TAS_Summary

TAS Descriptive statistics

We will be going through

Step 1: Loading Packages

library(tidyverse)
library(readxl)
library(ggplot2)
library (reshape2)
library(writexl)
library (lmerTest)
library(lme4)
library(dplyr)
library(ggpubr)
library(rstatix)
library(effectsize)
library(multcomp)
library(scales)

Step 2: Import the data

TAS_data_long_format_age <- read_excel("C:/ZZ_SherMay/BHP/TAS_data_long_format_age.xlsx") %>% unite("TAS_ID", c("1968 Interview Number", "Person Number")) %>% mutate(year_new = case_when(year == 2005 ~ -1, year == 2009 ~ 0,year == 2015 ~ 1))

Step 3: Preview the data

view(TAS_data_long_format_age)

knitr::kable(head(TAS_data_long_format_age))

TAS	TAS05	TAS09	TAS15	TAS_ID	Gender	Individual is sample	Year ID Number	Sequence Number	Relationship to Head	Release Number	B5A	B5D	B6C	C2D	C2E	C2F	D2D3_month	D2D3_year	E1_1st_mention	E1_2nd_mention	E1_3rd_mention	E3	G1	G2_month	G2_year	G10	G11	G30A	G41A	G41B	G41C	G41H	G41P	H1	L7_1st_mention	L7_2nd_mention	L7_3rd_mention	Age_17_graduate	Age_18_graduate	year	year_new
2	1	1	NA	4_180	2	3	771	2	22	5	3	5	7	3	1	1	2	2005	1	6	0	0	2	0	0	5	0	7	7	7	7	7	6	3	1	0	0	2022	2023	2005	-1
2	1	1	NA	5_32	2	2	624	3	30	5	5	5	5	7	7	7	0	0	1	7	0	0	1	5	2002	1	1	7	7	6	6	7	5	2	1	0	0	20	21	2005	-1
2	1	1	NA	6_34	1	2	1202	51	30	5	2	2	6	1	1	1	0	0	7	0	0	5	1	5	2002	1	1	0	7	5	7	5	3	1	1	0	0	20	21	2005	-1
2	1	1	NA	14_30	1	2	736	51	30	5	4	4	4	2	1	1	0	0	2	0	0	0	1	6	2003	1	5	6	5	6	6	5	5	2	1	0	0	19	20	2005	-1
1	1	NA	NA	18_38	2	2	5647	3	98	5	3	4	3	4	2	2	0	0	3	7	0	5	1	6	2004	1	5	6	5	5	5	7	5	2	1	0	0	18	19	2005	-1
2	1	1	NA	47_34	2	2	2516	3	30	5	4	5	6	4	5	2	0	0	1	0	0	0	1	5	2005	5	0	6	3	6	4	7	4	1	1	0	0	17	18	2005	-1

knitr::kable(TAS_data_long_format_age %>% count(year), align = "cc")

year	n
2005	745
2009	1554
2015	1641

knitr::kable(TAS_data_long_format_age %>% count(Age_18_graduate), align = "cc")

Age_18_graduate	n
18	465
19	504
20	366
21	287
22	331
23	256
24	208
25	121
26	33
27	16
28	1
29	1
32	1
2023	146
2027	264
2033	940

knitr::kable(TAS_data_long_format_age %>% group_by(year) %>%  count(Age_18_graduate) %>% ungroup(), align = "ccc")

year	Age_18_graduate	n
2005	18	179
2005	19	169
2005	20	164
2005	21	83
2005	22	2
2005	23	2
2005	2023	146
2009	18	195
2009	19	172
2009	20	166
2009	21	172
2009	22	183
2009	23	161
2009	24	153
2009	25	84
2009	26	2
2009	27	1
2009	28	1
2009	2027	264
2015	18	91
2015	19	163
2015	20	36
2015	21	32
2015	22	146
2015	23	93
2015	24	55
2015	25	37
2015	26	31
2015	27	15
2015	29	1
2015	32	1
2015	2033	940

Mean Age

mean_age <- TAS_data_long_format_age %>% group_by(year) %>% filter(Age_18_graduate <50) %>% summarize(mean_age = mean(Age_18_graduate, na.rm = TRUE)) %>% ungroup()
mean_age

## # A tibble: 3 × 2
##    year mean_age
##   <dbl>    <dbl>
## 1  2005     19.3
## 2  2009     21.2
## 3  2015     21.4

SD age

sd_age <- TAS_data_long_format_age %>% group_by(year) %>% filter(Age_18_graduate <50) %>% summarize(sd_age = sd(Age_18_graduate, na.rm = TRUE)) %>% ungroup()
sd_age

## # A tibble: 3 × 2
##    year sd_age
##   <dbl>  <dbl>
## 1  2005   1.07
## 2  2009   2.20
## 3  2015   2.54

###Education Status

TAS_data_long_format_age %>% group_by(year) %>% filter(G1 >0) %>% filter(G1<4) %>% count(G1) %>% ungroup()

## # A tibble: 9 × 3
##    year    G1     n
##   <dbl> <dbl> <int>
## 1  2005     1   601
## 2  2005     2    46
## 3  2005     3    96
## 4  2009     1  1292
## 5  2009     2   108
## 6  2009     3   153
## 7  2015     1   799
## 8  2015     2    52
## 9  2015     3    79

Attended College

TAS_data_long_format_age %>% group_by(year) %>% filter(G10 >0) %>% filter(G10<6) %>% count(G10) %>% ungroup()

## # A tibble: 6 × 3
##    year   G10     n
##   <dbl> <dbl> <int>
## 1  2005     1   487
## 2  2005     5   160
## 3  2009     1  1043
## 4  2009     5   352
## 5  2015     1   531
## 6  2015     5   348

Attending College

TAS_data_long_format_age %>% group_by(year) %>% filter(G11 >0) %>% filter(G11<6) %>% count(G11) %>% ungroup()

## # A tibble: 6 × 3
##    year   G11     n
##   <dbl> <dbl> <int>
## 1  2005     1   397
## 2  2005     5    90
## 3  2009     1   631
## 4  2009     5   412
## 5  2015     1   482
## 6  2015     5   704

TAS_data_long_format_age %>% group_by(year) %>% count(Gender) %>% ungroup()

## # A tibble: 6 × 3
##    year Gender     n
##   <dbl>  <dbl> <int>
## 1  2005      1   348
## 2  2005      2   397
## 3  2009      1   731
## 4  2009      2   823
## 5  2015      1   790
## 6  2015      2   851

TAS_data_long_format_age %>% group_by(year) %>% count(E3) %>% ungroup()

## # A tibble: 11 × 3
##     year    E3     n
##    <dbl> <dbl> <int>
##  1  2005     0   400
##  2  2005     1    86
##  3  2005     5   258
##  4  2005     9     1
##  5  2009     0   928
##  6  2009     1    96
##  7  2009     5   530
##  8  2015     0  1165
##  9  2015     1    51
## 10  2015     5   424
## 11  2015     9     1

---

2005 vs 2009 vs 2015 (between person; 18-19 y/o)

One-Way ANOVA

TAS_filtered_age <- TAS_data_long_format_age %>% group_by(TAS_ID) %>% mutate(Age_18_graduate = case_when(year == 2005 & Age_18_graduate == 2023 & any(year == 2009) ~ first(Age_18_graduate[year == 2009]) - 4, year == 2009 & Age_18_graduate == 2027 & any(year == 2005) ~ first(Age_18_graduate[year == 2005]) + 4, year == 2009 & Age_18_graduate == 2027 & any(year == 2015) ~ first(Age_18_graduate[year == 2015]) - 6, year == 2015 & Age_18_graduate == 2033 & any(year == 2009) ~ first(Age_18_graduate[year == 2009]) + 6,Age_18_graduate >= 1 & Age_18_graduate <= 50 ~ Age_18_graduate, TRUE ~ Age_18_graduate)) %>% ungroup() %>%  filter(Age_18_graduate == 18 | Age_18_graduate == 19) %>% mutate(year_new = factor(year_new, levels = c(-1, 0, 1), labels = c("2005", "2009", "2015")))

view(TAS_filtered_age)

knitr::kable(head(TAS_filtered_age[, 1:42]))

TAS	TAS05	TAS09	TAS15	TAS_ID	Gender	Individual is sample	Year ID Number	Sequence Number	Relationship to Head	Release Number	B5A	B5D	B6C	C2D	C2E	C2F	D2D3_month	D2D3_year	E1_1st_mention	E1_2nd_mention	E1_3rd_mention	E3	G1	G2_month	G2_year	G10	G11	G30A	G41A	G41B	G41C	G41H	G41P	H1	L7_1st_mention	L7_2nd_mention	L7_3rd_mention	Age_17_graduate	Age_18_graduate	year	year_new
1	1	NA	NA	18_38	2	2	5647	3	98	5	3	4	3	4	2	2	0	0	3	7	0	5	1	6	2004	1	5	6	5	5	5	7	5	2	1	0	0	18	19	2005	2005
2	1	1	NA	47_34	2	2	2516	3	30	5	4	5	6	4	5	2	0	0	1	0	0	0	1	5	2005	5	0	6	3	6	4	7	4	1	1	0	0	17	18	2005	2005
2	1	1	NA	53_36	2	2	1616	3	30	5	4	5	7	4	1	1	0	0	6	0	0	1	1	6	2005	1	5	7	7	7	5	7	6	2	1	0	0	17	18	2005	2005
2	1	1	NA	79_32	2	2	6520	2	30	5	3	4	6	7	5	3	0	0	1	7	0	0	1	5	2004	1	1	0	7	7	6	7	4	1	1	0	0	18	19	2005	2005
2	1	1	NA	88_35	1	2	3411	2	30	5	2	5	7	3	1	2	0	0	1	0	0	0	1	5	2005	1	1	7	2	6	5	6	7	2	1	0	0	17	18	2005	2005
2	1	1	NA	89_34	2	2	4527	3	30	5	2	4	5	2	3	1	0	0	3	7	0	5	1	5	2005	1	1	7	5	6	4	7	5	1	1	0	0	17	18	2005	2005

knitr::kable(TAS_filtered_age %>% group_by(year) %>% count(Age_18_graduate), align = "ccc")

year	Age_18_graduate	n
2005	18	185
2005	19	171
2009	18	196
2009	19	173
2015	18	91
2015	19	163

---

B5A: Responsibility for self

knitr::kable(TAS_filtered_age %>% count(B5A), align = "cc")

B5A	n
1	48
2	202
3	243
4	318
5	167
9	1

B5A_filtered <- TAS_filtered_age %>% filter(B5A < 8) %>% filter(B5A >0)

B5A_Anova <- aov(B5A ~ year_new, data = B5A_filtered)
summary(B5A_Anova)

##              Df Sum Sq Mean Sq F value Pr(>F)
## year_new      2    4.7   2.336   1.826  0.162
## Residuals   975 1247.2   1.279

B5A_pairwise <-TukeyHSD(B5A_Anova)
B5A_pairwise

##   Tukey multiple comparisons of means
##     95% family-wise confidence level
## 
## Fit: aov(formula = B5A ~ year_new, data = B5A_filtered)
## 
## $year_new
##                  diff         lwr       upr     p adj
## 2009-2005 -0.09379580 -0.29115264 0.1035610 0.5046582
## 2015-2005  0.08028842 -0.13776097 0.2983378 0.6629646
## 2015-2009  0.17408422 -0.04247977 0.3906482 0.1430000

B5A_glht <- glht(B5A_Anova, linfct = mcp(year_new = "Tukey")) 
summary(B5A_glht)

## 
##   Simultaneous Tests for General Linear Hypotheses
## 
## Multiple Comparisons of Means: Tukey Contrasts
## 
## 
## Fit: aov(formula = B5A ~ year_new, data = B5A_filtered)
## 
## Linear Hypotheses:
##                  Estimate Std. Error t value Pr(>|t|)
## 2009 - 2005 == 0 -0.09380    0.08408  -1.116    0.504
## 2015 - 2005 == 0  0.08029    0.09289   0.864    0.663
## 2015 - 2009 == 0  0.17408    0.09226   1.887    0.143
## (Adjusted p values reported -- single-step method)

B5A_pairwise_t_test <- pairwise.t.test(B5A_filtered$B5A, B5A_filtered$year_new, p.adjust.method = "BH")
print(B5A_pairwise_t_test)

## 
##  Pairwise comparisons using t tests with pooled SD 
## 
## data:  B5A_filtered$B5A and B5A_filtered$year_new 
## 
##      2005 2009
## 2009 0.39 -   
## 2015 0.39 0.18
## 
## P value adjustment method: BH

B5A_means <- B5A_filtered %>%  group_by(year_new) %>% summarise(Mean = mean(B5A),SD = sd(B5A),Count = n())
knitr::kable((B5A_means), align = "cccc")

year_new	Mean	SD	Count
2005	3.376405	1.105125	356
2009	3.282609	1.151481	368
2015	3.456693	1.136918	254

