Data analysis Income

Changes in the Justification of Pension Inequality in Chile (2016–2023) and its Relationship to Social Class and Beliefs in Meritocracy

Author

Andreas Laffert

Published

April 4, 2026

1 Presentation

This is the analysis code for the paper “Changes in the Justification of Pension Inequality in Chile (2016–2023) and its Relationship to Social Class and Beliefs in Meritocracy”. The dataset used is df_balanced_income.RData.

2 Libraries

Show the code
if (! require("pacman")) install.packages("pacman")

pacman::p_load(tidyverse, 
               sjmisc, 
               sjPlot, 
               lme4, 
               here, 
               performance,
               influence.ME, 
               srvyr,
               ordinal,
               texreg, 
               ggdist,
               misty,
               kableExtra,
               ggalluvial, 
               shadowtext,
               MetBrewer,
               patchwork,
               sjlabelled)


options(scipen=999)
rm(list = ls())

3 Data

Show the code
load(file = here("input/data/rob_income/df_balanced_rob_income.RData"))

glimpse(df_balanced)
Rows: 7,968
Columns: 17
$ idencuesta            <dbl> 1101011, 1101011, 1101011, 1101011, 1101011, 110…
$ ola                   <chr> "2016", "2017", "2018", "2019", "2022", "2023", …
$ muestra               <dbl> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, …
$ ponderador_long_total <dbl> 0.11821742, 0.16716225, 0.15261954, 0.33262422, …
$ segmento              <dbl> 110101, 110101, 110101, 110101, 110101, 110101, …
$ estrato               <dbl> 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, …
$ just_pension          <fct> Strongly disagree, Strongly disagree, Strongly d…
$ merit_effort          <fct> Agree, Disagree, Agree, Disagree, Disagree, Disa…
$ merit_talent          <fct> Agree, Disagree, Neither agree nor disagree, Dis…
$ educ                  <fct> Less than Universitary, Less than Universitary, …
$ educyear              <dbl> 4.30, 4.30, 7.50, 4.30, 4.30, 4.30, 9.80, 9.80, …
$ sex                   <fct> Female, Female, Female, Female, Female, Female, …
$ age                   <dbl> 64, 65, 66, 68, 70, 70, 60, 62, 62, 64, 67, 66, …
$ aget                  <fct> 50-64, 65 or more, 65 or more, 65 or more, 65 or…
$ ideo                  <fct> Does not identify, Does not identify, Right, Lef…
$ decile_eq             <fct> 3, 1, 1, 1, NA, 4, 7, 8, 8, 2, NA, 1, 9, NA, NA,…
$ decile_eq1            <fct> 3, 1, 1, 1, DNA, 4, 7, 8, 8, 2, DNA, 1, 9, DNA, …
Show the code
# Generate analytical sample

df_study1 <- df_balanced %>%
  select(-muestra) %>% 
  na.omit() %>% 
  mutate(ola = case_when(ola == "2016" ~ 1,
                         ola == "2017" ~ 2, 
                         ola == "2018" ~ 3,
                         ola == "2019" ~ 4,
                         ola == "2022" ~ 5,
                         ola == "2023" ~ 6)) %>% 
  mutate(ola = as.factor(ola),
         ola_num = as.numeric(ola),
         ola_2=as.numeric(ola)^2)

df_study1 <- df_study1 %>%
  group_by(idencuesta) %>%             # Agrupar por el identificador del participante
  mutate(n_participaciones = n()) %>%  # Contar el número de filas (participaciones) por participante
  ungroup()

df_study1 <- df_study1 %>% filter(n_participaciones>1)

# Corregir etiquetas

df_study1$just_pension <- sjlabelled::set_label(df_study1$just_pension, 
                        label = "Pension distributive justice")

df_study1$egp <- sjlabelled::set_label(df_study1$egp, 
                        label = "Social class")

df_study1$merit_effort <- sjlabelled::set_label(df_study1$merit_effort, 
                        label = "People are rewarded for their efforts")

df_study1$merit_talent <- sjlabelled::set_label(df_study1$merit_talent, 
                        label = "People are rewarded for their intelligence")

