Time1-Time2 regressions

Example from
Rogosa, D. R. (1995). Myths and methods: "Myths about longitudinal
research," plus supplemental questions. In The analysis of change, J. M.
Gottman, Ed. Hillsdale, New Jersey: Lawrence Erlbaum Associates, 3-65. 

MTB > describe c2-c5
        N     MEAN   MEDIAN     MIN      MAX       Q1       Q3
C2     40    44.16    43.90   30.45    59.35    38.18    49.56
C3     40    54.21    53.63   39.87    67.71    48.61    60.45
C4     40    64.27    63.21   42.06    81.84    57.90    73.01
C5     40   14.992   15.200   9.462   20.399   13.288   16.837


MTB > corr c2-c5
   
          C2       C3       C4
C3        0.842
C4        0.536    0.907
C5        0.766    0.765    0.598



MTB > note let's look at exog variable in c5


MTB > regress c4 2 c5 c3
The regression equation is
C4 = 0.68 - 0.757 C5 + 1.38 C3
Predictor       Coef       Stdev    t-ratio        p
Constant       0.683       4.555       0.15    0.882
C5           -0.7570      0.3329      -2.27    0.029
C3            1.3822      0.1290      10.72    0.000
s = 3.752       R-sq = 84.4%     R-sq(adj) = 83.5%




MTB > regress c4 2 c5 c2
The regression equation is
C4 = 31.2 + 1.50 C5 + 0.239 C2
Predictor       Coef       Stdev    t-ratio        p
Constant      31.213       7.546       4.14    0.000
C5            1.5004      0.6678       2.25    0.031
C2            0.2392      0.2587       0.92    0.361
s = 7.514       R-sq = 37.3%     R-sq(adj) = 33.9%



MTB > note use W with fallible data
MTB > regress c24 2 c25 c22
The regression equation is
C24 = 25.7 + 2.14 C25 + 0.145 C22
Predictor       Coef       Stdev    t-ratio        p
Constant      25.690       8.821       2.91    0.006
C25           2.1431      0.7181       2.98    0.005
C22           0.1447      0.2759       0.52    0.603
s = 8.770       R-sq = 38.8%     R-sq(adj) = 35.5%