Microstressors

1 A basic model

DRMacIver’s note on The costs of being understood is a good starting point model of how asymmetric costs in communication add up.

although the cost of the interaction is about the same for you and the other party, when you are interacting with many people, the cost of their behaviour on you is effectively much larger than any cost you can impose on them, because they only have to pay it once, while you have to pay it once for each of them. The cost of the interaction is effectively multiplied by the number of people you have to have it with, which for you is large and for them is you…

The problem with these sorts of asymmetric costs is that[…] they’re almost by definition invisible to those who impose them on you, because if they were there during That Conversation, even if it was initiated with someone else, they wouldn’t have to have it again.

…If you’re in a population with a 50/50 split between two groups (e.g. gender), there is no asymmetry in the obvious conversations you have—there is one man for every woman, more or less, so the number of obvious day to day conversations you have is about equal. This falls apart in a whole number of cases—e.g. in environments with large gender skews […], and in situations in which the size of your demographic is in some sense “artificially lowered” because of visibility (e.g. you’ll get to have all sorts of obvious conversations with men if you’re a woman whose tweet goes unexpectedly popular). But, by default, regardless of how bad communication tends to be between the two groups there tends not to be asymmetry of the sort that I’m talking about because the numbers are broadly equal.

In contrast, as the groups get smaller the asymmetry grows sharply. If your group is x% of the conversation, then there are \((100—x) / x\) people not in your group for every member who is. For example if you’re 20% of the population, there are four of them for every one of you, if you’re 1% of the population there are 99 of them for every one of you, etc. This means that, all else being equal, the smaller your group is, the higher the asymmetric costs you experience as a member of it interacting with people outside it are.

Ok, moving that forward slightly, let us suppose that you are visibly in a minority group in a workplace — say, the first woman in an all-male workplace, or the first Vietnamese worker in a mostly-white Australian business, or the first German working in an Indonesian bank, first person in a wheelchair in an able-bodied workplace, or something like that. Humans, being curious apes, will see you are different and will ask you about that. How many times will you need to explain yourself? Let us call the number of repeated explanations of why people like you are in fact present in this workplace, \(e\). How big is \(e\)? Suppose there are 100 people in the workplace, and your subgroup has \(n\leq 100\) people in it. If \(n<50\) then you are in a minority.

Suppose also that the more unusual you seem, the more likely people are to ask you “hey, what’s the deal? I haven’t seen any people like you here before,” so that the chance of any one person asking increases with your rarity in this group. If we use a very simple model for this, we can imagine this kind of dynamic:

\[\begin{aligned} e&=\text{number of people not in my group}\times\text{rarity of my group}\\ &=(100-n)\times\frac{(100-n)}{100} \end{aligned}\]

What does that look like?

N = 100
ns = 1:N
plot(ns, (100-ns)*(100-ns)/100,
    type = "l",
    xlim=c(1, N),
    # ylim=c(0, 1),
    xlab='Number of people in my group',
    ylab='Questions about my oddness')

# lines(ts, betap(T, 100, 10.0, 0.999),
#     , col = 4)
# legend("bottomright", c("lambda=0.1", "lambda=1", "lambda=10"),
#     lty = 1, col = 2:4)

Is this a perfect model? Dunno; for one thing we have allowed for the possibility of 50-50 “minorities” to ask one another about things; Is that reasonable?

This is just a stylized model, not a real data-driven research project, but I find this is helpful to cultivate empathy for things that are stressful just because of frequency. Anyway, it is a good for making something clear — if you are in a majority group, people in minority groups around you might be having an awkward question asked of them at a hundred times the rate (for example) that you get asked that awkward question. If you are the trailblazing visible minority member you might need to explain yourself 98 times in a 100-person firm. At 5 minutes per conversation, that is more than 8 hours of explaining your ethnic background. If you are from a majority group you usually don’t need to think about having that conversation on repeat every time you change workplace.

I have myself experienced this problem of being asked why I was there by virtue being a visible ethnic minority in a project. Indeed, explaining myself was tedious after a while, even though everyone who asks it is friendly and well meaning and curious; just saying the same thing too many times makes it a chore. At the same time, it was not a particularly major problem.

That experience, admittedly, was as a high-status outsider, which might be different to the experience of a low-status outsider from a vulnerable or oppressed group. If those conversations were more regularly hostile or even ambivalent it would be worse.

Listening to people talk about this, the kind of conversations that minority group members tend to experience as stressors are broader than explaining their own background. It is more about ambiguity and awkwardness and signals that can be misunderstood cross-culturally.

TBC: enthalpy model of stressors.

cf DRMacIver’s Notebook: The costs of being understood, and reasoning from lived experience.

2 How bad are microstressors for people in majority groups?

i.e. What is the baseline?

Important, since the relative badness of microstressors is part of the conversation about them, i.e. the inequity argument also arises from the fact that, e.g. white males might find some stressors less stressful, or might be more resilient in the face of stress etc. TBD.

3 Splaining

Another interesting and different case, is splaining. This is one that people who come from an identifiable group of low status often report adverse reactions to, and roughly corresponds to having spuriously confident people explain to you your own area of expertise despite your qualification in the area.

