First-differenced and de-meaned data
- Data is simply the differences (change scores) between time points (see Table 11.2)
- trust.06.07 = trust.2007 - trust.2006
- First-differenced: Individual’s change from t0 to t1 (outcome and treatment variable below)
- Same is done for covariates
Table 11.2: Data change scores (wide-format)
|
Name
|
trust.06.07
|
trust.07.08
|
threat.06.07
|
threat.07.08
|
|
Brittany
|
0
|
-2
|
0
|
1
|
|
Ethan
|
1
|
-2
|
-1
|
0
|
|
Kyle
|
7
|
-2
|
0
|
0
|
|
Jacob
|
-2
|
3
|
1
|
0
|
|
Jessica
|
2
|
-5
|
0
|
0
|
- Data is de-meaned for fixed effects approach (see Table 11.3)
- trust.2006.demeaned = trust.2006 - trust.mean (across all time points)
- De-meaned = individual’s deviation for each time point from its mean across time points (outcome and treatment variable below)
- Same is done for covariates
Table 11.3: Data demeaned (wide-format)
|
Name
|
trust.2006.demeaned
|
trust.2007.demeaned
|
trust.2008.demeaned
|
threat.2006.demeaned
|
threat.2007.demeaned
|
threat.2008.demeaned
|
|
Brittany
|
0.6666667
|
0.6666667
|
-1.333333
|
-0.3333333
|
-0.3333333
|
0.6666667
|
|
Ethan
|
0.0000000
|
1.0000000
|
-1.000000
|
0.6666667
|
-0.3333333
|
-0.3333333
|
|
Kyle
|
-4.0000000
|
3.0000000
|
1.000000
|
0.0000000
|
0.0000000
|
0.0000000
|
|
Jacob
|
0.3333333
|
-1.6666667
|
1.333333
|
-0.6666667
|
0.3333333
|
0.3333333
|
|
Jessica
|
0.3333333
|
2.3333333
|
-2.666667
|
0.0000000
|
0.0000000
|
0.0000000
|