WORKING + MEDICARE + MEDICAID,
tau = .x,
data = df_nonzero)
)## Warning in rq.fit.br(x, y, tau = tau, ...): Solution may be nonunique
## Warning in rq.fit.br(x, y, tau = tau, ...): Solution may be nonuniquequantile.model.sig.tidy <- map(quantile.model.sig, ~tidy(.x, se = "boot")) %>%
bind_rows()
lm.nonzero.sig <- lm(WEEKLY_WORK_HOURS ~ FAIR_HEALTH + POOR_HEALTH + AGE + WEEKS_PAID_VACATION + MALE + RETIRED + WORKING + MEDICARE + MEDICAID,
data=df_nonzero)
ols.tidy <- tidy(lm.nonzero.sig)
quantile.model.sig.tidy %>%
ggplot(aes(x = tau,
y = estimate
)
) + # the data we want to plot - in our case the tidied quantile regression output
geom_point(color = "#27408b",
size = 3
) + # plots the coefficient estimates
geom_line(color="#27408b",
size = 1
) + # adds a line connecting the coefficient estimates
geom_hline(data = ols.tidy,
aes(yintercept = estimate),
color = "red"
) + # add OLS estimate - note it is from another data set which is totally OK
facet_wrap(~term,
scales="free",
ncol=2
) + # one plot per explanatory variable, subplots over 2 columns
theme_bw()