4 Analysis

4.1 Descriptives

Show the code
datos.pension <- df_study1 %>% 
   mutate(just_pension = factor(just_pension, 
                              levels = c("Strongly agree",
                                         "Agree",
                                         "Neither agree nor disagree",
                                         "Disagree",
                                         "Strongly disagree"))) %>% 
  group_by(idencuesta, ola) %>% 
  count(just_pension) %>% 
  group_by(ola) %>% 
  mutate(porcentaje=n/sum(n)) %>% 
  ungroup() %>% 
  na.omit() %>% 
  mutate(wave = case_when(ola == 1 ~ "2016",
                          ola == 2 ~ "2017",
                          ola == 3 ~ "2018",
                          ola == 4 ~ "2019",
                          ola == 5 ~ "2022",
                          ola == 6 ~ "2023"),
         wave = factor(wave, levels = c("2016",
                                        "2017",
                                        "2018",
                                        "2019",
                                        "2022",
                                        "2023")))



etiquetas.pension <- df_study1 %>%
  mutate(just_pension = factor(just_pension, 
                              levels = c("Strongly agree",
                                         "Agree",
                                         "Neither agree nor disagree",
                                         "Disagree",
                                         "Strongly disagree"))) %>% 
  group_by(ola, just_pension) %>%
  summarise(count = n(), .groups = "drop") %>%
  group_by(ola) %>%
  mutate(porcentaje = count / sum(count)) %>% 
  na.omit() %>% 
  mutate(idencuesta = 1,
         wave = case_when(ola == 1 ~ "2016",
                          ola == 2 ~ "2017",
                          ola == 3 ~ "2018",
                          ola == 4 ~ "2019",
                          ola == 5 ~ "2022",
                          ola == 6 ~ "2023"),
         wave = factor(wave, levels = c("2016",
                                        "2017",
                                        "2018",
                                        "2019",
                                        "2022",
                                        "2023")))


datos.pension %>% 
  ggplot(aes(x = wave, fill = just_pension, stratum = just_pension,
             alluvium = idencuesta, y = porcentaje)) +
  ggalluvial::geom_flow(alpha = .4) + 
  ggalluvial::geom_stratum(linetype = 0) +
  scale_y_continuous(labels = scales::percent) + 
  scale_fill_manual(values =  c("#0571B0","#92C5DE","#b3b3b3ff","#F4A582","#CA0020")) +
  geom_shadowtext(data = etiquetas.pension,
                  aes(label = ifelse(porcentaje > 0 , scales::percent(porcentaje, accuracy = .1),"")),
                  position = position_stack(vjust = .5),
                  show.legend = FALSE,
                  size = 3,
                  color = rep('white'),
                  bg.colour='grey30')+
  labs(y = "%",
       x = NULL,
       fill = NULL,
       title = NULL) +
  theme_ggdist() +
  theme(legend.position = "bottom")
Figure 1: Changes in the justification of inequality in pensions over time (2016-2023)

4.2 Longitudinal multilevel models

4.3 ICC

Show the code
m0 <- clmm(just_pension ~ 1 + (1 | idencuesta), 
           link = "logit",
  Hess = TRUE, # calcula explícitamente la matriz varianza-covarianza de estimadores
                data = df_study1)

