1.資料管理

#載入資料 
data(klein2000)
dta <- klein2000

2.計算結果

#資料整理
dta_pay <- dta %>%
        group_by(GRPID) %>%
        mutate(mneglead=mean(PAY))
#計算工作滿意度的多少差異可以歸因於屬於同一群體的個人之間的薪酬差異
m1 <- lme4::lmer(JOBSAT ~ (PAY | GRPID), data=dta)
summary(m1)
## Linear mixed model fit by REML ['lmerMod']
## Formula: JOBSAT ~ (PAY | GRPID)
##    Data: dta
## 
## REML criterion at convergence: 3381
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -2.9050 -0.6367 -0.0033  0.6236  3.8665 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev. Corr 
##  GRPID    (Intercept) 0.8385   0.9157        
##           PAY         1.0682   1.0335   -0.35
##  Residual             4.4165   2.1016        
## Number of obs: 750, groups:  GRPID, 50
## 
## Fixed effects:
##             Estimate Std. Error t value
## (Intercept)   0.2806     0.1456   1.927
sjPlot::tab_model(m1, show.p=FALSE, show.r2=FALSE)
  JOBSAT
Predictors Estimates CI
(Intercept) 0.28 -0.01 – 0.57
Random Effects
σ2 4.42
τ00 GRPID 0.84
τ11 GRPID.PAY 1.07
ρ01 GRPID -0.35
ICC 0.16
N GRPID 50
Observations 750 
#圖2.1
ggplot(dta_pay, aes(JOBSAT, PAY, group= GRPID)) +
  geom_point(alpha=.2)+
  geom_line(alpha= .2) +
  stat_smooth(aes(group= GRPID), method="lm", formula=y~x, se=FALSE,
              size=rel(.5))+
  stat_smooth(aes(x=JOBSAT, y=PAY, 
                  group=1),
              method="lm", formula=y~x, 
              se=FALSE, size=rel(.5), col= "red")+
  labs(x="Job Satisfaction)",
       y="Pay")+
  theme_minimal()

#資料整理
dta_NEGL <- klein2000 %>% 
         group_by(GRPID) %>% 
         mutate(mneglead=mean(NEGLEAD))
#計算組間工作滿意度的差異有多大可歸因於組間(組級)消極領導的差異
m2 <- lme4::lmer(JOBSAT ~ (NEGLEAD | GRPID), data=dta)
summary(m2)
## Linear mixed model fit by REML ['lmerMod']
## Formula: JOBSAT ~ (NEGLEAD | GRPID)
##    Data: dta
## 
## REML criterion at convergence: 3431.7
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -2.8616 -0.6544 -0.0075  0.6254  3.2700 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev. Corr 
##  GRPID    (Intercept) 0.5517   0.7428        
##           NEGLEAD     0.4205   0.6484   -0.21
##  Residual             5.0311   2.2430        
## Number of obs: 750, groups:  GRPID, 50
## 
## Fixed effects:
##             Estimate Std. Error t value
## (Intercept) 0.004284   0.133794   0.032
sjPlot::tab_model(m2, show.p=FALSE, show.r2=FALSE)
  JOBSAT
Predictors Estimates CI
(Intercept) 0.00 -0.26 – 0.27
Random Effects
σ2 5.03
τ00 GRPID 0.55
τ11 GRPID.NEGLEAD 0.42
ρ01 GRPID -0.21
ICC 0.10
N GRPID 50
Observations 750 
#圖2.2
ggplot(dta_NEGL, aes(JOBSAT, NEGLEAD, group= GRPID)) +
  geom_point(alpha=.2)+
  geom_line(alpha= .2) +
  stat_smooth(aes(group= GRPID), method="lm", formula=y~x, se=FALSE,
              size=rel(.5))+
  stat_smooth(aes(x=JOBSAT, y=NEGLEAD, 
                  group=1),
              method="lm", formula=y~x, 
              se=FALSE, size=rel(.5), col= "red")+
  labs(x="Job Satisfaction)",
       y="Negative Leadership")+
  theme_minimal()

3.結論

  1. 由圖2.1可以看到，工作滿意度與組間薪酬呈正向關係，薪酬越高，工作滿意度越佳。
  2. 由圖2.2展示了工作滿意度與消極領導間的關係，呈負相關，消極領導程度越低，工作滿意度越佳。