>I think Tog's conclusions (the single set of studies put forth whenever this >thing >comes up) are poorly made ...
it turns out that there is rather older work that supports much the same conclusion, which i probably saw mentioned in HCI Remixed, since that's one i've read recently.