performance::icc(m0, by_group = T) # 0.23 es between, 0.77 within
# ICC by Group

Group      |   ICC
------------------
idencuesta | 0.226

4.4 Time effects

Show the code
#m1.1 <- clmm(just_pension ~ 1 + ola + (1 | idencuesta),
#                link = "logit",
#  Hess = TRUE,
#  data = df_study1)
#
#m1.2 <- clmm(just_pension ~ 1 + ola_num + (1 | idencuesta),
#                link = "logit",
#  Hess = TRUE, data = df_study1)
#
#m1.3 <- clmm(just_pension ~ 1 + ola_num + ola_2 + (1| idencuesta),
#                link = "logit",
#  Hess = TRUE, data = df_study1)
#
#m1.4 <- clmm(just_pension ~ 1 + ola_num + ola_2 + (1 + ola_num | #idencuesta),
#               link = "logit",
# Hess = TRUE, data = df_study1)
#
##save(m1.1,m1.2,m1.3,m1.4, file = here("output/rob_income/time_effects.RData"))

load(file = here("output/rob_income/time_effects.RData"))

ccoef <- list(
  "Strongly disagree|Disagree" = "Strongly disagree|Disagree",
  "Disagree|Neither agree nor disagree" = "Disagree|Neither agree nor disagree", 
  "Neither agree nor disagree|Agree" = "Neither agree nor disagree|Agree", 
  "Agree|Strongly agree" = "Agree|Strongly agree", 
  "ola2017" = "Wave 2017",
  "ola2018" = "Wave 2018",
  "ola2019" = "Wave 2019",
  "ola2022" = "Wave 2022",
  "ola2023" = "Wave 2023",
  ola_num = "Wave",
  ola_2 = "Wave^2")

texreg::htmlreg(list(m1.1,m1.2,m1.3,m1.4),
               caption.above = T,
               caption = NULL,
               stars = c(0.05, 0.01, 0.001),
               custom.coef.map = ccoef,
               digits = 3,
               groups = list("Wave (Ref.= 2016)" = 5:9),
               custom.note = "Note: Cells contain regression coefficients with standard errors in parentheses. %stars.",
               leading.zero = T,
               use.packages = F,
               booktabs = F,
               scalebox = 0.80,
               include.loglik = FALSE,
               include.aic = FALSE,
               center = T)
  Model 1 Model 2 Model 3 Model 4
Strongly disagree|Disagree -1.030*** -0.498*** -1.510*** -1.513***
  (0.065) (0.066) (0.123) (0.123)
Disagree|Neither agree nor disagree 1.327*** 1.820*** 0.842*** 0.838***
  (0.066) (0.071) (0.122) (0.121)
Neither agree nor disagree|Agree 1.912*** 2.396*** 1.425*** 1.422***
  (0.070) (0.075) (0.123) (0.123)
Agree|Strongly agree 4.419*** 4.879*** 3.927*** 3.931***
  (0.109) (0.113) (0.148) (0.148)
Wave (Ref.= 2016)        
         
     Wave 2017 -0.368***      
  (0.081)      
     Wave 2018 -0.024      
  (0.078)      
     Wave 2019 0.086      
  (0.076)      
     Wave 2023 0.842***      
  (0.077)      
     Wave   0.209*** -0.652*** -0.653***
    (0.017) (0.090) (0.090)
Wave^2     0.144*** 0.144***
      (0.015) (0.015)
BIC 15519.842 15604.978 15517.638 15534.327
Num. obs. 6123 6123 6123 6123
Groups (idencuesta) 1327 1327 1327 1327
Variance: idencuesta: (Intercept) 1.064 1.019 1.056 0.939
Variance: idencuesta: ola_num       0.000
Note: Cells contain regression coefficients with standard errors in parentheses. ***p < 0.001; **p < 0.01; *p < 0.05.

