Two decades of little change: An analysis of U.S. medical school basic science faculty by sex, race/ethnicity, and academic rank.

In the analysis of time alone, the positive linear increase across time became only marginally significant (β = 0.02, p = 0.0884) after adjusting for a significant positive AR1 autocorrelative term (β = 0.8716, p<0.0001; the Durbin-Watson test of lag 1 autocorrelation was significant before adjusting for the AR1, but not after); the autocorrelation term absorbed much of the variance due to the deviation from a straight-line trend.

also in 5,076 other papers