Data Ethics

A pollster's graph shows a bar twice as tall as its neighbor — 'support DOUBLED!' Look closer: the y-axis starts at 4.9 million instead of 0. The real increase is 4.9 to 5.0 million — a 2% blip, painted as a landslide. The math wasn't wrong; the PICTURE lied. Statistics gives enormous power to summarize, and with that power comes the temptation to mislead — by truncating axes, cherry-picking windows, or collapsing groups in ways that reverse the truth.