B5A_filtered$predicted_B5A <- predict(B5A_Anova)
B5A_filtered$year_factor_B5A <- factor(B5A_filtered$year)

ggplot(B5A_filtered, aes(x = Age_18_graduate, y = B5A, color = year_factor_B5A)) + geom_point(aes(shape = year_factor_B5A), alpha = 0.5) + geom_line(aes(y = predicted_B5A), size = 1) +  labs(title = "Responsibility for Self (B5A) by Age and Year",x = "Age", y = "Responsibility for Self (B5A)", color = "Year", shape = "Year") + theme_minimal()

## Warning: Using `size` aesthetic for lines was deprecated in ggplot2 3.4.0.
## ℹ Please use `linewidth` instead.
## This warning is displayed once every 8 hours.
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
## generated.

ggplot(B5A_filtered, aes(x = year_factor_B5A, y = B5A, fill = year_factor_B5A)) + geom_boxplot() + stat_summary(fun = "mean", geom = "crossbar", width = 0.75, color = "black", size = 0.2, linetype = "dashed") + labs(title = "Box Plot of Responsibility for Self (B5A) in 2005, 2009 and 2015",x = "Year", y = "Responsibility for Self (B5A)", fill = "Year") + theme_minimal()

B6C: Money management skills

knitr::kable(TAS_filtered_age %>% count(B6C), align = "cc")

B6C	n
1	9
2	23
3	52
4	122
5	276
6	239
7	258

B6C_filtered <- TAS_filtered_age %>% filter(B6C < 8) %>% filter(B6C >0)

B6C_Anova <- aov(B6C ~ year_new, data = B6C_filtered)
summary(B6C_Anova)

##              Df Sum Sq Mean Sq F value Pr(>F)
## year_new      2    4.6   2.297   1.271  0.281
## Residuals   976 1763.8   1.807

B6C_pairwise <-TukeyHSD(B6C_Anova)
B6C_pairwise

##   Tukey multiple comparisons of means
##     95% family-wise confidence level
## 
## Fit: aov(formula = B6C ~ year_new, data = B6C_filtered)
## 
## $year_new
##                  diff        lwr       upr     p adj
## 2009-2005  0.12987577 -0.1045438 0.3642953 0.3952346
## 2015-2005 -0.02430771 -0.2834785 0.2348631 0.9736358
## 2015-2009 -0.15418347 -0.4114462 0.1030793 0.3377514

B6C_glht <- glht(B6C_Anova, linfct = mcp(year_new = "Tukey")) 
summary(B6C_glht)

## 
##   Simultaneous Tests for General Linear Hypotheses
## 
## Multiple Comparisons of Means: Tukey Contrasts
## 
## 
## Fit: aov(formula = B6C ~ year_new, data = B6C_filtered)
## 
## Linear Hypotheses:
##                  Estimate Std. Error t value Pr(>|t|)
## 2009 - 2005 == 0  0.12988    0.09987   1.300    0.395
## 2015 - 2005 == 0 -0.02431    0.11041  -0.220    0.974
## 2015 - 2009 == 0 -0.15418    0.10960  -1.407    0.337
## (Adjusted p values reported -- single-step method)

B6C_pairwise_t_test <- pairwise.t.test(B6C_filtered$B6C, B6C_filtered$year_new, p.adjust.method = "BH")
print(B6C_pairwise_t_test)

## 
##  Pairwise comparisons using t tests with pooled SD 
## 
## data:  B6C_filtered$B6C and B6C_filtered$year_new 
## 
##      2005 2009
## 2009 0.29 -   
## 2015 0.83 0.29
## 
## P value adjustment method: BH

B6C_means <- B6C_filtered %>%  group_by(year_new) %>% summarise(Mean = mean(B6C),SD = sd(B6C),Count = n())
knitr::kable((B6C_means), align = "cccc")

year_new	Mean	SD	Count
2005	5.390449	1.434708	356
2009	5.520325	1.231491	369
2015	5.366142	1.370134	254

B6C_filtered$predicted_B6C <- predict(B6C_Anova)
B6C_filtered$year_factor_B6C <- factor(B6C_filtered$year)

ggplot(B6C_filtered, aes(x = Age_18_graduate, y = B6C, color = year_factor_B6C)) + geom_point(aes(shape = year_factor_B6C), alpha = 0.5) + geom_line(aes(y = predicted_B6C), size = 1) +  labs(title = "Money management skills (B6C) by Age and Year",x = "Age", y = "Money management skills (B6C)", color = "Year", shape = "Year") + theme_minimal()

ggplot(B6C_filtered, aes(x = year_factor_B6C, y = B6C, fill = year_factor_B6C)) + geom_boxplot() + stat_summary(fun = "mean", geom = "crossbar", width = 0.75, color = "black", size = 0.2, linetype = "dashed") + labs(title = "Box Plot of Money management skills (B6C) in 2005, 2009 and 2015",x = "Year", y = "Money management skills (B6C)", fill = "Year") + theme_minimal()

C2D: Worry about expenses

knitr::kable(TAS_filtered_age %>% count(C2D), align = "cc")

C2D	n
1	165
2	158
3	159
4	171
5	144
6	84
7	98

C2D_filtered <- TAS_filtered_age %>% filter(C2D < 8) %>% filter(C2D >0)

C2D_Anova <- aov(C2D ~ year_new, data = C2D_filtered)
summary(C2D_Anova)

##              Df Sum Sq Mean Sq F value Pr(>F)
## year_new      2      8   4.105   1.146  0.318
## Residuals   976   3494   3.580

C2D_pairwise <-TukeyHSD(C2D_Anova)
C2D_pairwise

##   Tukey multiple comparisons of means
##     95% family-wise confidence level
## 
## Fit: aov(formula = C2D ~ year_new, data = C2D_filtered)
## 
## $year_new
##                  diff        lwr       upr     p adj
## 2009-2005 -0.04100058 -0.3709610 0.2889599 0.9541902
## 2015-2005 -0.22586924 -0.5906686 0.1389302 0.3140844
## 2015-2009 -0.18486866 -0.5469824 0.1772451 0.4544183

C2D_glht <- glht(C2D_Anova, linfct = mcp(year_new = "Tukey")) 
summary(C2D_glht)

## 
##   Simultaneous Tests for General Linear Hypotheses
## 
## Multiple Comparisons of Means: Tukey Contrasts
## 
## 
## Fit: aov(formula = C2D ~ year_new, data = C2D_filtered)
## 
## Linear Hypotheses:
##                  Estimate Std. Error t value Pr(>|t|)
## 2009 - 2005 == 0  -0.0410     0.1406  -0.292    0.954
## 2015 - 2005 == 0  -0.2259     0.1554  -1.453    0.314
## 2015 - 2009 == 0  -0.1849     0.1543  -1.198    0.454
## (Adjusted p values reported -- single-step method)

C2D_pairwise_t_test <- pairwise.t.test(C2D_filtered$C2D, C2D_filtered$year_new, p.adjust.method = "BH")
print(C2D_pairwise_t_test)

## 
##  Pairwise comparisons using t tests with pooled SD 
## 
## data:  C2D_filtered$C2D and C2D_filtered$year_new 
## 
##      2005 2009
## 2009 0.77 -   
## 2015 0.35 0.35
## 
## P value adjustment method: BH

C2D_means <- C2D_filtered %>%  group_by(year_new) %>% summarise(Mean = mean(C2D),SD = sd(C2D),Count = n())
knitr::kable((C2D_means), align = "cccc")

year_new	Mean	SD	Count
2005	3.702247	1.884285	356
2009	3.661247	1.925696	369
2015	3.476378	1.853699	254

C2D_filtered$predicted_C2D <- predict(C2D_Anova)
C2D_filtered$year_factor_C2D <- factor(C2D_filtered$year)

ggplot(C2D_filtered, aes(x = Age_18_graduate, y = C2D, color = year_factor_C2D)) + geom_point(aes(shape = year_factor_C2D), alpha = 0.5) + geom_line(aes(y = predicted_C2D), size = 1) +  labs(title = "Worry about expense (C2D) by Age and Year",x = "Age", y = "Worry about expense (C2D)", color = "Year", shape = "Year") + theme_minimal()

ggplot(C2D_filtered, aes(x = year_factor_C2D, y = C2D, fill = year_factor_C2D)) + geom_boxplot() + stat_summary(fun = "mean", geom = "crossbar", width = 0.75, color = "black", size = 0.2, linetype = "dashed") + labs(title = "Box Plot of Worry about expense (C2D) in 2005, 2009 and 2015",x = "Year", y = "Worry about expense (C2D)", fill = "Year") + theme_minimal()

C2E: Worry about future job

knitr::kable(TAS_filtered_age %>% count(C2E), align = "cc")

C2E	n
1	174
2	175
3	158
4	139
5	133
6	88
7	112

C2E_filtered <- TAS_filtered_age %>% filter(C2E < 8) %>% filter(C2E >0)

C2E_Anova <- aov(C2E ~ year_new, data = C2E_filtered)
summary(C2E_Anova)

##              Df Sum Sq Mean Sq F value Pr(>F)
## year_new      2      1   0.433   0.112  0.894
## Residuals   976   3765   3.857

C2E_pairwise <-TukeyHSD(C2E_Anova)
C2E_pairwise

##   Tukey multiple comparisons of means
##     95% family-wise confidence level
## 
## Fit: aov(formula = C2E ~ year_new, data = C2E_filtered)
## 
## $year_new
##                  diff        lwr       upr     p adj
## 2009-2005 0.061310557 -0.2811726 0.4037937 0.9072802
## 2015-2005 0.062505530 -0.3161388 0.4411499 0.9205857
## 2015-2009 0.001194973 -0.3746618 0.3770518 0.9999693

C2E_glht <- glht(C2E_Anova, linfct = mcp(year_new = "Tukey")) 
summary(C2E_glht)

## 
##   Simultaneous Tests for General Linear Hypotheses
## 
## Multiple Comparisons of Means: Tukey Contrasts
## 
## 
## Fit: aov(formula = C2E ~ year_new, data = C2E_filtered)
## 
## Linear Hypotheses:
##                  Estimate Std. Error t value Pr(>|t|)
## 2009 - 2005 == 0 0.061311   0.145906   0.420    0.907
## 2015 - 2005 == 0 0.062506   0.161311   0.387    0.920
## 2015 - 2009 == 0 0.001195   0.160124   0.007    1.000
## (Adjusted p values reported -- single-step method)

C2E_pairwise_t_test <- pairwise.t.test(C2E_filtered$C2E, C2E_filtered$year_new, p.adjust.method = "BH")
print(C2E_pairwise_t_test)

## 
##  Pairwise comparisons using t tests with pooled SD 
## 
## data:  C2E_filtered$C2E and C2E_filtered$year_new 
## 
##      2005 2009
## 2009 0.99 -   
## 2015 0.99 0.99
## 
## P value adjustment method: BH

C2E_means <- C2E_filtered %>%  group_by(year_new) %>% summarise(Mean = mean(C2E),SD = sd(C2E),Count = n())
knitr::kable((C2E_means), align = "cccc")

year_new	Mean	SD	Count
2005	3.567416	1.945296	356
2009	3.628726	1.974109	369
2015	3.629921	1.975355	254

C2E_filtered$predicted_C2E <- predict(C2E_Anova)
C2E_filtered$year_factor_C2E <- factor(C2E_filtered$year)

ggplot(C2E_filtered, aes(x = Age_18_graduate, y = C2E, color = year_factor_C2E)) + geom_point(aes(shape = year_factor_C2E), alpha = 0.5) + geom_line(aes(y = predicted_C2E), size = 1) +  labs(title = "Worry about future job (C2E) by Age and Year",x = "Age", y = "Worry about future job (C2E)", color = "Year", shape = "Year") + theme_minimal()

ggplot(C2E_filtered, aes(x = year_factor_C2E, y = C2E, fill = year_factor_C2E)) + geom_boxplot() + stat_summary(fun = "mean", geom = "crossbar", width = 0.75, color = "black", size = 0.2, linetype = "dashed") + labs(title = "Box Plot of Worry about future job (C2E) in 2005, 2009 and 2015",x = "Year", y = "Worry about future job (C2E)", fill = "Year") + theme_minimal()

G41A: Importance of job status

knitr::kable(TAS_filtered_age %>% count(G41A), align = "cc")

G41A	n
1	38
2	36
3	76
4	116
5	239
6	199
7	274
9	1

G41A_filtered <- TAS_filtered_age %>% filter(G41A < 8) %>% filter(G41A >0)

G41A_Anova <- aov(G41A ~ year_new, data = G41A_filtered)
summary(G41A_Anova)

##              Df Sum Sq Mean Sq F value   Pr(>F)    
## year_new      2   42.2  21.091   8.046 0.000342 ***
## Residuals   975 2555.8   2.621                     
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

G41A_pairwise <-TukeyHSD(G41A_Anova)
G41A_pairwise

##   Tukey multiple comparisons of means
##     95% family-wise confidence level
## 
## Fit: aov(formula = G41A ~ year_new, data = G41A_filtered)
## 
## $year_new
##                  diff        lwr        upr     p adj
## 2009-2005 -0.03083781 -0.3133562  0.2516806 0.9644608
## 2015-2005 -0.48834380 -0.8004838 -0.1762038 0.0007404
## 2015-2009 -0.45750599 -0.7675196 -0.1474923 0.0016083

G41A_glht <- glht(G41A_Anova, linfct = mcp(year_new = "Tukey")) 
summary(G41A_glht)

