gious Harvard Educational Review, thisrnre-evaluation by E.R. House, G.V. Glass,rnL.F. McLean, and D.F. Walker arguedrnthat DFs superioritV' was not very statisticallyrnsignificant if the data were re-adjustedrnin various ways, was due more tornenvironment than to the method used,rnand that in any case, while . . .