> 'propaganda' leaflet has left public LESS well informed, study shows Two polls at different points in time, without any information about methods and samplin

> 'propaganda' leaflet has left public LESS well informed, study shows Two polls at different points in time, without any information about methods and sampling error, gave slightly differing results (23% vs. 21%). This leaflet campaign took place between them. This isn't a very good argument: Weak correlation in noisy data does not mean causation. (Not saying the government was right to do this this way. But calling the thing misinformation that confused voters based on this comparison of polls is a bit exaggerated.)