## 
##   Simultaneous Tests for General Linear Hypotheses
## 
## Multiple Comparisons of Means: Tukey Contrasts
## 
## 
## Fit: aov(formula = G41A ~ year_new, data = G41A_filtered)
## 
## Linear Hypotheses:
##                  Estimate Std. Error t value Pr(>|t|)    
## 2009 - 2005 == 0 -0.03084    0.12036  -0.256 0.964399    
## 2015 - 2005 == 0 -0.48834    0.13298  -3.672 0.000746 ***
## 2015 - 2009 == 0 -0.45751    0.13207  -3.464 0.001587 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- single-step method)

G41A_pairwise_t_test <- pairwise.t.test(G41A_filtered$G41A, G41A_filtered$year_new, p.adjust.method = "BH")
print(G41A_pairwise_t_test)

## 
##  Pairwise comparisons using t tests with pooled SD 
## 
## data:  G41A_filtered$G41A and G41A_filtered$year_new 
## 
##      2005    2009   
## 2009 0.79784 -      
## 2015 0.00076 0.00083
## 
## P value adjustment method: BH

G41A_means <- G41A_filtered %>%  group_by(year_new) %>% summarise(Mean = mean(G41A),SD = sd(G41A),Count = n())
knitr::kable((G41A_means), align = "cccc")

year_new	Mean	SD	Count
2005	5.362360	1.627454	356
2009	5.331522	1.530395	368
2015	4.874016	1.728588	254

G41A_filtered$predicted_G41A <- predict(G41A_Anova)
G41A_filtered$year_factor_G41A <- factor(G41A_filtered$year)

ggplot(G41A_filtered, aes(x = Age_18_graduate, y = G41A, color = year_factor_G41A)) + geom_point(aes(shape = year_factor_G41A), alpha = 0.5) + geom_line(aes(y = predicted_G41A), size = 1) +  labs(title = "Importance of job status (G41A) by Age and Year",x = "Age", y = "Importance of job status (G41A)", color = "Year", shape = "Year") + theme_minimal()

ggplot(G41A_filtered, aes(x = year_factor_G41A, y = G41A, fill = year_factor_G41A)) + geom_boxplot() + stat_summary(fun = "mean", geom = "crossbar", width = 0.75, color = "black", size = 0.2, linetype = "dashed") + labs(title = "Box Plot of Importance of job status (G41A) in 2005, 2009 and 2015",x = "Year", y = "Importance of job status (G41A)", fill = "Year") + theme_minimal()

2005 & 2009 & 2015 (between person; 18-19 y/o)

Demographics

Filter the data (2005 & 2009 & 2015)

Long_format_2005_2009_2015_new <-  TAS_data_long_format_age %>% filter(year==2005| year==2015| year == 2009) %>% filter(Age_18_graduate == 18|Age_18_graduate == 19) %>% mutate(year_new = case_when(year == 2005 ~ -1, year == 2009 ~ 0,year == 2015 ~ 1)) %>% mutate(year_new = case_when(year == 2005 ~ -1, year == 2009 ~ 0,year == 2015 ~ 1))

knitr::kable(head(Long_format_2005_2009_2015_new[, 1:42]))

TAS	TAS05	TAS09	TAS15	TAS_ID	Gender	Individual is sample	Year ID Number	Sequence Number	Relationship to Head	Release Number	B5A	B5D	B6C	C2D	C2E	C2F	D2D3_month	D2D3_year	E1_1st_mention	E1_2nd_mention	E1_3rd_mention	E3	G1	G2_month	G2_year	G10	G11	G30A	G41A	G41B	G41C	G41H	G41P	H1	L7_1st_mention	L7_2nd_mention	L7_3rd_mention	Age_17_graduate	Age_18_graduate	year	year_new
1	1	NA	NA	18_38	2	2	5647	3	98	5	3	4	3	4	2	2	0	0	3	7	0	5	1	6	2004	1	5	6	5	5	5	7	5	2	1	0	0	18	19	2005	-1
2	1	1	NA	47_34	2	2	2516	3	30	5	4	5	6	4	5	2	0	0	1	0	0	0	1	5	2005	5	0	6	3	6	4	7	4	1	1	0	0	17	18	2005	-1
2	1	1	NA	53_36	2	2	1616	3	30	5	4	5	7	4	1	1	0	0	6	0	0	1	1	6	2005	1	5	7	7	7	5	7	6	2	1	0	0	17	18	2005	-1
2	1	1	NA	79_32	2	2	6520	2	30	5	3	4	6	7	5	3	0	0	1	7	0	0	1	5	2004	1	1	0	7	7	6	7	4	1	1	0	0	18	19	2005	-1
2	1	1	NA	88_35	1	2	3411	2	30	5	2	5	7	3	1	2	0	0	1	0	0	0	1	5	2005	1	1	7	2	6	5	6	7	2	1	0	0	17	18	2005	-1
2	1	1	NA	89_34	2	2	4527	3	30	5	2	4	5	2	3	1	0	0	3	7	0	5	1	5	2005	1	1	7	5	6	4	7	5	1	1	0	0	17	18	2005	-1

knitr::kable(Long_format_2005_2009_2015_new %>% count(year), align = "cc")

year	n
2005	348
2009	367
2015	254

---

Demographics

Year

knitr::kable(count(Long_format_2005_2009_2015_new, year), align = "cc")

year	n
2005	348
2009	367
2015	254

Age

knitr::kable(count(Long_format_2005_2009_2015_new, Age_18_graduate), align = "cc")

Age_18_graduate	n
18	465
19	504

Age by year (2005)

knitr::kable(Long_format_2005_2009_2015_new %>% filter(year == 2005) %>% count(Age_18_graduate), align = "cc")

Age_18_graduate	n
18	179
19	169

Age by year (2009)

knitr::kable(Long_format_2005_2009_2015_new %>% filter(year == 2009) %>% count(Age_18_graduate), align = "cc")

Age_18_graduate	n
18	195
19	172

Age by year (2015)

knitr::kable(Long_format_2005_2009_2015_new %>% filter(year == 2015) %>% count(Age_18_graduate), align = "cc")

Age_18_graduate	n
18	91
19	163

Gender [1 = Male; 2 = Female]

knitr::kable(Long_format_2005_2009_2015_new %>% group_by(year) %>% count(Gender), align = "ccc")

year	Gender	n
2005	1	169
2005	2	179
2009	1	165
2009	2	202
2015	1	119
2015	2	135

Years graduated high school (by year)

knitr::kable(Long_format_2005_2009_2015_new %>% group_by(year) %>% count(G2_year), align = "ccc")

year	G2_year	n
2005	2004	169
2005	2005	179
2009	2008	172
2009	2009	195
2015	2014	163
2015	2015	91

mean_year_graduation <- Long_format_2005_2009_2015_new %>% group_by(year) %>% summarize(mean_graduation = mean(G2_year, na.rm = TRUE)) %>% ungroup()
knitr::kable((mean_year_graduation), align = "cc")

year	mean_graduation
2005	2004.514
2009	2008.531
2015	2014.358

sd_year_graduation <- Long_format_2005_2009_2015_new %>% group_by(year) %>% summarize(sd_graduation = sd(G2_year, na.rm = TRUE)) %>% ungroup()
knitr::kable((sd_year_graduation), align = "cc")

year	sd_graduation
2005	0.5005132
2009	0.4996984
2015	0.4804380

Race – 1st mention [1 = White; 2 = Black, African-American, or Negro; 3 = American Indian or Alaska Native; 4 = Asian; 5 = Native Hawaiian or Pacific Islander; 7 = Some other race; 8 = DK;9 = NA]

knitr::kable(Long_format_2005_2009_2015_new %>% group_by(year) %>% count(L7_1st_mention), align = "ccc")

year	L7_1st_mention	n
2005	1	192
2005	2	129
2005	3	1
2005	4	5
2005	5	3
2005	7	4
2005	8	1
2005	9	13
2009	1	184
2009	2	139
2009	3	6
2009	4	10
2009	5	1
2009	7	25
2009	8	1
2009	9	1
2015	1	139
2015	2	99
2015	3	1
2015	4	3
2015	5	2
2015	7	9
2015	9	1

Race – 2nd mention

knitr::kable(Long_format_2005_2009_2015_new %>% group_by(year) %>% count(L7_2nd_mention), align = "ccc")

year	L7_2nd_mention	n
2005	0	341
2005	2	2
2005	3	2
2005	5	2
2005	7	1
2009	0	347
2009	1	8
2009	2	1
2009	3	6
2009	5	2
2009	7	3
2015	0	228
2015	1	4
2015	2	3
2015	3	15
2015	4	1
2015	5	1
2015	7	2

Race – 3rd mention

knitr::kable(Long_format_2005_2009_2015_new %>% group_by(year) %>% count(L7_3rd_mention), align = "ccc")

year	L7_3rd_mention	n
2005	0	348
2009	0	364
2009	3	3
2015	0	251
2015	1	2
2015	7	1

Widowed – year

knitr::kable(Long_format_2005_2009_2015_new %>% group_by(year) %>% count(D2D3_year), align = "ccc")

year	D2D3_year	n
2005	0	348
2009	0	367
2015	0	254

Employment status – 1st mention [1 = Working now, including military; 2 = Only temporarily laid off; sick or maternity leave; 3 = Looking for work, unemployed; 4 = Retired; 5 = Disabled, permanently or temporarily; 6 = Keeping house; 7 = Student; 8 = Other]

knitr::kable(Long_format_2005_2009_2015_new %>% group_by(year) %>% count(E1_1st_mention), align = "ccc")

year	E1_1st_mention	n
2005	1	163
2005	2	1
2005	3	38
2005	6	6
2005	7	136
2005	8	4
2009	1	149
2009	3	73
2009	6	5
2009	7	139
2009	99	1
2015	1	133
2015	2	1
2015	3	46
2015	6	7
2015	7	67

Employment status – 2nd mention

knitr::kable(Long_format_2005_2009_2015_new %>% group_by(year) %>% count(E1_2nd_mention), align = "ccc")

year	E1_2nd_mention	n
2005	0	261
2005	1	13
2005	3	5
2005	6	3
2005	7	66
2009	0	214
2009	1	19
2009	3	23
2009	6	4
2009	7	107
2015	0	158
2015	1	14
2015	3	10
2015	6	2
2015	7	70

Employment status – 3rd mention

knitr::kable(Long_format_2005_2009_2015_new %>% group_by(year) %>% count(E1_3rd_mention), align = "ccc")

year	E1_3rd_mention	n
2005	0	345
2005	1	1
2005	3	1
2005	7	1
2009	0	367
2015	0	254

Work for money [1 = Yes; 5 = No; 8 = DK; 9 = NA]

knitr::kable(Long_format_2005_2009_2015_new %>% group_by(year) %>% count(E3), align = "ccc")

year	E3	n
2005	0	177
2005	1	45
2005	5	125
2005	9	1
2009	0	168
2009	1	18
2009	5	181
2015	0	149
2015	1	6
2015	5	99

Education status [1 = Graduated from high school; 2 = Got a GED; 3 = Neither]

knitr::kable(Long_format_2005_2009_2015_new %>% group_by(year) %>% count(G1), align = "ccc")

year	G1	n
2005	1	348
2009	1	367
2015	1	254

Attended College [1 = Yes; 5 = No]

knitr::kable(Long_format_2005_2009_2015_new %>% group_by(year) %>% count(G10), align = "ccc")

year	G10	n
2005	1	255
2005	5	92
2005	9	1
2009	1	265
2009	5	102
2015	0	1
2015	1	173
2015	5	80

Attending College [1 = Yes; 5 = No]

knitr::kable(Long_format_2005_2009_2015_new %>% group_by(year) %>% count(G11), align = "ccc")

year	G11	n
2005	0	93
2005	1	221
2005	5	34
2009	0	102
2009	1	242
2009	5	23
2015	0	80
2015	1	139
2015	5	35

---

---

2005 & 2009 (within person)

t-test (2005 & 2009)

Filter the data (2005 & 2009)

Long_format_2005_2009 <- TAS_data_long_format_age %>% filter(year < 2010) %>% filter (TAS05 == 1) %>%  filter (TAS09 == 1) %>% mutate(year_new = case_when(year == 2005 ~ -1, year == 2009 ~ 0,year == 2015 ~ 1)) %>%  group_by(TAS_ID) %>% mutate(Age_18_graduate = case_when(Age_18_graduate == 2027 ~ Age_18_graduate[year == 2005] + 4, Age_18_graduate == 2023 ~ Age_18_graduate[year == 2009] - 4, TRUE ~ Age_18_graduate)) %>% ungroup() %>% filter (Age_18_graduate <100) %>%  group_by(TAS_ID) %>% mutate(age_difference = Age_18_graduate[year == 2009] - Age_18_graduate[year == 2005]) %>% filter(age_difference < 6) %>% filter(age_difference > 2) %>% ungroup()

view(Long_format_2005_2009)

knitr::kable(head(Long_format_2005_2009[, 1:43]))

