I’m no expert, but this seems like a fairly large oversight:

The authors failed to notice a curious aspect of their results. The difference between the numbers of cigarettes smoked daily by the high- and low-nicotine groups was attributable not to a reduction in the number of cigarettes smoked by the latter group by comparison with their normal, pre-experiment levels, but by an increase in the number of cigarettes smoked by the former compared with their pre-experiment levels. Why this should have been the case is mysterious, but it was so.