Does anyone know how I go about performing a oneway ANOVA with a linear monotonic contrast when I only have summary data for each group (i.e. n, mean, sd).

I know how to do a standard oneway ANOVA in this situation:
(see http://www-01.ibm.com/support/docvie...id=swg21476127)

But how do I extend this approach to include the situation where I seek a relationship between groups (e.g. most of my scenarios have 4 groups, so I would be seeking something like 1 > 2 > 3 > 4)