TAS	TAS05	TAS09	TAS15	TAS_ID	Gender	Individual is sample	Year ID Number	Sequence Number	Relationship to Head	Release Number	B5A	B5D	B6C	C2D	C2E	C2F	D2D3_month	D2D3_year	E1_1st_mention	E1_2nd_mention	E1_3rd_mention	E3	G1	G2_month	G2_year	G10	G11	G30A	G41A	G41B	G41C	G41H	G41P	H1	L7_1st_mention	L7_2nd_mention	L7_3rd_mention	Age_17_graduate	Age_18_graduate	year	year_new	age_difference
2	1	1	NA	5_32	2	2	624	3	30	5	5	5	5	7	7	7	0	0	1	7	0	0	1	5	2002	1	1	7	7	6	6	7	5	2	1	0	0	20	21	2005	-1	4
2	1	1	NA	6_34	1	2	1202	51	30	5	2	2	6	1	1	1	0	0	7	0	0	5	1	5	2002	1	1	0	7	5	7	5	3	1	1	0	0	20	21	2005	-1	4
2	1	1	NA	14_30	1	2	736	51	30	5	4	4	4	2	1	1	0	0	2	0	0	0	1	6	2003	1	5	6	5	6	6	5	5	2	1	0	0	19	20	2005	-1	4
2	1	1	NA	47_34	2	2	2516	3	30	5	4	5	6	4	5	2	0	0	1	0	0	0	1	5	2005	5	0	6	3	6	4	7	4	1	1	0	0	17	18	2005	-1	4
2	1	1	NA	53_35	2	2	1392	3	33	5	4	5	5	3	1	1	0	0	1	0	0	0	1	6	2002	1	1	7	6	7	7	7	5	1	1	0	0	20	21	2005	-1	4
2	1	1	NA	53_36	2	2	1616	3	30	5	4	5	7	4	1	1	0	0	6	0	0	1	1	6	2005	1	5	7	7	7	5	7	6	2	1	0	0	17	18	2005	-1	4

knitr::kable(count(Long_format_2005_2009, year), align = "cc")

year	n
2005	542
2009	542

knitr::kable(count(Long_format_2005_2009, Age_18_graduate), align = "ccc")

Age_18_graduate	n
14	1
17	7
18	166
19	150
20	141
21	83
22	167
23	150
24	140
25	76
26	2
27	1

Age count - 2005

knitr::kable(Long_format_2005_2009 %>% filter(year == 2005) %>% count(Age_18_graduate), align = "cc")

Age_18_graduate	n
14	1
17	7
18	165
19	150
20	141
21	76
22	1
23	1

Age count - 2009

knitr::kable(Long_format_2005_2009 %>% filter(year == 2009) %>% count(Age_18_graduate), align = "cc")

Age_18_graduate	n
18	1
21	7
22	166
23	149
24	140
25	76
26	2
27	1

---

data_2005 <- Long_format_2005_2009 %>% filter(year_new == -1)
data_2009 <- Long_format_2005_2009 %>% filter(year_new == 0)

B5A: Responsibility for self

B5A_t_test_05_09 <- t.test(data_2005$B5A, data_2009$B5A, paired = TRUE)
B5A_t_test_05_09

## 
##  Paired t-test
## 
## data:  data_2005$B5A and data_2009$B5A
## t = -14.895, df = 541, p-value < 2.2e-16
## alternative hypothesis: true mean difference is not equal to 0
## 95 percent confidence interval:
##  -0.9376670 -0.7191595
## sample estimates:
## mean difference 
##      -0.8284133

knitr::kable(Long_format_2005_2009 %>% group_by(year) %>% count(B5A), align = "ccc")

year	B5A	n
2005	1	25
2005	2	105
2005	3	119
2005	4	167
2005	5	126
2009	1	11
2009	2	29
2009	3	47
2009	4	146
2009	5	309

mean_B5A_0509 <- Long_format_2005_2009 %>% group_by(year) %>% summarize(average_B5A = mean(B5A, na.rm = TRUE)) %>% ungroup()
knitr::kable((mean_B5A_0509), align = "cc")

year	average_B5A
2005	3.487085
2009	4.315498

sd_B5A_0509 <- Long_format_2005_2009 %>% group_by(year) %>% summarize(sd_B5A = sd(B5A, na.rm = TRUE)) %>% ungroup()
knitr::kable((sd_B5A_0509), align = "cc")

year	sd_B5A
2005	1.1753865
2009	0.9776143

Long_format_2005_2009 %>% ggplot(aes(x = factor(year), y = B5A, fill = factor(year))) + geom_boxplot() + stat_summary(fun = "mean", geom = "crossbar", width = 0.75, color = "black", size = 0.2, linetype = "dashed") + labs(title = "Boxplot of Responsibility for Self (B5A) in 2005 and 2009", x = "Year", y = "Responsibility for Self (B5A)") + scale_fill_manual(values = c("2005" = "skyblue", "2009" = "salmon")) + theme_minimal()

effect_size_B5A_05_09 <- cohens_d(data_2005$B5A, data_2009$B5A, paired = TRUE)

## For paired samples, 'repeated_measures_d()' provides more options.

effect_size_B5A_05_09

## Cohen's d |         95% CI
## --------------------------
## -0.64     | [-0.73, -0.55]

B6C: Money management skills

B6C_t_test_05_09 <- t.test(data_2005$B6C, data_2009$B6C, paired = TRUE)
B6C_t_test_05_09

## 
##  Paired t-test
## 
## data:  data_2005$B6C and data_2009$B6C
## t = -1.4667, df = 541, p-value = 0.143
## alternative hypothesis: true mean difference is not equal to 0
## 95 percent confidence interval:
##  -0.20285228  0.02942055
## sample estimates:
## mean difference 
##     -0.08671587

knitr::kable(Long_format_2005_2009 %>% group_by(year) %>% count(B6C), align = "ccc")

year	B6C	n
2005	1	9
2005	2	21
2005	3	30
2005	4	71
2005	5	163
2005	6	115
2005	7	133
2009	1	4
2009	2	10
2009	3	22
2009	4	88
2009	5	163
2009	6	134
2009	7	121

mean_B6C_0509 <- Long_format_2005_2009 %>% group_by(year) %>% summarize(average_B6C = mean(B6C, na.rm = TRUE)) %>% ungroup()
knitr::kable((mean_B6C_0509), align = "cc")

year	average_B6C
2005	5.278598
2009	5.365314

sd_B6C_0509 <- Long_format_2005_2009 %>% group_by(year) %>% summarize(sd_B6C = sd(B6C, na.rm = TRUE)) %>% ungroup()
knitr::kable((sd_B6C_0509), align = "cc")

year	sd_B6C
2005	1.444559
2009	1.272246

Long_format_2005_2009 %>% ggplot(aes(x = factor(year), y = B6C, fill = factor(year))) + geom_boxplot() + stat_summary(fun = "mean", geom = "crossbar", width = 0.75, color = "black", size = 0.2, linetype = "dashed") + labs(title = "Boxplot of Money management skills (B6C) in 2005 and 2009", x = "Year", y = "Money management skills (B6C)") + scale_fill_manual(values = c("2005" = "skyblue", "2009" = "salmon")) + theme_minimal()

effect_size_B6C_05_09 <- cohens_d(data_2005$B6C, data_2009$B6C, paired = TRUE)

## For paired samples, 'repeated_measures_d()' provides more options.

effect_size_B6C_05_09

## Cohen's d |        95% CI
## -------------------------
## -0.06     | [-0.15, 0.02]

C2D: Worry about expenses

C2D_t_test_05_09 <- t.test(data_2005$C2D, data_2009$C2D, paired = TRUE)
C2D_t_test_05_09

## 
##  Paired t-test
## 
## data:  data_2005$C2D and data_2009$C2D
## t = -3.0212, df = 541, p-value = 0.002636
## alternative hypothesis: true mean difference is not equal to 0
## 95 percent confidence interval:
##  -0.46582598 -0.09874967
## sample estimates:
## mean difference 
##      -0.2822878

knitr::kable(Long_format_2005_2009 %>% group_by(year) %>% count(C2D), align = "ccc")

year	C2D	n
2005	1	82
2005	2	87
2005	3	96
2005	4	98
2005	5	81
2005	6	48
2005	7	50
2009	1	60
2009	2	85
2009	3	90
2009	4	97
2009	5	82
2009	6	62
2009	7	66

mean_C2D_0509 <- Long_format_2005_2009 %>% group_by(year) %>% summarize(average_C2D = mean(C2D, na.rm = TRUE)) %>% ungroup()
knitr::kable((mean_C2D_0509), align = "cc")

year	average_C2D
2005	3.651292
2009	3.933579

sd_C2D_0509 <- Long_format_2005_2009 %>% group_by(year) %>% summarize(sd_C2D = sd(C2D, na.rm = TRUE)) %>% ungroup()
knitr::kable((sd_C2D_0509), align = "cc")

year	sd_C2D
2005	1.843756
2009	1.869894

ggplot(Long_format_2005_2009, aes(x = factor(year), y = C2D, fill = factor(year))) + geom_boxplot() + stat_summary(fun = "mean", geom = "crossbar", width = 0.75, color = "black", size = 0.2, linetype = "dashed") + labs(title = "Boxplot of Worry about expenses (C2D) in 2005 and 2009", x = "Year", y = "Worry about expenses (C2D)") + scale_fill_manual(values = c("2005" = "skyblue", "2009" = "salmon")) + theme_minimal()

effect_size_C2D_05_09 <- cohens_d(data_2005$C2D, data_2009$C2D, paired = TRUE)

## For paired samples, 'repeated_measures_d()' provides more options.

effect_size_C2D_05_09

## Cohen's d |         95% CI
## --------------------------
## -0.13     | [-0.21, -0.05]

C2E: Worry about future job

C2E_t_test_05_09 <- t.test(data_2005$C2E, data_2009$C2E, paired = TRUE)
C2E_t_test_05_09

## 
##  Paired t-test
## 
## data:  data_2005$C2E and data_2009$C2E
## t = -2.5726, df = 541, p-value = 0.01036
## alternative hypothesis: true mean difference is not equal to 0
## 95 percent confidence interval:
##  -0.42299611 -0.05670869
## sample estimates:
## mean difference 
##      -0.2398524

knitr::kable(Long_format_2005_2009 %>% group_by(year) %>% count(C2E), align = "ccc")

year	C2E	n
2005	1	103
2005	2	101
2005	3	89
2005	4	65
2005	5	86
2005	6	59
2005	7	39
2009	1	75
2009	2	100
2009	3	90
2009	4	90
2009	5	64
2009	6	67
2009	7	56

mean_C2E_0509 <- Long_format_2005_2009 %>% group_by(year) %>% summarize(average_C2E = mean(C2E, na.rm = TRUE)) %>% ungroup()
knitr::kable((mean_C2E_0509), align = "cc")

year	average_C2E
2005	3.485240
2009	3.725092

sd_C2E_0509 <- Long_format_2005_2009 %>% group_by(year) %>% summarize(sd_C2E = sd(C2E, na.rm = TRUE)) %>% ungroup()
knitr::kable((sd_C2E_0509), align = "cc")

year	sd_C2E
2005	1.898234
2009	1.903416

ggplot(Long_format_2005_2009, aes(x = factor(year), y = C2E, fill = factor(year))) + geom_boxplot() + stat_summary(fun = "mean", geom = "crossbar", width = 0.75, color = "black", size = 0.2, linetype = "dashed") + labs(title = "Boxplot of Worry about future job (C2E) in 2005 and 2009", x = "Year", y = "Worry about future job (C2E)") + scale_fill_manual(values = c("2005" = "skyblue", "2009" = "salmon")) + theme_minimal()

effect_size_C2E_05_09 <- cohens_d(data_2005$C2E, data_2009$C2E, paired = TRUE)

## For paired samples, 'repeated_measures_d()' provides more options.

effect_size_C2E_05_09

## Cohen's d |         95% CI
## --------------------------
## -0.11     | [-0.19, -0.03]

G41A: Importance of job status

G41A_t_test_05_09 <- t.test(data_2005$G41A, data_2009$G41A, paired = TRUE)
G41A_t_test_05_09

## 
##  Paired t-test
## 
## data:  data_2005$G41A and data_2009$G41A
## t = 8.1936, df = 541, p-value = 1.841e-15
## alternative hypothesis: true mean difference is not equal to 0
## 95 percent confidence interval:
##  0.4614852 0.7525369
## sample estimates:
## mean difference 
##       0.6070111

knitr::kable(Long_format_2005_2009 %>% group_by(year) %>% count(G41A), align = "ccc")

year	G41A	n
2005	1	22
2005	2	24
2005	3	37
2005	4	61
2005	5	127
2005	6	117
2005	7	154
2009	1	48
2009	2	42
2009	3	42
2009	4	84
2009	5	140
2009	6	85
2009	7	101

mean_G41A_0509 <- Long_format_2005_2009 %>% group_by(year) %>% summarize(average_G41A = mean(G41A, na.rm = TRUE)) %>% ungroup()
knitr::kable((mean_G41A_0509), align = "cc")

year	average_G41A
2005	5.239852
2009	4.632841

sd_G41A_0509 <- Long_format_2005_2009 %>% group_by(year) %>% summarize(sd_G41A = sd(G41A, na.rm = TRUE)) %>% ungroup()
knitr::kable((sd_G41A_0509), align = "cc")

year	sd_G41A
2005	1.653328
2009	1.831101

ggplot(Long_format_2005_2009, aes(x = factor(year), y = G41A, fill = factor(year))) + geom_boxplot() + stat_summary(fun = "mean", geom = "crossbar", width = 0.75, color = "black", size = 0.2, linetype = "dashed") + labs(title = "Boxplot of Importance of job status (G41A) in 2005 and 2009", x = "Year", y = "Importance of job status (G41A)") + scale_fill_manual(values = c("2005" = "skyblue", "2009" = "salmon")) + theme_minimal()

effect_size_G41A_05_09 <- cohens_d(data_2005$G41A, data_2009$G41A, paired = TRUE)

