Linear Statistical Models: Regression

Raw Scores, Deviation Scores & Standard Scores


Raw Scores

The observations in a data set. Usually denoted in upper case letters; X, Y, etc.

Sample Mean

Deviation Scores

Deviation scores are raw scores with the sample mean subtracted from them. Deviation scores are usually denoted with a lower case letter; x, y, etc. The mean of deviation scores is always zero.

Standard Scores

Standard scores, also called z-scores, are deviation scores divided by the standard deviation. Standard scores are usually denoted as z. The mean of standard scores is always zero and the standard deviation is always one.

Comparing Raw Scores, Deviation Scores & Standard Scores

Example:

input X
96
87
89
79
92
91
end

summarize X

Variable |     Obs        Mean   Std. Dev.       Min        Max
---------+-----------------------------------------------------
       X |       6          89   5.761944         79         96 

return list

scalars:
       r(N)        =  6
       r(sum_w)    =  6
       r(mean)     =  89
       r(Var)      =  33.2
       r(sd)       =  5.761944116355173
       r(min)      =  79
       r(max)      =  96
       r(sum)      =  534

generate x = X - r(mean)

generate z = x/r(sd)

format X x z %4.2f

list

       X       x      z 
  1. 96.00    7.00   1.21  
  2. 87.00   -2.00  -0.35  
  3. 89.00    0.00   0.00  
  4. 79.00  -10.00  -1.74  
  5. 92.00    3.00   0.52  
  6. 91.00    2.00   0.35   
  
summarize, format

Variable |     Obs        Mean   Std. Dev.       Min        Max
---------+-----------------------------------------------------
       X |       6       89.00       5.76      79.00      96.00  
       x |       6        0.00       5.76     -10.00       7.00  
       z |       6        0.00       1.00      -1.74       1.21 


Linear Statistical Models Course

Phil Ender, 5Jan98