We have a Masters student at York working on the psychometrics of complicating sex and gender beyond M/F for psychological measures. I'm excited to see what they develop.
Correspondence analysis <http://www.datavis.ca/courses/grcat/grc5.html> (for categorical data) or multilevel modeling <http://www.biostat.jhsph.edu/~fdominic/teaching/bio656/lectures/1.intro.pdf> (for mixed designs) would offer viable solutions to include n levels of a gender variable. Or, you could run it as a regression, if the researcher had reason to consider (and ability to measure) gender identity as a continuum or spectrum. I absolutely would not be interested in interpreting a 5 way (or even 56 way <http://www.slate.com/blogs/future_tense/2014/02/13/facebook_custom_gender_options_here_are_all_56_custom_options.html>) interaction ;) Arlie -- Arlie R. Belliveau, MA History & Theory of Psychology Doctoral Student York University Department of Psychology 059 Behavioural Science Building 4700 Keele St. Toronto, ON ar...@yorku.ca www.arliebelliveau.com On Tue, May 12, 2015 at 1:53 PM, Mike Palij <m...@nyu.edu> wrote: > On Tue, 12 May 2015 09:48:08 -0700, Christopher Green wrote: > >> And, since ANOVA is merely a constrained special case of the >> General Linear Model, there is not even really a practical limit >> on the number of values a continuous "gender" variable might >> assume. >> > > Okay, this is getting a little bit silly. Not the General Linear Model > or how many levels ANOVA can handle (Advice: use few independent > levels with few levels; you don't know what a headache is until you > try to explain a five-way interaction) but the more fundamental issue > of "what is the question that one wants to get answer to?" Michael > Scoles below clearly points in this direction and Jim Clark adds > to it with the concept of asexuality but is one concerned with: > > (1) Traditional gender role self-identification (i.e., male or female > or some mixture of the two). Note: this is primarily a psychological > variable created through a developmental history within certain > cultural environments. That is, a personally defined gender > role consistent with one's experience with one's body and > with the cultural strictures concerning sex. > > (2) Genetically based sex differences (XX vs XY vs variations). > Do genes have an influence on sexual orientation and behavior? > Maybe, maybe not, see: > http://www.councilforresponsiblegenetics.org/ViewPage.aspx?pageId=66 > > (3) Sexual behavior (Males having Sex with Females or MSF, > Males having Sex with Males or MSM, Females having Sex > with Females or FSF, Sex with both Same Sex and Opposite > Sex, Sorry No Sex -- We're British, etc.). > Note: Sexual behavior does not have to reflect sexual orientation > as AIDS/HIV research have shown: males who define themselves > as traditional heterosexuals may engage in sex with other > males ["on the down low"] and deny any homosexual interests or > attractions > Note #2: Given the adaptability of human behavior, including > sexual behavior, it should come as no surprise that males and > females may define themselves one way ("straight","gay/lesbian", > etc.) but still behave in ways that are inconsistent with that > definition. For example, "straight" males and females when put > into situations where access to the opposite sex is limited, may > engage in homosexual behavior without changing their sexual > orientation. While in prison, some people become "gay for > the stay", that is, will engaged in homosexual behavior while in > prison but return to heterosexual behavior when released. > There is "gay for pay", that is, one engages in same sex acts > to earn money but engages in heterosexual personal relationships. > See also paragraph #3 in the CRG article I link to above: > During one phase of one's life one is one way but during a > later phase they're another. > > So, which is it: > (a) Self-definition/cultural definition of sexual role and identity > (b) Genetic definition of sexual role and identity > (c) Behavioral definition of sexual role and identity > > Or are all three involved but, as is the human tendency, we tend > to oversimplify things just to be able to get through the day because > we have too many other things to consider and do as well. > > By the way, I've taken to saying in my research methods and statistics > classes "Let's use 'Sex/Gender' as a grouping variable and, for > purposes of simplification, let's just assume there are two levels. > We can complicate things after we get through the basics." > > -Mike Palij > New York University > m...@nyu.edu > > P.S. I'll leave the category of "bi-curious" to someone else. > I'm tuckered out from thinking so much about sex. ;-) > > > On May 12, 2015, at 10:11 AM, Jim Clark wrote: >> Hi >> >> Since ANOVA handles any number of levels of a factor (2, 3, 7, 24? .), >> ANOVA >> need make no adjustment to accommodate more complex models of human sexual >> orientation. It is the researcher who must decide what is the appropriate >> number of levels and expected pattern of results, all of which is happily >> handled by ANOVA. >> >> To further complicate matters, there is a growing literature on Asexuality >> that might be relevant here, depending on how one is defining the >> different >> levels of the sexual orientation factor. That is, if it is defined by the >> targets of people's sexual attraction, then clearly Asexuality needs to be >> accommodated. Here's a link to Tony Bogaert's book: >> >> http://www.amazon.com/Anthony-F.-Bogaert/e/B007LU54TY >> >> On Tuesday, May 12, 2015 9:03 AM, Michael Scoles wrote: >> >> Dear Sir or Madam: >> >> You raise an interesting question, but in order to answer it, more >> information is needed. >> (1) What is the dependent variable and what is the reason for >> investigating >> its association with gender? >> (2) From your description, it appears that there is a second independent >> variable, with two levels. What is this variable and what are the levels? >> (3) LGBT is an acronym with four letters (some would add Q), but you only >> include 3 levels for gender in your proposed design. Which group are you >> choosing to offend? >> >> Thank you for any clarification you can provide. >> >> >> >> On Tue, May 12, 2015 at 3:46 AM, michael sylvester wrote: >> >> With Bruce Jenner in mind.ANOVA may have to make some adjustmes to >> accommodate >> LGBT subjects. Currently the gender variable N2 male/female but with LGBT >> subjects coming into the subject pool we may be lookimg not at a 2x2 >> design >> but a 3x2 design.Of course ANOVA >> measures simple,main,and interaction effects.Those analyses would be very >> interestingWith LGBT a subset of the gender paradigm,when does the he or >> the >> she kicks in. >> michael >> daytona beach.florida >> 'going beyond where no tipster has gone before.' >> >> > > > --- > You are currently subscribed to tips as: arliebelliv...@gmail.com. > To unsubscribe click here: > http://fsulist.frostburg.edu/u?id=13091.5fbcba6b94d471416a45a34246e4403b&n=T&l=tips&o=44735 > or send a blank email to > leave-44735-13091.5fbcba6b94d471416a45a34246e44...@fsulist.frostburg.edu > --- You are currently subscribed to tips as: arch...@mail-archive.com. To unsubscribe click here: http://fsulist.frostburg.edu/u?id=13090.68da6e6e5325aa33287ff385b70df5d5&n=T&l=tips&o=45198 or send a blank email to leave-45198-13090.68da6e6e5325aa33287ff385b70df...@fsulist.frostburg.edu