4.4.1 Anova

Show the code
anova(m1.3, m1.4) # quedarse con tiempo continua y con pendiente aleatoria
Likelihood ratio tests of cumulative link models:
 
     formula:                                                        link:
m1.3 just_pension ~ 1 + ola_num + ola_2 + (1 | idencuesta)           logit
m1.4 just_pension ~ 1 + ola_num + ola_2 + (1 + ola_num | idencuesta) logit
     threshold:
m1.3 flexible  
m1.4 flexible  

     no.par   AIC  logLik LR.stat df Pr(>Chisq)
m1.3      7 15471 -7728.3                      
m1.4      9 15474 -7727.9  0.7511  2     0.6869

4.5 WE and BE main effects

Table 1: Cumulative link longitudinal multilevel models for pension market justice preferences
  Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7
Wave 2017 -0.368***            
  (0.081)            
Wave 2018 -0.024            
  (0.078)            
Wave 2019 0.086            
  (0.076)            
Wave 2023 0.842***            
  (0.077)            
Wave   -0.653*** -0.675*** -0.678*** -0.680*** -0.681*** -0.685***
    (0.090) (0.090) (0.090) (0.090) (0.090) (0.090)
Wave^2   0.144*** 0.148*** 0.148*** 0.148*** 0.148*** 0.148***
    (0.015) (0.015) (0.015) (0.015) (0.015) (0.015)
Merit: Effort (WE)     0.145*** 0.147*** 0.148*** 0.148*** 0.142***
      (0.039) (0.039) (0.039) (0.039) (0.039)
Merit: Talent (WE)     0.071 0.071 0.070 0.070 0.066
      (0.039) (0.039) (0.039) (0.039) (0.039)
Merit: Effort (BE)       0.333*** 0.358*** 0.351*** 0.335***
        (0.090) (0.089) (0.088) (0.086)
Merit: Talent (BE)       0.089 0.088 0.112 0.066
        (0.092) (0.090) (0.089) (0.087)
Decile 10 (BE)         0.707*** 0.537*** 0.478***
          (0.096) (0.104) (0.102)
Universitary education (BE)           0.385*** 0.344***
            (0.096) (0.097)
Controls No No No No No No Yes
BIC 15519.842 15534.327 15510.926 15479.716 15434.695 15427.404 15399.703
Numb. obs. 6123 6123 6123 6123 6123 6123 6123
Num. groups: individuals 1327 1327 1327 1327 1327 1327 1327
Var: individuals (Intercept) 1.064 0.939 0.962 0.832 0.726 0.704 0.656
Var: individuals, wave   0.000 0.000 0.001 0.001 0.001 0.001
Note: Cells contain regression coefficients with standard errors in parentheses. ***p < 0.001; **p < 0.01; *p < 0.05.

