Accuracy problem with deltaT

I posted a number of accuracy test lately. But now some
results have appeared that deviate. It concerns the data
on 96-072A (USA 129) of the 8th.

The problem is in the delta T values. They come out some
0.3-0.5s too early compared to other data gathered by
other stations around that time. It concerns data from
two separate images.

In below table, the asterisk-marked points are mine. The
other points are by David Brierley and Peter Wakelin.
The given delta T values are relative to Ted Molczan's
elset 05253.13803400.

deltaT
-0.10
-0.11
-0.37 *
-0.29 *
-0.48 *
0.04
0.06
0.07
0.07
0.08
0.11
0.05
0.10
-0.02
-0.11
-0.08
-0.03
-0.03
-0.03
-0.03
0.04

I am at a loss to explain this deviation.
Something in the timing is not right, but what?

The crosstrack value of the second point initially was
somewhat off (0.07 degree). This turned out to be because
I had switched the image reference time and object position
time in Astrorecord. When this was corrected, the crosstrack
value came out fine (but the delta T difference remained).