Compare the effect of pooling data by year vs. averaging across the decade (year beginning in 0 to year ending in 9), using feminism as an example. Without the data from 2000 - 2002, the yearly graph still shows a clear downward turn, while the graph by decade would suggest purely exponential increase.