## For paired samples, 'repeated_measures_d()' provides more options.

effect_size_G41A_05_09

## Cohen's d |       95% CI
## ------------------------
## 0.35      | [0.27, 0.44]

Demographics (2005 & 2009) (within person)

Filter the data (2005 & 2009)

Long_format_2005_2009 <- TAS_data_long_format_age %>% filter(year < 2010) %>% filter (TAS05 == 1) %>%  filter (TAS09 == 1) %>% mutate(year_new = case_when(year == 2005 ~ -1, year == 2009 ~ 0,year == 2015 ~ 1)) %>%  group_by(TAS_ID) %>% mutate(Age_18_graduate = case_when(Age_18_graduate == 2027 ~ Age_18_graduate[year == 2005] + 4, Age_18_graduate == 2023 ~ Age_18_graduate[year == 2009] - 4, TRUE ~ Age_18_graduate)) %>% ungroup() %>% filter (Age_18_graduate <100) %>%  group_by(TAS_ID) %>% mutate(age_difference = Age_18_graduate[year == 2009] - Age_18_graduate[year == 2005]) %>% filter(age_difference < 6) %>% filter(age_difference > 2) %>% ungroup()

view(Long_format_2005_2009)

knitr::kable(head(Long_format_2005_2009[, 1:43]))

TAS	TAS05	TAS09	TAS15	TAS_ID	Gender	Individual is sample	Year ID Number	Sequence Number	Relationship to Head	Release Number	B5A	B5D	B6C	C2D	C2E	C2F	D2D3_month	D2D3_year	E1_1st_mention	E1_2nd_mention	E1_3rd_mention	E3	G1	G2_month	G2_year	G10	G11	G30A	G41A	G41B	G41C	G41H	G41P	H1	L7_1st_mention	L7_2nd_mention	L7_3rd_mention	Age_17_graduate	Age_18_graduate	year	year_new	age_difference
2	1	1	NA	5_32	2	2	624	3	30	5	5	5	5	7	7	7	0	0	1	7	0	0	1	5	2002	1	1	7	7	6	6	7	5	2	1	0	0	20	21	2005	-1	4
2	1	1	NA	6_34	1	2	1202	51	30	5	2	2	6	1	1	1	0	0	7	0	0	5	1	5	2002	1	1	0	7	5	7	5	3	1	1	0	0	20	21	2005	-1	4
2	1	1	NA	14_30	1	2	736	51	30	5	4	4	4	2	1	1	0	0	2	0	0	0	1	6	2003	1	5	6	5	6	6	5	5	2	1	0	0	19	20	2005	-1	4
2	1	1	NA	47_34	2	2	2516	3	30	5	4	5	6	4	5	2	0	0	1	0	0	0	1	5	2005	5	0	6	3	6	4	7	4	1	1	0	0	17	18	2005	-1	4
2	1	1	NA	53_35	2	2	1392	3	33	5	4	5	5	3	1	1	0	0	1	0	0	0	1	6	2002	1	1	7	6	7	7	7	5	1	1	0	0	20	21	2005	-1	4
2	1	1	NA	53_36	2	2	1616	3	30	5	4	5	7	4	1	1	0	0	6	0	0	1	1	6	2005	1	5	7	7	7	5	7	6	2	1	0	0	17	18	2005	-1	4

---

Demographics

Year

knitr::kable(count(Long_format_2005_2009, year), align = "cc")

year	n
2005	542
2009	542

Age

knitr::kable(count(Long_format_2005_2009, Age_18_graduate), align = "cc")

Age_18_graduate	n
14	1
17	7
18	166
19	150
20	141
21	83
22	167
23	150
24	140
25	76
26	2
27	1

Age by year(2005)

knitr::kable(Long_format_2005_2009 %>% filter(year == 2005) %>% count(Age_18_graduate), align = "cc")

Age_18_graduate	n
14	1
17	7
18	165
19	150
20	141
21	76
22	1
23	1

Age by year (2009)

knitr::kable(Long_format_2005_2009 %>% filter(year == 2009) %>% count(Age_18_graduate), align = "cc")

Age_18_graduate	n
18	1
21	7
22	166
23	149
24	140
25	76
26	2
27	1

Gender [1 = Male; 2 = Female]

knitr::kable(Long_format_2005_2009 %>% group_by(year) %>% count(Gender), align = "ccc")

year	Gender	n
2005	1	245
2005	2	297
2009	1	245
2009	2	297

Years graduated high school

knitr::kable(Long_format_2005_2009 %>% filter(G2_year != 0) %>% group_by(TAS_ID) %>% filter(n_distinct(year) == 2) %>% ungroup() %>% group_by(year) %>% count(G2_year) %>% ungroup(), align = "ccc")

year	G2_year	n
2005	2001	1
2005	2002	76
2005	2003	139
2005	2004	145
2005	2005	157
2009	2001	2
2009	2002	76
2009	2003	138
2009	2004	144
2009	2005	158

knitr::kable(Long_format_2005_2009 %>% filter(G2_year != 0) %>% group_by(TAS_ID) %>% filter(n_distinct(year) == 2) %>% ungroup() %>% count(year), align = "cc")

year	n
2005	518
2009	518

mean_year_graduation <- Long_format_2005_2009 %>% filter(G2_year != 0) %>% group_by(TAS_ID) %>% filter(n_distinct(year) == 2) %>% ungroup() %>% group_by(year) %>% summarize(mean_graduation = mean(G2_year, na.rm = TRUE)) %>% ungroup()
knitr::kable((mean_year_graduation), align = "cc")

year	mean_graduation
2005	2003.736
2009	2003.734

sd_year_graduation <- Long_format_2005_2009 %>% filter(G2_year != 0) %>% group_by(TAS_ID) %>% filter(n_distinct(year) == 2) %>% ungroup() %>% group_by(year) %>% summarize(sd_graduation = sd(G2_year, na.rm = TRUE)) %>% ungroup()
knitr::kable((sd_year_graduation), align = "cc")

year	sd_graduation
2005	1.052552
2009	1.060305

Race – 1st mention [1 = White; 2 = Black, African-American, or Negro; 3 = American Indian or Alaska Native; 4 = Asian; 5 = Native Hawaiian or Pacific Islander; 7 = Some other race; 8 = DK; 9 = NA]

knitr::kable(Long_format_2005_2009 %>% group_by(year) %>% count(L7_1st_mention), align = "ccc")

year	L7_1st_mention	n
2005	1	293
2005	2	208
2005	3	3
2005	4	6
2005	5	3
2005	7	6
2005	8	2
2005	9	21
2009	1	306
2009	2	206
2009	3	5
2009	4	7
2009	7	17
2009	9	1

Race – 2nd mention

knitr::kable(Long_format_2005_2009 %>% group_by(year) %>% count(L7_2nd_mention), align = "ccc")

year	L7_2nd_mention	n
2005	0	529
2005	1	1
2005	2	4
2005	3	4
2005	5	2
2005	7	2
2009	0	514
2009	1	1
2009	2	6
2009	3	17
2009	4	1
2009	7	3

Race – 3rd mention

knitr::kable(Long_format_2005_2009 %>% group_by(year) %>% count(L7_3rd_mention), align = "ccc")

year	L7_3rd_mention	n
2005	0	540
2005	3	1
2005	5	1
2009	0	536
2009	3	3
2009	5	1
2009	7	2

Widowed – year

knitr::kable(Long_format_2005_2009 %>% group_by(year) %>% count(D2D3_year), align = "ccc")

year	D2D3_year	n
2005	0	542
2009	0	538
2009	2006	2
2009	2007	1
2009	2009	1

Employment status – 1st mention [1 = Working now, including military; 2 = Only temporarily laid off; sick or maternity leave; 3 = Looking for work, unemployed; 4 = Retired; 5 = Disabled, permanently or temporarily; 6 = Keeping house; 7 = Student; 8 = Other]

knitr::kable(Long_format_2005_2009 %>% group_by(year) %>% count(E1_1st_mention), align = "ccc")

year	E1_1st_mention	n
2005	1	260
2005	2	3
2005	3	50
2005	5	1
2005	6	10
2005	7	213
2005	8	5
2009	1	376
2009	2	1
2009	3	77
2009	5	2
2009	6	20
2009	7	65
2009	8	1

Employment status – 2nd mention

knitr::kable(Long_format_2005_2009 %>% group_by(year) %>% count(E1_2nd_mention), align = "ccc")

year	E1_2nd_mention	n
2005	0	399
2005	1	27
2005	3	8
2005	5	1
2005	6	2
2005	7	105
2009	0	405
2009	1	23
2009	3	10
2009	5	2
2009	6	10
2009	7	91
2009	8	1

Employment status – 3rd mention

knitr::kable(Long_format_2005_2009 %>% group_by(year) %>% count(E1_3rd_mention), align = "ccc")

year	E1_3rd_mention	n
2005	0	538
2005	1	1
2005	3	1
2005	6	1
2005	7	1
2009	0	541
2009	1	1

Work for money [1 = Yes; 5 = No; 8 = DK; 9 = NA]

knitr::kable(Long_format_2005_2009 %>% group_by(year) %>% count(E3), align = "ccc")

year	E3	n
2005	0	290
2005	1	71
2005	5	180
2005	9	1
2009	0	401
2009	1	20
2009	5	121

Education status [1 = Graduated from high school; 2 = Got a GED; 3 = Neither]

knitr::kable(Long_format_2005_2009 %>% group_by(year) %>% count(G1), align = "ccc")

year	G1	n
2005	1	526
2005	2	1
2005	3	13
2005	9	2
2009	1	537
2009	2	4
2009	3	1

Attended College [1 = Yes; 5 = No]

knitr::kable(Long_format_2005_2009 %>% group_by(year) %>% count(G10), align = "ccc")

year	G10	n
2005	0	14
2005	1	414
2005	5	113
2005	9	1
2009	0	1
2009	1	443
2009	5	98

Attending College [1 = Yes; 5 = No]

knitr::kable(Long_format_2005_2009 %>% group_by(year) %>% count(G11), align = "ccc")

year	G11	n
2005	0	128
2005	1	347
2005	5	67
2009	0	99
2009	1	166
2009	5	277

---

---

2009 & 2015 (within person)

t-test (2009 & 2015)

Filter the data (2009 & 2015)

Long_format_2009_2015 <- TAS_data_long_format_age %>% filter (TAS09 == 1) %>%  filter (TAS15 == 1) %>% mutate(year_new = case_when(year == 2005 ~ -1, year == 2009 ~ 0,year == 2015 ~ 1)) %>%  group_by(TAS_ID) %>% mutate(Age_18_graduate = case_when(Age_18_graduate == 2033 ~ Age_18_graduate[year == 2009] + 6, Age_18_graduate == 2027 ~ Age_18_graduate[year == 2015] - 6, TRUE ~ Age_18_graduate)) %>% ungroup() %>% filter (Age_18_graduate <100) %>%  group_by(TAS_ID) %>% mutate(age_difference = Age_18_graduate[year == 2015] - Age_18_graduate[year == 2009]) %>% filter(age_difference < 8) %>% filter(age_difference > 4) %>% ungroup()

view(Long_format_2009_2015)

knitr::kable(head(Long_format_2009_2015[, 1:43]))

TAS	TAS05	TAS09	TAS15	TAS_ID	Gender	Individual is sample	Year ID Number	Sequence Number	Relationship to Head	Release Number	B5A	B5D	B6C	C2D	C2E	C2F	D2D3_month	D2D3_year	E1_1st_mention	E1_2nd_mention	E1_3rd_mention	E3	G1	G2_month	G2_year	G10	G11	G30A	G41A	G41B	G41C	G41H	G41P	H1	L7_1st_mention	L7_2nd_mention	L7_3rd_mention	Age_17_graduate	Age_18_graduate	year	year_new	age_difference
2	NA	1	1	4_39	2	2	13	3	60	3	4	5	4	6	7	5	0	0	1	0	0	0	1	5	2008	1	5	5	6	5	2	7	6	2	1	0	0	18	19	2009	0	6
2	NA	1	1	7_40	2	2	3836	2	22	3	2	2	7	7	3	4	0	0	6	0	0	5	1	6	2007	5	0	5	5	2	5	6	5	3	1	0	0	19	20	2009	0	6
2	NA	1	1	7_41	1	2	576	2	30	3	3	4	7	4	5	4	0	0	3	0	0	5	1	5	2009	5	0	7	5	6	5	7	5	2	1	0	0	17	18	2009	0	6
2	NA	1	1	10_34	2	2	3276	3	30	3	4	5	6	4	1	1	0	0	1	0	0	0	1	6	2008	1	5	7	7	5	4	7	5	2	2	0	0	18	19	2009	0	6
2	NA	1	1	14_31	2	2	713	1	10	3	5	5	7	4	4	4	0	0	1	0	0	0	1	6	2005	5	0	6	5	7	6	7	2	4	1	0	0	21	22	2009	0	6
2	NA	1	1	22_30	2	2	907	2	30	3	5	1	4	3	1	1	0	0	1	0	0	0	1	5	2006	1	1	7	6	6	6	6	6	1	2	0	0	20	21	2009	0	6

knitr::kable(count(Long_format_2009_2015, year), align = "cc")

year	n
2009	515
2015	515

knitr::kable(count(Long_format_2009_2015, Age_18_graduate), align = "ccc")