This is the point where we mention Rebecca Solnit’s classic essay which, via the internet meme machine, coined mansplaining about the real and often irritating statistical tendency for men to be overwilling to talk over people, in particular with spurious confidence in their expertise. Later as a racially-charged version we acquired whitesplaining and then, generically, splaining.

I work in a culture where people will casually explain stuff to me until I ask them to stop, rather than wasting time gingerly testing to waters to see what level of explanation is perfect. Some of my colleagues explain stuff to me all the time. I find this to be a mixture of irritating and charming, depending on circumstance. It is not because I look female, or brown. I look like the white cis male that I am. I get splained to a lot. This small datum leads me to wonder about the distribution of splaining.

It seems like part of the cultural phenomenon of splaining is that the splainers are usually male and/or autism spectrum people. Do splainers splain more readily to people from other groups than their own? Or is it that they splain to everyone equally but splainees from particular subgroups are sensitive to it? I imagine both effects could be in play.

If you are in a minority sensitive to having your expertise dismissed, a culture of speculative mutual explaining, and a culture of talking down and ignoring might look similar, and any given splanation might be due to either possibility.

This is an example of a thing which is possibly both unevenly distributed and unevenly interpreted. What does that mean in practice?

TBC

4 Micro-inequity

I am late to the party on this one. It turns out there are other terms for microstressors, which I had missed because of all the attention-sucking arguments online about microaggressions. An important one with a long publication history is Micro-inequity, intended by author Mary Rowe (Rowe 1977) to describe the wearing-down effect of many small stressors and in particular how they fall inequitably upon people in under-represented groups. That terminology did not go viral, unlike the more contentious “micro-aggressions” terminology of Chester Pierce, as resurrected by Derald Sue and colleagues (Sue et al. 2007), and which spread the polarised social media of 2016, possibly because it was contentious. I would be curious to see a philological study of this history. Microinequity does seem to be helpfully less loaded.

5 As microaggressions

tl;dr Casual bigotry is real, but the term microaggressions is IMO an unfortunate and possibly counter-productive term to address it.

Microaggression seems to be a popular term for what I would call a diversity microstressor, or what is called a microinequity elsewhere. It is a stressor that disproportionately experienced by people with minority status. AFAICS calling something a micro-aggression conflates several things into one concept. In particular, the term seems to be contentious because in the popular imagination it seems to many to imply intent, that these stressors are a deliberate attempt by one person to irritate another, possibly deniably. Or at the very least that they should be treated as such. Clearly, people can and are intentionally bigoted in subtle ways sometimes. Whether to use this very loaded terminology to which seems to presume bigotry per default is another question.

I am generally not keen to argue about details of terminology but this one does seem to me to be worth trying to shift. Have you ever seen anyone accused of a microaggression respond productively? I have not. So far I have only seen this term lead to argument about intent, which seems to give the term a toxoplasmic force, with flavours of a tokenism/tablestakes logjam.

Side question: I understand that this might be culturally specific, in the sense that attaching high importance to intention is a particularly Western value. Perhaps the distinction between intent and outcome is less important outside my world?

The criticism of this term that everyone likes to cite, Lilienfeld (2017) (crib notes) argues that there is not much empirical basis for any consistent natural category of action being microaggressive. There are a few arguments there, but the most basic one to me is that it conflates two phenomena into one thing. Calling pasta micropoison is a helpful category for coeliacs but possibly people with gluten tolerance might prefer the bad effects of pasta on coeliacs and the nourishing effects of pasta upon other humans, be separated rather than demanding that pasta be classified universally as a micropoison.

Likewise, microaggressions seem to conflate the dynamics of small stressors and the interpretation and meaning of the stressors for the person who experiences them.

Things can be frequent and asymmetric and malign, or not, and can vary on each of these axes independently. But maybe the category is useful nonetheless because it provides a way of analysing the problem, e.g. because the presumption of guilt is a helpful heuristic in handling the problem, i.e. it is better to start from a presumption of aggression. Lilienfeld (2017) argue “no” to that too. Rather, they argue that

1. The class of thing that are microaggressions is muddy
2. We lack compelling evidence that labelling and treating bad interactions as microaggressions is a helpful intervention

Nagai (2017) makes the even stronger claim that microaggression research is pseudo-science.

Monnica T Williams by contrast defends the microaggression concept on the basis that these problems can be solved at a psychometric level (i.e. ‘we can in fact measure these things’). She does not argue that the implied category is clear as such (Williams 2020):

Because microaggressions are subtle and must be interpreted in context, a person can believe they experienced a microaggression but be mistaken. That being said, when Person A describes a microaggression, and Person B approaches this situation with the notion that Person A “could be wrong,” it sets the stage for more microaggressions and relationship damage (or further relationship damage if the offender is the one claiming the target may be wrong). The default response should be belief of a person’s experience, just as we would believe them if they said that we had accidentally slammed the door in their face, left something in the hall that they later tripped over, or mispronounced their last name

It seems these researchers could all be correct in their main claims; the stances that they publicly defend have surprisingly small contradictory overlap. People with stances that we might characterise as “microaggression-skeptical” usually give examples of behaviours being classed as microaggressive which are extremely specific and essentially impossible for the purported aggressor to know ahead of time, which they argue is in itself an aggressive thing to do. People more “microaggression-supportive” give examples of casual, small-scale bigotry which would clearly be burdensome to experience, and sometimes provide cover for intentional bigotry. Both these seem to point to microstressors being real, but have, it seems, different models of the typical contexts in which they arise, and thus how to address them.