4.6 Interactions with controls (direct effect)

Show the code
 ## WE and BE Interactions with controls

#
## income
#m11 <- clmm(just_pension ~ 1 + ola_num +  ola_2 + 
#               merit_effort_cwc + merit_talent_cwc + 
#               merit_effort_mean + merit_talent_mean + 
#               dummy_decile10_mean*merit_effort_cwc + 
#               dummy_educ_mean + ideo + sex + age + 
#               (1 + ola_num + merit_effort_cwc| idencuesta), 
#            link = "logit",
#  Hess = TRUE,
#           data = df_study1)
#
#
#m12 <- clmm(just_pension ~ 1 + ola_num +  ola_2 + 
#               merit_effort_cwc + merit_talent_cwc + 
#               merit_effort_mean + merit_talent_mean + 
#               dummy_decile10_mean*merit_effort_mean + 
#               dummy_educ_mean + ideo + sex + age + 
#               (1 + ola_num + merit_effort_mean| idencuesta), 
#            link = "logit",
#  Hess = TRUE,
#           data = df_study1)
#
#m13 <- clmm(just_pension ~ 1 + ola_num +  ola_2 + 
#               merit_effort_cwc + merit_talent_cwc + 
#               merit_effort_mean + merit_talent_mean + 
#               dummy_decile10_mean*merit_talent_cwc + 
#               dummy_educ_mean + ideo + sex + age + 
#               (1 + ola_num + merit_talent_cwc| idencuesta), 
#            link = "logit",
#  Hess = TRUE,
#           data = df_study1)
#
#
#m14 <- clmm(just_pension ~ 1 + ola_num +  ola_2 + 
#               merit_effort_cwc + merit_talent_cwc + 
#               merit_effort_mean + merit_talent_mean + 
#               dummy_decile10_mean*merit_talent_mean + 
#               dummy_educ_mean + ideo + sex + age + 
#               (1 + ola_num + merit_talent_mean| idencuesta), 
#            link = "logit",
#  Hess = TRUE,
#           data = df_study1)
#
## educ
#
#
#m15 <- clmm(just_pension ~ 1 + ola_num +  ola_2 + 
#              merit_effort_cwc + merit_talent_cwc + 
#              merit_effort_mean + merit_talent_mean + 
#              dummy_educ_mean*merit_effort_cwc + 
#              dummy_decile10_mean + ideo + sex + age + 
#              (1 + ola_num + merit_effort_cwc| idencuesta), 
#            link = "logit",
#            Hess = TRUE,
#            data = df_study1)
#
#
#m16 <- clmm(just_pension ~ 1 + ola_num +  ola_2 + 
#              merit_effort_cwc + merit_talent_cwc + 
#              merit_effort_mean + merit_talent_mean + 
#              dummy_educ_mean*merit_effort_mean + 
#              dummy_decile10_mean + ideo + sex + age + 
#              (1 + ola_num + merit_effort_mean| idencuesta), 
#            link = "logit",
#            Hess = TRUE,
#            data = df_study1)
#
#m17 <- clmm(just_pension ~ 1 + ola_num +  ola_2 + 
#              merit_effort_cwc + merit_talent_cwc + 
#              merit_effort_mean + merit_talent_mean + 
#              dummy_educ_mean*merit_talent_cwc + 
#              dummy_decile10_mean + ideo + sex + age + 
#              (1 + ola_num + merit_talent_cwc| idencuesta), 
#            link = "logit",
#            Hess = TRUE,
#            data = df_study1)
#
#
#m18 <- clmm(just_pension ~ 1 + ola_num +  ola_2 + 
#              merit_effort_cwc + merit_talent_cwc + 
#              merit_effort_mean + merit_talent_mean + 
#              dummy_educ_mean*merit_talent_mean + 
#              dummy_decile10_mean + ideo + sex + age + 
#              (1 + ola_num + merit_talent_mean| idencuesta), 
#            link = "logit",
#            Hess = TRUE,
#            data = df_study1)
#
#save(m11,m12,m13,m14,m15,m16,m17,m18,file = #here("output/rob_income/interactions_direct.RData"))

load(file = here("output/rob_income/interactions_direct.RData"))

htmlreg(list(m11,m12,m13,m14))
Statistical models
  Model 1 Model 2 Model 3 Model 4
ola_num -0.69*** -0.69 -0.69*** -0.69***
  (0.09)   (0.09) (0.09)
ola_2 0.15*** 0.15 0.15*** 0.15***
  (0.02)   (0.02) (0.01)
merit_effort_cwc 0.15*** 0.14 0.15*** 0.14***
  (0.04)   (0.04) (0.04)
merit_talent_cwc 0.06 0.07 0.08 0.07
  (0.04)   (0.04) (0.04)
merit_effort_mean 0.33*** 0.27 0.34*** 0.34***
  (0.09)   (0.09) (0.08)
merit_talent_mean 0.06 0.08 0.07 0.02
  (0.09)   (0.09) (0.09)
dummy_decile10_mean 0.49*** -0.40 0.49*** -0.22
  (0.10)   (0.10) (0.44)
dummy_educ_mean 0.35*** 0.35 0.35*** 0.37***
  (0.10)   (0.10) (0.10)
ideoCenter 0.32*** 0.31 0.32*** 0.32***
  (0.08)   (0.08) (0.08)
ideoRight 0.66*** 0.65 0.67*** 0.65***
  (0.10)   (0.10) (0.10)
ideoDoes not identify 0.23** 0.22 0.24** 0.23**
  (0.08)   (0.08) (0.08)
sexFemale -0.38*** -0.36 -0.38*** -0.36***
  (0.08)   (0.08) (0.08)
age -0.00 -0.00 -0.00 -0.00
  (0.00)   (0.00) (0.00)
merit_effort_cwc:dummy_decile10_mean -0.03      
  (0.09)      
Strongly disagree|Disagree -0.40 -0.53 -0.37 -0.50
  (0.25)   (0.25) (0.26)
Disagree|Neither agree nor disagree 2.02*** 1.83 2.06*** 1.86***
  (0.25)   (0.25) (0.26)
Neither agree nor disagree|Agree 2.62*** 2.42 2.66*** 2.45***
  (0.25)   (0.25) (0.26)
Agree|Strongly agree 5.18*** 4.95 5.23*** 4.98***
  (0.27)   (0.27) (0.27)
merit_effort_mean:dummy_decile10_mean   0.34    
         
