Okay, I know that, for most normal people in the real world, “interaction” involves nice things like meeting other people, being sociable, building up relationships, conversation.
In the exciting world of live research though, the interactions one finds oneself most interested in are those of the statistical variety.
Over the last few grim and miserable months of incomplete datasets and partial results (yeah, you have no idea), I’d basically given up hope of seeing the interaction I wanted. I’d resigned myself to a the prospect of a write-up based on a long series of ambiguous and unsatisfactory outcomes.
Now, I don’t want to count my chickens or anything, but this past week I finished the data collection phase and finalised those spreadsheet tables containing all my precious reaction time results and accuracy values, and did the statistical analysis.
It wasn’t the interaction I’d been wanting, but it’s an amazingly good second-best. It’s back to the old difference you get when you say out loud the word hotdog (ie, food) versus hot dog (animal). My participants were divided into two groups according to various criteria, and the group which the theory said should have been particularly bad at telling the difference between these two sound patterns, turned out to be better, in fact, than the other group. (More specifically, although both groups were equally accurate in telling the difference, this group was significantly faster than the other one to make their decision.)
Admittedly, again, I need to hedge all this round with many many reminders to self about counting and hatching – I may well have enthusiastically put the wrong figures into the stats program or, indeed, horrid thought, labelled the groups back to front.
Anyway, in the hope that it isn’t too good to be true, I just though I’d share.