I am analyzing survey data, of which I am a part. it is causing something of a philosophical quandary. because, in this data, there are at least 30 self-identified identity categories. on the other hand, if I use more than five, the models needed to analyze the data explode or die. in respect to the complexity and dynamism of identity as related to gender and sexuality (and the beauty of the human mind as model builder), I am interested in each of these identifications and their inter-relations--and it seems reasonable to give each of them their shot, analytically. in respect to making a set of results that are human-readable, and statistically valid in at least a vaguely defensible way, the number of categories really cannot be more than five. the contents of my re-code category are enough to start a hell-war in any gender studies department. this is the analyst’s lament, whenever one is dealing with large-N data. what is a “country”? what is X ethnic category? in reality: both of these categories are very messy. irony #1: I am in this survey data, and I chose to use the textbox, to demonstrate to whoever would analyze the data that the subject matter is complex, and draw their mind into a space of complexity during analysis. irony #2: I am one of the people who is analyzing this survey data.
