it looks like for the menu study we have users with dramatically different menu use patterns. It might be helpful to look at each intense, moderate, and infrequent menu users independently and study there interactions as separate analysis. for example we could test theories that: -intense menu users use bookmarks a lot and that drives the bulk of their menu use and/or -infrequent menu users don't use bookmarks. to get a this kind of data easily it would be good to add total count of menu interactions for each user to the data files. the following is an example for user zero having 4 menu interactions, user one having 2 interactions, and user two having 3 interactions user, os, fx_version, ui_method, start_menu_id, explore_ms, explore_num, menu_id, item_id, timestamp , total_interactions 0, WINNT Windows NT 6.1, 3.6, 0,4,28691,5,-1,-1,1266044868218,4 0, WINNT Windows NT 6.1, 3.6, 0,0,0,0,1,18,1266045660234,4 0, WINNT Windows NT 6.1, 3.6, 0,0,0,0,1,17,1266045665637,4 0, WINNT Windows NT 6.1, 3.6, 0,0,0,0,1,18,1266047306684,4 1, WINNT Windows NT 6.1, 3.6, 0,4,4332,2,5,64,1266047468720,2 1, WINNT Windows NT 6.1, 3.6, 0,0,0,0,1,18,1266047562462,2 2, WINNT Windows NT 6.1, 3.6, 0,4,3098,2,-1,-1,1266047798503,3 2, WINNT Windows NT 6.1, 3.6, 0,5,2124,1,5,64,1266047802801,3 2, WINNT Windows NT 6.1, 3.6, 0,5,1464,1,5,64,1266047815459,3 post processing scripts could then easily fliter the data by the number of interactions. This kind of group by interaction frequency/count might also apply to other studies.
Product: Mozilla Labs → Mozilla Labs Graveyard
You need to log in before you can comment on or make changes to this bug.