Nothing meaningful is ever simple: The danger of dichotomizing variables

Quantitative research looks for mathematical relationships and statistically significant differences between neatly defined concepts, called variables.  It is often described as elegant, simple, and therefore definitive and scientific, and here’s where I’m going to start throwing down. Partly because the idea that simpler equates to more scientific, and therefore more correct, strikes me as inherently patriarchal, racist and oppressive, and also because as a researcher, I find it inaccurate and misleading.  And when quantitative research depends on forcing complex realities into mutually exclusive categories, I have to ask the question: is anything worthwhile in life ever binary?  Other than the clear fact that my nephew is not allowed to climb into the seal tank at the aquarium, no matter how much he wants to, no matter how angry he is with me for holding him back, no matter how much he wants to pet the seals and begin a new life with less mean, restrictive and earthbound companions, I’m having trouble fitting much about our world into such narrow boxes.  And my nephew was 2 when I made this binary decision on his behalf.

But let’s talk patriarchy, white supremacy and oppression first, ok? Let’s talk about the imperialist idea of one simple truth, belonging solely to those who have grabbed power.  And let’s talk about the horrific systems that have been built to maintain the fiction of one simple truth, enforced and policed by those same people in positions of unearned power.  This is the same appropriation of science that brought us researchers filling kidnapped skulls with marbles to “prove” that race was more than a social construct used to oppress and enslave; race was a single elegant truth that divided a small group of “superior” producers of knowledge from the “inferior” masses whose homes then became ripe for the taking, because science.

Race as a single, elegant biological reality justified racism, and allowed white supremacy to grow, unchecked, sold as the single truth in a colonial world, arbitrated and enforced by those who benefitted from it, and thus clung it, eschewing complexity. Complexity might have demanded re-thinking imperialism, and the imperialists’ claims to all the resources, including biological superiority, and including the multiple conflicting truths that lay buried just under the ground they stole.  Complexity and multiple truths would have acknowledged genocide, smallpox infested blankets, Indian schools and broken treaties as contributors to the declining number of Indigenous people in the United States, but race as a single, elegant biological reality propelled the myth that biological inferiority alone explained the deaths. And with this simple, elegant explanation, those in power maintained their false claims to the superiority which justified land grabs and smallpox in the blankets.

Paul Broca, the scientist with the marbles, is widely hailed for joining anthropology and medicine to make genius discoveries, including the new tools he invented for measuring bodies.  The fact that he used his supposed genius to promulgate the ridiculous idea that skulls from Black people hold less marbles than white skulls and are therefore less human is generally not mentioned, or relegated to a footnote about different times. Different times when a single, simple idea that reduced complexities to enforce dominant, oppressive ideologies was hailed as genius. Different times, if by different we mean exactly the same, such as this study by Gapstur et al from 2002, which found that young cis Black men had a 3% higher rate of testosterone than their white counterparts, though this difference went away once researchers looked at the men’s waist size, which was then used by a racist (I’ll spare you a link to his blog) to explain mass incarceration of Black men. That’s right, it’s not institutional racism in law enforcement and courts, or the next clever morphing of Jim Crow (as explained by actual genius Michelle Alexander), it’s a 3% difference in average testosterone rates causing men of color to be significantly more violent and thus get arrested more.  In other words, this simplistic, inaccurate, ridiculous thinking that Paul Broca’s biographers consider a historical footnote, if that, is alive and well, especially when we consider that young cis men’s testosterone levels  fluctuate throughout the day, with higher variation than Gapstur et al noted among Black and white men in his sample. And lest you think that the use of quantitative measurements of bodies as a means of promoting the simple elegance of white supremacist bullshit, a quick search led me to peer reviewed studies of racial differences in body mass index that researchers explained as solely due to race, not inequitable distribution of resources, not institutionalized racism, not food desserts, but race as a simple, elegant biological fact makes people of color bigger, and thus deserving of health inequities.

And let’s throw in gender-based oppression, and the Victorian ideal of binary white gender roles, where emotional, fanciful women suffered from hysteria whenever we noticed or voiced our oppression, and rational, intellectual men were always there to comfort us with their elegantly straightforward thinking.  And thus enforce the idea that gender was a binary of superior and inferior, with mediocre men in positions of unearned power continuing to call the shots because they owned the singular, simple, elegant truth.  And then the binary concept of gender was used to further enforce white supremacy and ableism, by casting white women who didn’t accept the simple, elegant truth as singular and godlike as morally disabled, and then as less than white, and therefore even less able to grasp the “truth” of mediocre white men insisting on maintaining unearned power.