Age_18_graduate	n
15	1
18	154
19	134
20	136
21	89
22	2
24	155
25	134
26	136
27	87
28	2

Age count - 2009

knitr::kable(Long_format_2009_2015 %>% filter(year == 2009) %>% count(Age_18_graduate), align = "cc")

Age_18_graduate	n
15	1
18	154
19	134
20	136
21	88
22	2

Age count - 2015

knitr::kable(Long_format_2009_2015 %>% filter(year == 2015) %>% count(Age_18_graduate), align = "cc")

Age_18_graduate	n
21	1
24	155
25	134
26	136
27	87
28	2

---

data_2009 <- Long_format_2009_2015 %>% filter(year_new == 0)
data_2015 <- Long_format_2009_2015 %>% filter(year_new == 1)

B5A: Responsibility for self (SPECIAL)

knitr::kable(Long_format_2009_2015 %>% group_by(year) %>% count(B5A), align = "ccc")

year	B5A	n
2009	1	34
2009	2	94
2009	3	113
2009	4	157
2009	5	116
2009	9	1
2015	1	12
2015	2	15
2015	3	35
2015	4	92
2015	5	359
2015	8	2

remove_id_B5A_09 <- Long_format_2009_2015 %>% filter(year_new == 0,B5A > 7) %>% pull(TAS_ID)
remove_id_B5A_15 <- Long_format_2009_2015 %>% filter(year_new == 1, B5A > 7) %>% pull(TAS_ID)
data_2009_B5A <- Long_format_2009_2015 %>% filter(year_new == 0) %>% filter(B5A < 7) %>% filter(!(TAS_ID %in% remove_id_B5A_15))
data_2015_B5A <- Long_format_2009_2015 %>% filter(year_new == 1) %>% filter(B5A < 7) %>% filter(!(TAS_ID %in% remove_id_B5A_09))

B5A_t_test_09_15 <- t.test(data_2009_B5A$B5A, data_2015_B5A$B5A, paired = TRUE)
B5A_t_test_09_15

## 
##  Paired t-test
## 
## data:  data_2009_B5A$B5A and data_2015_B5A$B5A
## t = -17.111, df = 511, p-value < 2.2e-16
## alternative hypothesis: true mean difference is not equal to 0
## 95 percent confidence interval:
##  -1.1823174 -0.9387764
## sample estimates:
## mean difference 
##       -1.060547

knitr::kable(Long_format_2009_2015 %>% filter(B5A < 7) %>% group_by(TAS_ID) %>% filter(n_distinct(year) == 2) %>% ungroup() %>% group_by(year) %>% count(B5A), align = "ccc")

year	B5A	n
2009	1	34
2009	2	94
2009	3	111
2009	4	157
2009	5	116
2015	1	12
2015	2	15
2015	3	35
2015	4	91
2015	5	359

knitr::kable(Long_format_2009_2015 %>% filter(B5A < 7) %>% group_by(TAS_ID) %>% filter(n_distinct(year) == 2) %>% ungroup() %>% count(year), align = "cc")

year	n
2009	512
2015	512

mean_B5A_0915 <- Long_format_2009_2015 %>% filter(B5A < 7) %>% group_by(TAS_ID) %>% filter(n_distinct(year) == 2) %>% ungroup() %>% group_by(year) %>% summarize(average_B5A = mean(B5A, na.rm = TRUE)) %>% ungroup()
knitr::kable((mean_B5A_0915), align = "cc")

year	average_B5A
2009	3.443359
2015	4.503906

sd_B5A_0915 <- Long_format_2009_2015 %>% filter(B5A < 7) %>% group_by(TAS_ID) %>% filter(n_distinct(year) == 2) %>% ungroup() %>% group_by(year) %>% summarize(sd_B5A = sd(B5A, na.rm = TRUE)) %>% ungroup()
knitr::kable((sd_B5A_0915), align = "cc")

year	sd_B5A
2009	1.2117796
2015	0.9194487

Long_format_2009_2015 %>% filter(B5A < 7) %>% group_by(TAS_ID) %>% filter(n_distinct(year) == 2) %>% ungroup() %>% ggplot(aes(x = factor(year), y = B5A, fill = factor(year))) + geom_boxplot() + stat_summary(fun = "mean", geom = "crossbar", width = 0.75, color = "black", size = 0.2, linetype = "dashed") + labs(title = "Boxplot of Responsibility for Self (B5A) in 2009 and 2015", x = "Year", y = "Responsibility for Self (B5A)") + scale_fill_manual(values = c("2009" = "skyblue", "2015" = "salmon")) + theme_minimal()

effect_size_B5A_09_15 <- cohens_d(data_2009_B5A$B5A, data_2015_B5A$B5A, paired = TRUE)

## For paired samples, 'repeated_measures_d()' provides more options.

effect_size_B5A_09_15

## Cohen's d |         95% CI
## --------------------------
## -0.76     | [-0.85, -0.66]

B6C: Money management skills

B6C_t_test_09_15 <- t.test(data_2009$B6C, data_2015$B6C, paired = TRUE)
B6C_t_test_09_15

## 
##  Paired t-test
## 
## data:  data_2009$B6C and data_2015$B6C
## t = 1.5235, df = 514, p-value = 0.1282
## alternative hypothesis: true mean difference is not equal to 0
## 95 percent confidence interval:
##  -0.02586079  0.20450156
## sample estimates:
## mean difference 
##      0.08932039

knitr::kable(Long_format_2009_2015 %>% group_by(year) %>% count(B6C), align = "ccc")

year	B6C	n
2009	1	4
2009	2	11
2009	3	17
2009	4	62
2009	5	154
2009	6	139
2009	7	128
2015	1	4
2015	2	7
2015	3	27
2015	4	74
2015	5	140
2015	6	157
2015	7	106

mean_B6C_0915 <- Long_format_2009_2015 %>% group_by(year) %>% summarize(average_B6C = mean(B6C, na.rm = TRUE)) %>% ungroup()
knitr::kable((mean_B6C_0915), align = "cc")

year	average_B6C
2009	5.485437
2015	5.396116

sd_B6C_0915 <- Long_format_2009_2015 %>% group_by(year) %>% summarize(sd_B6C = sd(B6C, na.rm = TRUE)) %>% ungroup()
knitr::kable((sd_B6C_0915), align = "cc")

year	sd_B6C
2009	1.265096
2015	1.254713

ggplot(Long_format_2009_2015, aes(x = factor(year), y = B6C, fill = factor(year))) + geom_boxplot() + stat_summary(fun = "mean", geom = "crossbar", width = 0.75, color = "black", size = 0.2, linetype = "dashed") + labs(title = "Boxplot of Money management skills (B6C) in 2009 and 2015", x = "Year", y = "Money management skills (B6C)") + scale_fill_manual(values = c("2009" = "skyblue", "2015" = "salmon")) + theme_minimal()

effect_size_B6C_09_15 <- cohens_d(data_2009$B6C, data_2015$B6C, paired = TRUE)

## For paired samples, 'repeated_measures_d()' provides more options.

effect_size_B6C_09_15

## Cohen's d |        95% CI
## -------------------------
## 0.07      | [-0.02, 0.15]

C2D: Worry about expenses

C2D_t_test_09_15 <- t.test(data_2009$C2D, data_2015$C2D, paired = TRUE)
C2D_t_test_09_15

## 
##  Paired t-test
## 
## data:  data_2009$C2D and data_2015$C2D
## t = 3.2447, df = 514, p-value = 0.001252
## alternative hypothesis: true mean difference is not equal to 0
## 95 percent confidence interval:
##  0.1256330 0.5112602
## sample estimates:
## mean difference 
##       0.3184466

knitr::kable(Long_format_2009_2015 %>% group_by(year) %>% count(C2D), align = "ccc")

year	C2D	n
2009	1	77
2009	2	97
2009	3	65
2009	4	89
2009	5	79
2009	6	49
2009	7	59
2015	1	98
2015	2	99
2015	3	84
2015	4	78
2015	5	73
2015	6	46
2015	7	37

mean_C2D_0915 <- Long_format_2009_2015 %>% group_by(year) %>% summarize(average_C2D = mean(C2D, na.rm = TRUE)) %>% ungroup()
knitr::kable((mean_C2D_0915), align = "cc")

year	average_C2D
2009	3.735922
2015	3.417476

sd_C2D_0915 <- Long_format_2009_2015 %>% group_by(year) %>% summarize(sd_C2D = sd(C2D, na.rm = TRUE)) %>% ungroup()
knitr::kable((sd_C2D_0915), align = "cc")

year	sd_C2D
2009	1.930749
2015	1.859481

ggplot(Long_format_2009_2015, aes(x = factor(year), y = C2D, fill = factor(year))) + geom_boxplot() + stat_summary(fun = "mean", geom = "crossbar", width = 0.75, color = "black", size = 0.2, linetype = "dashed") + labs(title = "Boxplot of Worry about expenses (C2D) in 2009 and 2015", x = "Year", y = "Worry about expenses (C2D)") + scale_fill_manual(values = c("2009" = "skyblue", "2015" = "salmon")) + theme_minimal()

effect_size_C2D_09_15 <- cohens_d(data_2009$C2D, data_2015$C2D, paired = TRUE)

## For paired samples, 'repeated_measures_d()' provides more options.

effect_size_C2D_09_15

## Cohen's d |       95% CI
## ------------------------
## 0.14      | [0.06, 0.23]

C2E: Worry about future job

C2E_t_test_09_15 <- t.test(data_2009$C2E, data_2015$C2E, paired = TRUE)
C2E_t_test_09_15

## 
##  Paired t-test
## 
## data:  data_2009$C2E and data_2015$C2E
## t = 5.1874, df = 514, p-value = 3.073e-07
## alternative hypothesis: true mean difference is not equal to 0
## 95 percent confidence interval:
##  0.2907333 0.6451890
## sample estimates:
## mean difference 
##       0.4679612

knitr::kable(Long_format_2009_2015 %>% group_by(year) %>% count(C2E), align = "ccc")

year	C2E	n
2009	1	89
2009	2	89
2009	3	77
2009	4	86
2009	5	76
2009	6	36
2009	7	62
2015	1	109
2015	2	111
2015	3	96
2015	4	74
2015	5	67
2015	6	25
2015	7	33

mean_C2E_0915 <- Long_format_2009_2015 %>% group_by(year) %>% summarize(average_C2E = mean(C2E, na.rm = TRUE)) %>% ungroup()
knitr::kable((mean_C2E_0915), align = "cc")

year	average_C2E
2009	3.634952
2015	3.166990

sd_C2E_0915 <- Long_format_2009_2015 %>% group_by(year) %>% summarize(sd_C2E = sd(C2E, na.rm = TRUE)) %>% ungroup()
knitr::kable((sd_C2E_0915), align = "cc")

year	sd_C2E
2009	1.944448
2015	1.779498

ggplot(Long_format_2009_2015, aes(x = factor(year), y = C2E, fill = factor(year))) + geom_boxplot() + stat_summary(fun = "mean", geom = "crossbar", width = 0.75, color = "black", size = 0.2, linetype = "dashed") + labs(title = "Boxplot of Worry about future job (C2E) in 2009 and 2015", x = "Year", y = "Worry about future job (C2E)") + scale_fill_manual(values = c("2009" = "skyblue", "2015" = "salmon")) + theme_minimal()

effect_size_C2E_09_15 <- cohens_d(data_2009$C2E, data_2015$C2E, paired = TRUE)

## For paired samples, 'repeated_measures_d()' provides more options.

effect_size_C2E_09_15

## Cohen's d |       95% CI
## ------------------------
## 0.23      | [0.14, 0.32]

G41A: Importance of job status (SPECIAL)

knitr::kable(Long_format_2009_2015 %>% group_by(year) %>% count(G41A), align = "ccc")

year	G41A	n
2009	1	30
2009	2	14
2009	3	39
2009	4	70
2009	5	128
2009	6	89
2009	7	144
2009	9	1
2015	1	54
2015	2	50
2015	3	50
2015	4	73
2015	5	115
2015	6	75
2015	7	98

remove_id_G41A <- Long_format_2009_2015 %>% filter(year_new == 0,G41A == 9) %>% pull(TAS_ID)
data_2009_G41A <- Long_format_2009_2015 %>% filter(year_new == 0) %>% filter(G41A != 9) 
data_2015_G41A <- Long_format_2009_2015 %>% filter(year_new == 1) %>% filter(!(TAS_ID %in% remove_id_G41A))

knitr::kable(Long_format_2009_2015 %>% filter(G41A != 9) %>% group_by(TAS_ID) %>% filter(n_distinct(year) == 2) %>% ungroup() %>% group_by(year) %>% count(G41A), align = "ccc")

year	G41A	n
2009	1	30
2009	2	14
2009	3	39
2009	4	70
2009	5	128
2009	6	89
2009	7	144
2015	1	54
2015	2	50
2015	3	50
2015	4	73
2015	5	115
2015	6	74
2015	7	98

knitr::kable(Long_format_2009_2015 %>% filter(G41A != 9) %>% group_by(TAS_ID) %>% filter(n_distinct(year) == 2) %>% ungroup() %>% count(year), align = "cc")

year	n
2009	514
2015	514

G41A_t_test_09_15 <- t.test(data_2009_G41A$G41A, data_2015_G41A$G41A, paired = TRUE)
G41A_t_test_09_15

