Recent work in regression analysis have embraised an approach to high dimensional data analysis that consists of selecting at random subsets with a relatively small number of predictors, doing variable selection and/or statistical inference on each subset, and then merging the results from the subsets. The merging may involve further variable selection and/or statistical inference on the the merged subsets. This approach makes it possible to construct methods for high dimensional data analysis using methods that were disigned for small dimensional data. This talk will present such constructions and examine their properties.