1. Before the project start

2. what did I do?

3. And what will happen?

There are two main parts in this project

1. Continuous Data

2. Categorical Data

This dataset contains 2042 observers with 27 variables

1. Individual information: COMPID & SUB

2. Individual Scores from each question of three different different scales

3. Mean Scores of individual's each scale

4. Mean Scores of individual's each scale Aggregated by company

## Linear mixed model fit by REML ['lmerMod']
## Formula: Score ~ Type - 1 + (1 | COMPID) + (1 | COMPID:SUB)
##    Data: dta2L
## 
## REML criterion at convergence: 15926.5
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -3.4461 -0.5144  0.1037  0.6007  2.8497 
## 
## Random effects:
##  Groups     Name        Variance Std.Dev.
##  COMPID:SUB (Intercept) 0.32419  0.5694  
##  COMPID     (Intercept) 0.05413  0.2327  
##  Residual               0.55410  0.7444  
## Number of obs: 6126, groups:  COMPID:SUB, 2042; COMPID, 49
## 
## Fixed effects:
##              Estimate Std. Error t value
## TypeLEAD      3.04488    0.04059   75.01
## TypeRHOSTILE  4.09716    0.04059  100.94
## TypeTSIG      3.17721    0.04059   78.27
## 
## Correlation of Fixed Effects:
##             TyLEAD TRHOST
## TypRHOSTILE 0.835        
## TypeTSIG    0.835  0.835

## Linear mixed model fit by REML ['lmerMod']
## Formula: Score ~ (1 | Type) + (1 | COMPID) + (1 | COMPID:SUB)
##    Data: dta2L
## 
## REML criterion at convergence: 15931.1
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -3.4454 -0.5147  0.1039  0.6003  2.8494 
## 
## Random effects:
##  Groups     Name        Variance Std.Dev.
##  COMPID:SUB (Intercept) 0.32419  0.5694  
##  COMPID     (Intercept) 0.05413  0.2327  
##  Type       (Intercept) 0.32822  0.5729  
##  Residual               0.55410  0.7444  
## Number of obs: 6126, groups:  COMPID:SUB, 2042; COMPID, 49; Type, 3
## 
## Fixed effects:
##             Estimate Std. Error t value
## (Intercept)    3.440      0.333   10.33

## [[1]]
## Linear mixed model
##  Family: gaussian (identity)
## Formula: Score ~ Type - 1 + (1 | COMPID) + (1 | COMPID:SUB)
## 
##   ICC (COMPID:SUB): 0.347688
##       ICC (COMPID): 0.058051
## 
## [[2]]
## Linear mixed model
##  Family: gaussian (identity)
## Formula: Score ~ (1 | Type) + (1 | COMPID) + (1 | COMPID:SUB)
## 
##   ICC (COMPID:SUB): 0.257164
##       ICC (COMPID): 0.042937
##         ICC (Type): 0.260360

## Call: ICC(x = dta2[, c(3, 4, 5)])
## 
## Intraclass correlation coefficients 
##                          type  ICC   F  df1  df2 p lower bound upper bound
## Single_raters_absolute   ICC1 0.23 1.9 2041 4084 0        0.20        0.26
## Single_random_raters     ICC2 0.30 3.0 2041 4082 0        0.13        0.44
## Single_fixed_raters      ICC3 0.40 3.0 2041 4082 0        0.38        0.43
## Average_raters_absolute ICC1k 0.47 1.9 2041 4084 0        0.43        0.51
## Average_random_raters   ICC2k 0.56 3.0 2041 4082 0        0.31        0.70
## Average_fixed_raters    ICC3k 0.67 3.0 2041 4082 0        0.64        0.69
## 
##  Number of subjects = 2042     Number of Judges =  3

This dataset contains 434 observers with 100 variables

1. Individual information: ID, SEX, Class, Age

2. Individual Response from each question of three different different

3. Three different scales are :Anxiety Sensitivity Index & Anxiety Sensitivity Profile & State Trait Anxiety Inventory Trait Version

##  Fleiss' Kappa for m Raters
## 
##  Subjects = 242 
##    Raters = 12 
##     Kappa = 0.138 
## 
##         z = 32.7 
##   p-value = 0

This dataset contains 30 observers with 6 variables

1. 30 subject

2. 6 raters

3. 5 levels:Depression, Personality Disorder, Schizophrenia, Neurosis, Other

##  Fleiss' Kappa for m Raters
## 
##  Subjects = 30 
##    Raters = 6 
##     Kappa = 0.43 
## 
##         z = 17.7 
##   p-value = 0

1. The results from multilevel approach are similar to ICC caculated from data

2. Multilevel approaches can give us more information

3. But multilevel approaches are more complicated and may need more data to get the results