## 
##  Paired t-test
## 
## data:  data_2009_G41A$G41A and data_2015_G41A$G41A
## t = 8.2779, df = 513, p-value = 1.093e-15
## alternative hypothesis: true mean difference is not equal to 0
## 95 percent confidence interval:
##  0.4985536 0.8088394
## sample estimates:
## mean difference 
##       0.6536965

mean_G41A_0915 <- Long_format_2009_2015 %>% filter(G41A != 9) %>% group_by(TAS_ID) %>% filter(n_distinct(year) == 2) %>% ungroup() %>% group_by(year) %>% summarize(average_G41A = mean(G41A, na.rm = TRUE)) %>% ungroup()
knitr::kable((mean_G41A_0915), align = "cc")

year	average_G41A
2009	5.130350
2015	4.476654

sd_G41A_0915 <- Long_format_2009_2015 %>% filter(G41A != 9) %>% group_by(TAS_ID) %>% filter(n_distinct(year) == 2) %>% ungroup() %>% group_by(year) %>% summarize(sd_G41A = sd(G41A, na.rm = TRUE)) %>% ungroup()
knitr::kable((sd_G41A_0915), align = "cc")

year	sd_G41A
2009	1.703263
2015	1.930680

Long_format_2009_2015 %>% filter(G41A != 9) %>% group_by(TAS_ID) %>% filter(n_distinct(year) == 2) %>% ungroup() %>% ggplot(aes(x = factor(year), y = G41A, fill = factor(year))) + geom_boxplot() + stat_summary(fun = "mean", geom = "crossbar", width = 0.75, color = "black", size = 0.2, linetype = "dashed") + labs(title = "Boxplot of Importance of job status (G41A) in 2009 and 2015", x = "Year", y = "Importance of job status (G41A)") + scale_fill_manual(values = c("2009" = "skyblue", "2015" = "salmon")) + theme_minimal()

effect_size_G41A_09_15 <- cohens_d(data_2009_G41A$G41A, data_2015_G41A$G41A, paired = TRUE)

## For paired samples, 'repeated_measures_d()' provides more options.

effect_size_G41A_09_15

## Cohen's d |       95% CI
## ------------------------
## 0.37      | [0.28, 0.45]

##2009 & 2015

Demographics (2009 & 2015) (within person)

Filter the data (2009 & 2015)

Long_format_2009_2015 <- TAS_data_long_format_age %>% filter (TAS09 == 1) %>%  filter (TAS15 == 1) %>% mutate(year_new = case_when(year == 2005 ~ -1, year == 2009 ~ 0,year == 2015 ~ 1)) %>%  group_by(TAS_ID) %>% mutate(Age_18_graduate = case_when(Age_18_graduate == 2033 ~ Age_18_graduate[year == 2009] + 6, Age_18_graduate == 2027 ~ Age_18_graduate[year == 2015] - 6, TRUE ~ Age_18_graduate)) %>% ungroup() %>% filter (Age_18_graduate <100) %>%  group_by(TAS_ID) %>% mutate(age_difference = Age_18_graduate[year == 2015] - Age_18_graduate[year == 2009]) %>% filter(age_difference < 8) %>% filter(age_difference > 4) %>% ungroup()

view(Long_format_2009_2015)

knitr::kable(head(Long_format_2009_2015[, 1:43]))

TAS	TAS05	TAS09	TAS15	TAS_ID	Gender	Individual is sample	Year ID Number	Sequence Number	Relationship to Head	Release Number	B5A	B5D	B6C	C2D	C2E	C2F	D2D3_month	D2D3_year	E1_1st_mention	E1_2nd_mention	E1_3rd_mention	E3	G1	G2_month	G2_year	G10	G11	G30A	G41A	G41B	G41C	G41H	G41P	H1	L7_1st_mention	L7_2nd_mention	L7_3rd_mention	Age_17_graduate	Age_18_graduate	year	year_new	age_difference
2	NA	1	1	4_39	2	2	13	3	60	3	4	5	4	6	7	5	0	0	1	0	0	0	1	5	2008	1	5	5	6	5	2	7	6	2	1	0	0	18	19	2009	0	6
2	NA	1	1	7_40	2	2	3836	2	22	3	2	2	7	7	3	4	0	0	6	0	0	5	1	6	2007	5	0	5	5	2	5	6	5	3	1	0	0	19	20	2009	0	6
2	NA	1	1	7_41	1	2	576	2	30	3	3	4	7	4	5	4	0	0	3	0	0	5	1	5	2009	5	0	7	5	6	5	7	5	2	1	0	0	17	18	2009	0	6
2	NA	1	1	10_34	2	2	3276	3	30	3	4	5	6	4	1	1	0	0	1	0	0	0	1	6	2008	1	5	7	7	5	4	7	5	2	2	0	0	18	19	2009	0	6
2	NA	1	1	14_31	2	2	713	1	10	3	5	5	7	4	4	4	0	0	1	0	0	0	1	6	2005	5	0	6	5	7	6	7	2	4	1	0	0	21	22	2009	0	6
2	NA	1	1	22_30	2	2	907	2	30	3	5	1	4	3	1	1	0	0	1	0	0	0	1	5	2006	1	1	7	6	6	6	6	6	1	2	0	0	20	21	2009	0	6

---

Demographics

Year

knitr::kable(count(Long_format_2009_2015, year), align = "cc")

year	n
2009	515
2015	515

Age

knitr::kable(count(Long_format_2009_2015, Age_18_graduate), align = "cc")

Age_18_graduate	n
15	1
18	154
19	134
20	136
21	89
22	2
24	155
25	134
26	136
27	87
28	2

Age by year(2009)

knitr::kable(Long_format_2009_2015 %>% filter(year == 2009) %>% count(Age_18_graduate), align = "cc")

Age_18_graduate	n
15	1
18	154
19	134
20	136
21	88
22	2

Age by year (2015)

knitr::kable(Long_format_2009_2015 %>% filter(year == 2015) %>% count(Age_18_graduate), align = "cc")

Age_18_graduate	n
21	1
24	155
25	134
26	136
27	87
28	2

Gender [1 = Male; 2 = Female]

knitr::kable(Long_format_2009_2015 %>% group_by(year) %>% count(Gender), align = "ccc")

year	Gender	n
2009	1	219
2009	2	296
2015	1	219
2015	2	296

Years graduated high school

knitr::kable(Long_format_2009_2015 %>% filter(G2_year != 0) %>% group_by(TAS_ID) %>% filter(n_distinct(year) == 2) %>% ungroup() %>% group_by(year) %>% count(G2_year) %>% ungroup(), align = "ccc") 

year	G2_year	n
2009	2006	14
2009	2007	29
2009	2008	32
2009	2009	49
2015	2006	13
2015	2007	29
2015	2008	32
2015	2009	50

knitr::kable(Long_format_2009_2015 %>% filter(G2_year != 0) %>% group_by(TAS_ID) %>% filter(n_distinct(year) == 2) %>% ungroup() %>% count(year), align = "ccc") 

year	n
2009	124
2015	124

mean_year_graduation <- Long_format_2009_2015 %>% filter(G2_year != 0) %>% group_by(TAS_ID) %>% filter(n_distinct(year) == 2) %>% ungroup() %>% group_by(year) %>% summarize(mean_graduation = mean(G2_year, na.rm = TRUE)) %>% ungroup()
knitr::kable((mean_year_graduation), align = "cc")

year	mean_graduation
2009	2007.935
2015	2007.960

sd_year_graduation <- Long_format_2009_2015 %>% filter(G2_year != 0) %>% group_by(TAS_ID) %>% filter(n_distinct(year) == 2) %>% ungroup() %>% group_by(year) %>% summarize(sd_graduation = sd(G2_year, na.rm = TRUE)) %>% ungroup()
knitr::kable((sd_year_graduation), align = "cc")

year	sd_graduation
2009	1.041746
2015	1.031214

Race – 1st mention [1 = White; 2 = Black, African-American, or Negro; 3 = American Indian or Alaska Native; 4 = Asian; 5 = Native Hawaiian or Pacific Islander; 7 = Some other race; 8 = DK; 9 = NA]

knitr::kable(Long_format_2009_2015 %>% group_by(year) %>% count(L7_1st_mention), align = "ccc")

year	L7_1st_mention	n
2009	1	261
2009	2	198
2009	3	7
2009	4	10
2009	7	37
2009	8	1
2009	9	1
2015	1	274
2015	2	195
2015	3	8
2015	4	8
2015	7	26
2015	9	4

Race – 2nd mention

knitr::kable(Long_format_2009_2015 %>% group_by(year) %>% count(L7_2nd_mention), align = "ccc")

year	L7_2nd_mention	n
2009	0	489
2009	1	8
2009	2	3
2009	3	10
2009	4	1
2009	5	2
2009	7	2
2015	0	497
2015	1	4
2015	2	4
2015	3	8
2015	4	2

Race – 3rd mention

knitr::kable(Long_format_2009_2015 %>% group_by(year) %>% count(L7_3rd_mention), align = "ccc")

year	L7_3rd_mention	n
2009	0	513
2009	3	2
2015	0	513
2015	1	1
2015	3	1

Widowed – year

knitr::kable(Long_format_2009_2015 %>% group_by(year) %>% count(D2D3_year), align = "ccc")

year	D2D3_year	n
2009	0	512
2009	2009	3
2015	0	503
2015	2009	2
2015	2012	3
2015	2014	7

Employment status – 1st mention [1 = Working now, including military; 2 = Only temporarily laid off; sick or maternity leave; 3 = Looking for work, unemployed; 4 = Retired; 5 = Disabled, permanently or temporarily; 6 = Keeping house; 7 = Student; 8 = Other]

knitr::kable(Long_format_2009_2015 %>% group_by(year) %>% count(E1_1st_mention), align = "ccc")

year	E1_1st_mention	n
2009	1	242
2009	3	93
2009	6	12
2009	7	167
2009	99	1
2015	1	406
2015	2	2
2015	3	49
2015	5	4
2015	6	22
2015	7	32

Employment status – 2nd mention

knitr::kable(Long_format_2009_2015 %>% group_by(year) %>% count(E1_2nd_mention), align = "ccc")

year	E1_2nd_mention	n
2009	0	301
2009	1	35
2009	2	1
2009	3	28
2009	6	3
2009	7	147
2015	0	445
2015	1	6
2015	3	4
2015	6	16
2015	7	44

E1: Employment status – 3rd mention

knitr::kable(Long_format_2009_2015 %>% group_by(year) %>% count(E1_3rd_mention), align = "ccc")

year	E1_3rd_mention	n
2009	0	514
2009	6	1
2015	0	515

Work for money [1 = Yes; 5 = No; 8 = DK; 9 = NA]

knitr::kable(Long_format_2009_2015 %>% group_by(year) %>% count(E3), align = "ccc")

year	E3	n
2009	0	277
2009	1	31
2009	5	207
2015	0	414
2015	1	8
2015	5	93

Education status [1 = Graduated from high school; 2 = Got a GED; 3 = Neither]

knitr::kable(Long_format_2009_2015 %>% group_by(year) %>% count(G1), align = "ccc")

year	G1	n
2009	1	511
2009	2	3
2009	3	1
2015	0	370
2015	1	139
2015	2	5
2015	3	1

Attended College [1 = Yes; 5 = No]

knitr::kable(Long_format_2009_2015 %>% group_by(year) %>% count(G10), align = "ccc")

year	G10	n
2009	0	1
2009	1	385
2009	5	129
2015	0	403
2015	1	35
2015	5	77

Attending College [1 = Yes; 5 = No]

knitr::kable(Long_format_2009_2015 %>% group_by(year) %>% count(G11), align = "ccc")

year	G11	n
2009	0	130
2009	1	323
2009	5	62
2015	0	79
2015	1	86
2015	5	350

---

2005 vs 2009 vs 2015 (between person; 18-19 y/o)

df_long <- bind_rows(
  mutate(B5A_means, Variable = "B5A"),
  mutate(B6C_means, Variable = "B6C"),
  mutate(C2D_means, Variable = "C2D"),
  mutate(C2E_means, Variable = "C2E"),
  mutate(G41A_means, Variable = "G41A"))
year_colors <- c("2005" = "#990000", "2009" = "#0A2240", "2015" = "#BEBEBE")

knitr::kable(df_long, align = "ccccc")