merit_talent_cwc:dummy_decile10_mean     -0.07  
      (0.09)  
merit_talent_mean:dummy_decile10_mean       0.25
        (0.15)
Log Likelihood -7605.70 -7606.31 -7603.75 -7607.18
AIC 15259.40 15260.62 15255.51 15262.35
BIC 15420.67 15421.90 15416.78 15423.63
Num. obs. 6123 6123 6123 6123
Groups (idencuesta) 1327 1327 1327 1327
Variance: idencuesta: (Intercept) 0.76 2.56 0.79 3.10
Variance: idencuesta: ola_num 0.01 0.00 0.01 0.01
Variance: idencuesta: merit_effort_cwc 0.14      
Variance: idencuesta: merit_effort_mean   0.30    
Variance: idencuesta: merit_talent_cwc     0.16  
Variance: idencuesta: merit_talent_mean       0.34
***p < 0.001; **p < 0.01; *p < 0.05
Show the code
htmlreg(list(m15,m16,m17,m18))
Statistical models
  Model 1 Model 2 Model 3 Model 4
ola_num -0.69*** -0.69*** -0.69*** -0.69***
  (0.09) (0.09) (0.09) (0.09)
ola_2 0.15*** 0.15*** 0.15*** 0.15***
  (0.02) (0.01) (0.02) (0.01)
merit_effort_cwc 0.14** 0.14*** 0.15*** 0.14***
  (0.05) (0.04) (0.04) (0.04)
merit_talent_cwc 0.06 0.07 0.06 0.07
  (0.04) (0.04) (0.05) (0.04)
merit_effort_mean 0.33*** 0.28** 0.34*** 0.31***
  (0.09) (0.09) (0.09) (0.08)
merit_talent_mean 0.06 0.07 0.07 0.04
  (0.09) (0.09) (0.09) (0.09)
dummy_educ_mean 0.35*** -0.21 0.35*** -0.28
  (0.10) (0.38) (0.10) (0.41)
dummy_decile10_mean 0.49*** 0.47*** 0.49*** 0.47***
  (0.10) (0.10) (0.10) (0.10)
ideoCenter 0.32*** 0.32*** 0.32*** 0.31***
  (0.08) (0.08) (0.08) (0.08)
ideoRight 0.66*** 0.65*** 0.67*** 0.65***
  (0.10) (0.10) (0.10) (0.10)
ideoDoes not identify 0.23** 0.22** 0.24** 0.22**
  (0.08) (0.08) (0.08) (0.08)
sexFemale -0.38*** -0.36*** -0.38*** -0.36***
  (0.08) (0.08) (0.08) (0.08)
age -0.00 -0.00 -0.00 -0.00
  (0.00) (0.00) (0.00) (0.00)
merit_effort_cwc:dummy_educ_mean 0.05      
  (0.08)      
Strongly disagree|Disagree -0.40 -0.49 -0.37 -0.51
  (0.25) (0.26) (0.25) (0.26)
Disagree|Neither agree nor disagree 2.02*** 1.88*** 2.06*** 1.86***
  (0.25) (0.26) (0.25) (0.26)
Neither agree nor disagree|Agree 2.62*** 2.46*** 2.66*** 2.44***
  (0.25) (0.26) (0.25) (0.26)
Agree|Strongly agree 5.18*** 4.99*** 5.23*** 4.97***
  (0.27) (0.27) (0.27) (0.28)
merit_effort_mean:dummy_educ_mean   0.22    
    (0.14)    
merit_talent_cwc:dummy_educ_mean     0.04  
      (0.08)  
merit_talent_mean:dummy_educ_mean       0.24
        (0.14)
Log Likelihood -7605.57 -7607.56 -7603.92 -7607.22
AIC 15259.15 15263.12 15255.84 15262.44
BIC 15420.42 15424.40 15417.12 15423.71
Num. obs. 6123 6123 6123 6123
Groups (idencuesta) 1327 1327 1327 1327
Variance: idencuesta: (Intercept) 0.76 2.56 0.79 3.21
Variance: idencuesta: ola_num 0.01 0.00 0.01 0.01
Variance: idencuesta: merit_effort_cwc 0.14      
Variance: idencuesta: merit_effort_mean   0.31    
Variance: idencuesta: merit_talent_cwc     0.16  
Variance: idencuesta: merit_talent_mean       0.36
***p < 0.001; **p < 0.01; *p < 0.05

