Often misinterpretations of instrument display represent differing views between design and operational use. Additionally, differences in how pilots have been taught to use the instrument or the meaning of specific indications for special situations.
If the inference is that a commercial sky pointer is used to recover roll attitude, I offer a different view, in that the primary roll reference is the horizon. Furthermore, a sky pointer should be used against its own reference (not the sky), in some cases this could be considered to be the aircraft fin – the fixed reference on the outside of the instrument – which is part of the aircraft.
Thus without the details of the ‘experiment’ – pilot experience, beliefs about system operation, or the task, then the results can be used out of context.
The objective of the report (Singer / Dekker) appears to seek further understanding of the Saab 340 accident involving LOC after takeoff and during a departure turn. IIRC the official report cited East/West differences in display type against the pilot’s experience (East) as a contributory cause. This problem as I understand, is between ‘inside’ and ‘outside’ types of displays and not specifically about sky pointers.
Also, IIRC the Saab 340 was one of the first aircraft with EFIS which used a full screen display of attitude vs a containing instrument. I found that at larger bank angles this type of display had some distracting features when the horizon transitioned from the side edge of the EFIS to the top/bottom edge.