year_new	Mean	SD	Count	Variable
2005	3.376405	1.105125	356	B5A
2009	3.282609	1.151481	368	B5A
2015	3.456693	1.136918	254	B5A
2005	5.390449	1.434708	356	B6C
2009	5.520325	1.231491	369	B6C
2015	5.366142	1.370134	254	B6C
2005	3.702247	1.884285	356	C2D
2009	3.661247	1.925696	369	C2D
2015	3.476378	1.853699	254	C2D
2005	3.567416	1.945296	356	C2E
2009	3.628726	1.974109	369	C2E
2015	3.629921	1.975355	254	C2E
2005	5.362360	1.627454	356	G41A
2009	5.331522	1.530395	368	G41A
2015	4.874016	1.728588	254	G41A

ggplot(df_long, aes(x = Variable, y = Mean, fill = year_new, group = year_new)) + geom_bar(stat = "identity", position = position_dodge(width = 0.8), color = "black", width = 0.5) + geom_errorbar(aes(ymin = Mean - SD, ymax = Mean + SD), position = position_dodge(width = 0.8), width = 0.25, colour = "#000000", size = 1) + scale_fill_manual(values = year_colors) + labs(title = "2005 vs 2009 vs 2015 (between person; 18-19 y/o)", x = "Variables", y = "Mean Value", fill = "Year") + theme_minimal() + theme(axis.text.x = element_text(angle = 0, hjust = 0.5, vjust = 0.5), axis.ticks.length = unit(0.3, "cm")) + theme(panel.grid.major = element_blank(),panel.grid.minor = element_blank(),plot.title = element_text(size = 16, face = "bold"),axis.title.x = element_text(vjust = -2)) + scale_x_discrete(labels = c("B5A" = "B5A: \n Responsibility \n for Self", "B6C" = "B6C:\n Money Management \n Skills", "C2D" = "C2D:\n Worry about \n Expenses", "C2E" = "C2E:\n Worry about \n Future Job", "G41A" = "G41A:\n Importance of \n Job Status")) + scale_y_continuous(breaks = seq(1, 7, by = 1), expand = c(0, 0)) + coord_cartesian(ylim = c(1, 9))  + 
 
   geom_segment(aes(x = 5 - 0.28, xend = 5 + 0.28, y = 8.3, yend = 8.3), color = "black") + 
  annotate("text", x = 5, y = 8.65, label = "p = 0.001", size = 3.2) + 
  geom_segment(aes(x = 5 - 0.28, xend = 5 - 0.28, y = 8.3, yend = 8), color = "black") + 
  geom_segment(aes(x = 5 + 0.28, xend = 5 + 0.28, y = 8.3, yend = 8), color = "black") +
  
  geom_segment(aes(x = 5.15 - 0.17, xend = 5.15 + 0.17, y = 7.4, yend = 7.4), color = "black") + 
  annotate("text", x = 5.15, y = 7.7, label = "p = 0.002", size = 3.2) + 
  geom_segment(aes(x = 5.15 - 0.17, xend = 5.15 - 0.17, y = 7.4, yend = 7.1), color = "black") + 
  geom_segment(aes(x = 5.15 + 0.17, xend = 5.15 + 0.17, y = 7.4, yend = 7.1), color = "black")

2005 vs 2009 (within person)

df_long_0509_mess <- bind_rows(mutate(mean_B5A_0509, SD = sd_B5A_0509$sd_B5A, Variable = "B5A"), mutate(mean_B6C_0509, SD = sd_B6C_0509$sd_B6C, Variable = "B6C"), mutate(mean_C2D_0509, SD = sd_C2D_0509$sd_C2D, Variable = "C2D"), mutate(mean_C2E_0509, SD = sd_C2E_0509$sd_C2E, Variable = "C2E"), mutate(mean_G41A_0509, SD = sd_G41A_0509$sd_G41A, Variable = "G41A")) %>% gather(key = "Measure", value = "Average", average_B5A, average_B6C, average_C2D, average_C2E, average_G41A)
df_long_0509 <- df_long_0509_mess %>% filter(!is.na(Average), !is.na(SD)) %>% filter(Average >= 0 & Average <= 7)
year_colors_0509 <- c("2005" = "#990000", "2009" = "#0A2240")

knitr::kable(df_long_0509, align = "ccccc")

year	SD	Variable	Measure	Average
2005	1.1753865	B5A	average_B5A	3.487085
2009	0.9776143	B5A	average_B5A	4.315498
2005	1.4445587	B6C	average_B6C	5.278598
2009	1.2722460	B6C	average_B6C	5.365314
2005	1.8437556	C2D	average_C2D	3.651292
2009	1.8698942	C2D	average_C2D	3.933579
2005	1.8982344	C2E	average_C2E	3.485240
2009	1.9034159	C2E	average_C2E	3.725092
2005	1.6533276	G41A	average_G41A	5.239852
2009	1.8311009	G41A	average_G41A	4.632841

ggplot(df_long_0509, aes(x = Variable, y = Average, fill = as.factor(year), group = year)) + geom_bar(stat = "identity", position = position_dodge(width = 0.8), color = "black", width = 0.5) + geom_errorbar(aes(ymin = Average - SD, ymax = Average + SD), position = position_dodge(width = 0.8), width = 0.25, colour = "#000000", size = 1) + scale_fill_manual(values = year_colors_0509) + labs(title = "2005 vs 2009 (within person)", x = "Variables", y = "Mean Value", fill = "Year") + theme_minimal() + theme(axis.text.x = element_text(angle = 0, hjust = 0.5, vjust = 0.5), axis.ticks.length = unit(0.3, "cm"))+ theme(panel.grid.major = element_blank(),panel.grid.minor = element_blank(),plot.title = element_text(size = 16, face = "bold"),axis.title.x = element_text(vjust = -2)) + scale_x_discrete(labels = c("B5A" = "B5A: \n Responsibility \n for Self", "B6C" = "B6C:\n Money Management \n Skills", "C2D" = "C2D:\n Worry about \n Expenses", "C2E" = "C2E:\n Worry about \n Future Job", "G41A" = "G41A:\n Importance of \n Job Status")) + scale_y_continuous(breaks = seq(1, 7, by = 1), expand = c(0, 0)) + coord_cartesian(ylim = c(1, 9))  +
  
  geom_segment(aes(x = 1 - 0.24, xend = 1 + 0.24, y = 5.8, yend = 5.8), color = "black") + 
  annotate("text", x = 1, y = 6.15, label = "p = 0.000", size = 3.2) + 
  geom_segment(aes(x = 1 - 0.24, xend = 1 - 0.24, y = 5.8, yend = 5.5), color = "black") + 
  geom_segment(aes(x = 1 + 0.24, xend = 1 + 0.24, y = 5.8, yend = 5.5), color = "black") +

  geom_segment(aes(x = 3 - 0.24, xend = 3 + 0.24, y = 6.4, yend = 6.4), color = "black") + 
  annotate("text", x = 3, y = 6.75, label = "p = 0.003", size = 3.2) + 
  geom_segment(aes(x = 3 - 0.24, xend = 3 - 0.24, y = 6.4, yend = 6.1), color = "black") + 
  geom_segment(aes(x = 3 + 0.24, xend = 3 + 0.24, y = 6.4, yend = 6.1), color = "black") +

  geom_segment(aes(x = 4 - 0.24, xend = 4 + 0.24, y = 6.2, yend = 6.2), color = "black") + 
  annotate("text", x = 4, y = 6.55, label = "p = 0.010", size = 3.2) + 
  geom_segment(aes(x = 4 - 0.24, xend = 4 - 0.24, y = 6.2, yend = 5.9), color = "black") + 
  geom_segment(aes(x = 4 + 0.24, xend = 4 + 0.24, y = 6.2, yend = 5.9), color = "black") +

  geom_segment(aes(x = 5 - 0.24, xend = 5 + 0.24, y = 7.5, yend = 7.5), color = "black") + 
  annotate("text", x = 5, y = 7.85, label = "p = 0.000", size = 3.2) + 
  geom_segment(aes(x = 5 - 0.24, xend = 5 - 0.24, y = 7.5, yend = 7.2), color = "black") + 
  geom_segment(aes(x = 5 + 0.24, xend = 5 + 0.24, y = 7.5, yend = 7.2), color = "black")

2009 vs 2015 (within person)

df_long_0915_mess <- bind_rows(mutate(mean_B5A_0915, SD = sd_B5A_0915$sd_B5A, Variable = "B5A"),mutate(mean_B6C_0915, SD = sd_B6C_0915$sd_B6C, Variable = "B6C"), mutate(mean_C2D_0915, SD = sd_C2D_0915$sd_C2D, Variable = "C2D"),mutate(mean_C2E_0915, SD = sd_C2E_0915$sd_C2E, Variable = "C2E"), mutate(mean_G41A_0915, SD = sd_G41A_0915$sd_G41A, Variable = "G41A")) %>%  gather(key = "Measure", value = "Average", average_B5A, average_B6C, average_C2D, average_C2E, average_G41A)
df_long_0915 <- df_long_0915_mess %>% filter(!is.na(Average), !is.na(SD)) %>% filter(Average >= 0 & Average <= 7)
year_colors_0915 <- c("2009" = "#990000", "2015" = "#0A2240")

knitr::kable(df_long_0915, align = "ccccc")

year	SD	Variable	Measure	Average
2009	1.2117796	B5A	average_B5A	3.443359
2015	0.9194487	B5A	average_B5A	4.503906
2009	1.2650962	B6C	average_B6C	5.485437
2015	1.2547127	B6C	average_B6C	5.396116
2009	1.9307493	C2D	average_C2D	3.735922
2015	1.8594813	C2D	average_C2D	3.417476
2009	1.9444478	C2E	average_C2E	3.634952
2015	1.7794978	C2E	average_C2E	3.166990
2009	1.7032629	G41A	average_G41A	5.130350
2015	1.9306797	G41A	average_G41A	4.476654

ggplot(df_long_0915, aes(x = Variable, y = Average, fill = as.factor(year), group = year)) + geom_bar(stat = "identity", position = position_dodge(width = 0.8), color = "black", width = 0.5) + geom_errorbar(aes(ymin = Average - SD, ymax = Average + SD), position = position_dodge(width = 0.8), width = 0.25, colour = "#000000", size = 1) + scale_fill_manual(values = year_colors_0915) + labs(title = "2009 vs 2015 (within person)", x = "Variables", y = "Mean Value", fill = "Year") + theme_minimal() + theme(axis.text.x = element_text(angle = 0, hjust = 0.5, vjust = 0.5), axis.ticks.length = unit(0.3, "cm")) + theme(panel.grid.major = element_blank(),panel.grid.minor = element_blank(),plot.title = element_text(size = 16, face = "bold"),axis.title.x = element_text(vjust = -2)) + scale_x_discrete(labels = c("B5A" = "B5A: \n Responsibility \n for Self", "B6C" = "B6C:\n Money Management \n Skills", "C2D" = "C2D:\n Worry about \n Expenses", "C2E" = "C2E:\n Worry about \n Future Job", "G41A" = "G41A:\n Importance of \n Job Status")) + scale_y_continuous(breaks = seq(1, 7, by = 1), expand = c(0, 0)) + coord_cartesian(ylim = c(1, 9))  +

geom_segment(aes(x = 1 - 0.24, xend = 1 + 0.24, y = 6, yend = 6), color = "black") + 
  annotate("text", x = 1, y = 6.33, label = "p = 0.000", size = 3.2) + 
  geom_segment(aes(x = 1 - 0.24, xend = 1 - 0.24, y = 6, yend = 5.7), color = "black") + 
  geom_segment(aes(x = 1 + 0.24, xend = 1 + 0.24, y = 6, yend = 5.7), color = "black") +
  
  geom_segment(aes(x = 3 - 0.24, xend = 3 + 0.24, y = 6.4, yend = 6.4), color = "black") + 
  annotate("text", x = 3, y = 6.75, label = "p = 0.001", size = 3.2) + 
  geom_segment(aes(x = 3 - 0.24, xend = 3 - 0.24, y = 6.4, yend = 6.1), color = "black") + 
  geom_segment(aes(x = 3 + 0.24, xend = 3 + 0.24, y = 6.4, yend = 6.1), color = "black") +
  
  geom_segment(aes(x = 4 - 0.24, xend = 4 + 0.24, y = 6.2, yend = 6.2), color = "black") + 
  annotate("text", x = 4, y = 6.55, label = "p = 0.000", size = 3.2) + 
  geom_segment(aes(x = 4 - 0.24, xend = 4 - 0.24, y = 6.2, yend = 5.9), color = "black") + 
  geom_segment(aes(x = 4 + 0.24, xend = 4 + 0.24, y = 6.2, yend = 5.9), color = "black") +
  
  geom_segment(aes(x = 5 - 0.24, xend = 5 + 0.24, y = 7.5, yend = 7.5), color = "black") + 
  annotate("text", x = 5, y = 7.85, label = "p = 0.000", size = 3.2) + 
  geom_segment(aes(x = 5 - 0.24, xend = 5 - 0.24, y = 7.5, yend = 7.2), color = "black") + 
  geom_segment(aes(x = 5 + 0.24, xend = 5 + 0.24, y = 7.5, yend = 7.2), color = "black")