4.7 Interactions WE x Time

Show the code
# ## WE and BE Interactions with controls
#
#df_study1 <- df_study1 |>
#  mutate(
#    time  = ola_num - min(ola_num),
#    time2 = time^2
#  )
#
#m19 <- clmm(just_pension ~ 1 + time*merit_effort_cwc +  
#              time2 + merit_talent_cwc + 
#               merit_effort_mean + merit_talent_mean + 
#               dummy_decile10_mean + 
#               dummy_educ_mean + ideo + sex + age + 
#               (1 + time + merit_effort_cwc| idencuesta), 
#            link = "logit",
#  Hess = TRUE,
#           data = df_study1)
#
#m20 <- clmm(just_pension ~ 1 + time*merit_talent_cwc +  
#              time2 +  merit_effort_cwc + 
#               merit_effort_mean + merit_talent_mean + 
#               dummy_decile10_mean + 
#               dummy_educ_mean + ideo + sex + age + 
#               (1 + time + merit_talent_cwc| idencuesta), 
#            link = "logit",
#  Hess = TRUE,
#           data = df_study1)
#
#
#screenreg(list(m19, m20))
#
#m19_flex <- clmm(
#  just_pension ~ 1 + factor(ola_num)*merit_effort_cwc +
#    merit_talent_cwc +
#    merit_effort_mean + merit_talent_mean +
#    dummy_decile10_mean + dummy_educ_mean +
#    ideo + sex + age +
#    (1 + merit_effort_cwc | idencuesta),
#  link = "logit",
#  Hess = TRUE,
#  data = df_study1
#)
#
#m20_flex <- clmm(
#  just_pension ~ 1 + factor(ola_num)*merit_talent_cwc + 
#    merit_effort_cwc +
#    merit_effort_mean + merit_talent_mean +
#    dummy_decile10_mean + dummy_educ_mean +
#    ideo + sex + age +
#    (1 + merit_talent_cwc | idencuesta),
#  link = "logit",
#  Hess = TRUE,
#  data = df_study1
#)
#
#screenreg(list(m19_flex, m20_flex))
#
#save(m19,m20, m19_flex, m20_flex, file = #here("output/rob_income/interactions_withinxtime.RData"))

load(file = here("output/rob_income/interactions_withinxtime.RData"))

