and why were they slightly east of the centre line?
They weren't.
The trees they hit and where they first impacted the ground are both exactly on the extended centre line of RWY18.
On the question of what they could or could not see once they were below the scattered cloud, a witness in the tower (possibly a controller) could see the aircraft's landing lights between the first impact and the impact with the ground.
The distance from the tower to a position between Mrs Benson's house and the ground impact position is roughly double the distance from Mrs Benson's house to the PAPIs. Does anyone have any data on relative intensities between PAPI and landing lights? I do know that maintenance staff are advised to wear welding goggles when doing a daylight ground verification check on a PAPI but have never heard of that being recommended when within 20-30 metres of an aircraft landing light during daylight, so I will make the assumption that PAPIs are quite a bit brighter than a landing light.
Anyway, the point I am making is that in the absence of any other limiting factors, if the controller could see the landing lights from the tower, the flight crew should have been able to see the PAPIs at half that distance.
We know the PAPIs were switched on and working post crash, and we know that the were correctly calibrated (to within 1/100th of a degree).
Do we know that they were switched on during the approach?
If they were operating and we assume that the CVR recording ended around the time of impact with the ground, then the call "runway in sight", 13 seconds before that, would have been made half a mile from impact at 140KTS and a couple of seconds before they hit the trees.
Is it possible that they were so low that their view of the runway was obscured by the line of trees at the Tarrant-Huffman/Treadwell Road junction until 3 seconds before they hit those trees?