So at this point, I am heartily and fully rejecting the idea of simple, stripped of complexity and elegant as equivalent to scientific and definitive.  And I am defining scientific as it was before it was appropriated by people like Paul Broca and his few friends in positions of unearned power, as the intelligent, messy, complex, holistic, intersectional and community-owned process of understanding our messy, complex, intersectional world it’s always been.

Moving on to the inaccuracies inherent in simplifying complex realities,  much of quantitative research rests on mutually exclusive categories and binaries, grouping people and their responses into simple yes or no categories: smoker/non-smoker, male/female, disabled/healthy, white/non-white, superior/inferior etc.  And how many of these binaries actually exist in the real world? Other than the binary rules we apply to children, like my binary, fixed decision about my nephew and the seal tank, nothing real is binary.  Not gender, when participants in my brilliant friend Kimberlynn Acevedo’s study self-described their gender in 70 distinct categories; not disabled/healthy, because health itself is an ableist, capitalistic construct that defines well-being in relationship to how much one produces, meaning that the definitions of disability and ablebodiedness are both personal and infinite.  Even a binary as seemingly straightforward as smoker/non-smoker begs questions: when do I become a smoker, exactly? Are we talking ever taken a puff, Bill Clinton’s exemption of never having inhaled, quit 10 years ago, less than a pack a day, and who is counting our packs? How many puffs make me a smoker, and when do I cross the line from non-smoker sitting on a porch in the summer with an old friend I’ve missed, relieving our wilder youth as we exchange memories and matches, sharing cigarettes and stories in the warm breeze?

And then there is the binary of race, which perpetuates white supremacy by defining white people as the norm and everyone else as a collective other. This is an example of a binary that marginalizes and erases.  First of all, are there not people who are both white and non-white?  Must children of white and non-white parents always pick just one identity? Does my nephew (of seal tank fame) need to disavow one of his parents to be counted in research?  Secondly, what else are we missing in this simplistic dichotomy? And finally, hard stop, we get better data when we see more of people, not less.  We come to better, more reliable conclusions when we stop erasing.  Our world is complicated, and so our variables need to be, too.

For right now, when our research methods are so elementary, complex variables make analysis harder. Many of the common quantitative statistical tests will not work unless they are fed binary variables.  For your information, these are odds ratios (whenever you see the abbreviation OR= whatever in study findings), independent t-tests and paired sample t-tests (whenever you see the abbreviation t=whatever in study findings) and ANOVA analysis (whenever you see the abbreviation f=whatever in study findings).  Some regression models also depend on binary variables, and authors will generally tell you when they are using regression.

And so in order to use these commonly accepted (and often demanded by reviewers and publishers) analyses, researchers are forced to do what’s called dichotomizing their data.  Essentially this means taking more complex or expansive categories and collapsing them down into two options, only.  So in Kimberlynn’s ground breaking study with over 70 genders present, she would have had to force all participants’ identities into two false categories, say femme-identified and other.  There would have been a whole lot of other, and a whole lot of misclassification and loss. Kimberlynn and her team made the revolutionary and important decision not to do this, and thus held on to their rich findings, a brave integrity move that respected their participants.  (Serious shout out to Kimberlynn and her team, with love and admiration.)  Other researchers will ask people to rank their satisfaction with their care provider on a scale of 1 to 10, then divide the answers into low and high satisfaction groups, without exploring how a 1 might be different from a 5, or the rich data behind why someone gives their provider an 7 instead of a 9.  And this means we are losing a lot of knowledge to an oppressive convention.

To me, this suggests that we desperately need new analytic approaches that are powered to deal with the actual complexity of the real world.  We need expansive variables, with a lot of fill in the blanks, instead of multiple choice questions. And as I’m always saying, we need qualitative, mixed methods, visual, participatory, story-based, art-based, methods not yet imagined, community-led research.  We desperately need it.  We need to reclaim science, and all the multiple truths and identities we all carry, without false dichotomies and dismissive binaries.

So bottom line, if you are reading research and you notice that there are mutually exclusive categories and very limited binaries, be suspicious.  Unless the research addresses one of the world’s very few truly binary questions related to whether two year old humans who can’t swim are allowed in seal tanks, be cynical.  Demand better, and less oppressive, from researchers.  And always question the narrative.

In love and solidarity,

Nechama

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s