htmlreg(list(m19, m20, m19_flex, m20_flex))
Statistical models
  Model 1 Model 2 Model 3 Model 4
time -0.39*** -0.39***    
  (0.06) (0.06)    
merit_effort_cwc 0.15* 0.15*** 0.28*** 0.15***
  (0.06) (0.04) (0.08) (0.04)
time2 0.15*** 0.15***    
  (0.02) (0.02)    
merit_talent_cwc 0.06 0.07 0.06 0.16*
  (0.04) (0.06) (0.04) (0.08)
merit_effort_mean 0.33*** 0.34*** 0.33*** 0.34***
  (0.09) (0.09) (0.09) (0.09)
merit_talent_mean 0.06 0.07 0.06 0.07
  (0.09) (0.09) (0.09) (0.09)
dummy_decile10_mean 0.49*** 0.49*** 0.48*** 0.49***
  (0.10) (0.10) (0.10) (0.10)
dummy_educ_mean 0.35*** 0.35*** 0.35*** 0.35***
  (0.10) (0.10) (0.10) (0.10)
ideoCenter 0.32*** 0.32*** 0.31*** 0.32***
  (0.08) (0.08) (0.08) (0.08)
ideoRight 0.66*** 0.67*** 0.66*** 0.67***
  (0.10) (0.10) (0.10) (0.10)
ideoDoes not identify 0.23** 0.24** 0.24** 0.25**
  (0.08) (0.08) (0.08) (0.08)
sexFemale -0.38*** -0.38*** -0.38*** -0.38***
  (0.08) (0.08) (0.08) (0.08)
age -0.00 -0.00 -0.00 -0.00
  (0.00) (0.00) (0.00) (0.00)
time:merit_effort_cwc -0.00      
  (0.02)      
Strongly disagree|Disagree 0.13 0.16 0.09 0.11
  (0.22) (0.22) (0.22) (0.22)
Disagree|Neither agree nor disagree 2.55*** 2.60*** 2.51*** 2.54***
  (0.23) (0.23) (0.23) (0.23)
Neither agree nor disagree|Agree 3.15*** 3.20*** 3.11*** 3.14***
  (0.23) (0.23) (0.23) (0.23)
Agree|Strongly agree 5.71*** 5.77*** 5.66*** 5.70***
  (0.25) (0.25) (0.25) (0.25)
time:merit_talent_cwc   -0.00    
    (0.02)    
factor(ola_num)2     -0.38*** -0.39***
      (0.08) (0.08)
factor(ola_num)3     -0.05 -0.05
      (0.08) (0.08)
factor(ola_num)4     0.09 0.09
      (0.08) (0.08)
factor(ola_num)5     0.84*** 0.85***
      (0.08) (0.08)
factor(ola_num)2:merit_effort_cwc     -0.24*  
      (0.11)  
factor(ola_num)3:merit_effort_cwc     -0.26*  
      (0.11)  
factor(ola_num)4:merit_effort_cwc     -0.11  
      (0.11)  
factor(ola_num)5:merit_effort_cwc     -0.08  
      (0.11)  
factor(ola_num)2:merit_talent_cwc       -0.21
        (0.11)
factor(ola_num)3:merit_talent_cwc       -0.17
        (0.11)
factor(ola_num)4:merit_talent_cwc       -0.07
        (0.10)
factor(ola_num)5:merit_talent_cwc       -0.08
        (0.10)
Log Likelihood -7605.74 -7604.04 -7597.43 -7597.95
AIC 15259.47 15256.09 15246.86 15247.90
BIC 15420.75 15417.36 15421.57 15422.61
Num. obs. 6123 6123 6123 6123
Groups (idencuesta) 1327 1327 1327 1327
Variance: idencuesta: (Intercept) 0.79 0.80 0.87 0.88
Variance: idencuesta: time 0.01 0.01    
Variance: idencuesta: merit_effort_cwc 0.14   0.13  
Variance: idencuesta: merit_talent_cwc   0.16   0.16
***p < 0.001; **p < 0.01; *p < 0.05