Williams’ door slam example draws that out nicely: If I slammed a door in someone’s face, either

1. I did it accidentally, in which case it was not aggressive but it was unfortunate but I need to learn to be more careful with doors, or
2. I did it deliberately, and I should be reported to HR.

Most employers have a well-understood workplace safety procedure for determining this and resolving it. Navigating the mysterious intermediate category of accidental-but-real-aggression seems to be where the disagreements begin. Conor Friedersdorf argues that this is the core of the misconception:

Many acknowledge that there are minor slights that do cumulative harm, but believe there are better ways to address them. The scores of emails I’ve received in response to the article include people on both sides of the larger debate on whether “microaggressions” are a sound or unsound framework. Its defenders often fail to realize how many of its critics share their desired ends, if not their preferred means.

Amitai Etzioni, Don’t Sweat the Microaggressions argues that microaggression fail to minimise the usual frictions of subcultures.

When you cannot tell if you are aggressive before the other person responds, and anybody can declare he or she has been abused by anything we say, communion between members of different groups becomes even more difficult. What we need is more contact and fewer grounds for mutual accusations and sense of being victimized.

Sure, OK. But this conversation seems to be deep into the weeds, which is in itself evidence for me that this term fails to cleave reality at the joints.

6 Sensitisation

Musa al-Gharbi asks Who gets to define what is racist? and proposes some ideas. There is an interesting wedge in there about the preferences of progressive activists supporting multiculturalism and the People-of-Colour for whom they advocate, who may not be especially progressive, and how the dynamics of this representation plays out.

For instance, a recent metanalysis on microaggressions found little empirical substantiation for the harm claims advanced in the literature. Yet there is abundant research demonstrating harm caused by heightened perceptions of racism (Anderson 2013), discrimination (Pascoe and Richman 2009), racialized violence, and racial inequality (Shedd 2015). There are very well-established and highly-adverse impacts on the psychological (Pieterse et al. 2012) (and even physical) well-being of people of color when they perceive more racism, racial inequality, and discrimination. That is, we have not (yet) been able to empirically verify that microaggressions are typically harmful, nor have we been able to effectively measure the extent of that harm. (Again, the vast majority of Blacks and Hispanics find even the paradigm cases to be inoffensive). However, we have ample reason to believe that sensitizing people to better perceive and to take greater offense at these “slights” actually would cause harm

I suspect there is a generational angle too, but I have not run across any good studies of it yet. TBD. I think it is likely there are studies of this, I just haven’t run into them yet.

This argument is both interesting and dangerous. We, or at least I, want to be very careful of arguments that make it likely to blame people for bringing an inequity for light. On the other hand, we know intuitively that validating and magnifying every potential slight is considered poor therapeutic practice and does not help cultivating resilience.

I do not have a great prescription here; I think it is a hard problem to devise hypothetical social mechanism that manages stressors for all the various stripes of people in the modern world and their different needs and desires with overall minimal stress. We want to be alive to the potential trade-offs here, between ignoring too much and silencing suffering, and validating too much, hence manufacturing suffering. If a plan does not include recognition of both these extremes, it is not credible. As to how to honestly trade off between subjective, frequently weaponised interests, I do not know in general.

7 Incoming

• ’Where are you from?’

So what I am going to advise my Chuvash, Chickasaw Freedman, Scotch-Irish, Austrian, French Canadian daughter? She’s into KPop, and talking about learning Korean—you know she’d get the question there. I’ll tell her that the question is natural and human, and to be ready for it with a set of funny answers (her grandparents already thought up the label ‘Chickachu’). She should be ready for unfriendly intentions, but to assume the best. And, like everyone today, to walk on eggshells.

Coleman Hughes on Colorblindness discusses a classic example:

Mounk: How do you think we should deal with the problems that arise even among broadly well-intentioned people? I believe you’re 12 or 13 when you move from a public school that is very diverse to a private school that is somewhat less diverse, and in which people treat you, as you describe it, very positively and nicely. But a bunch of people do things like touch your hair because you have a huge Afro at the time. Not that any one of these people was particularly obnoxious, they didn’t tease you or bully you for it, but it did sort of take on this psychic harm. It was a burden, right? It was annoying. You’re pissed off with it, very understandably. And I think it’s sort of interesting to reflect on how our culture teaches us to deal with these things.[…]

What’s the right attitude we should teach people towards those kinds of experiences where we take this seriously, where we don’t completely trivialize it, but we also don’t encourage people to lean into an identity of victimhood on the basis of those kinds of things when they’re not ill-intentioned?

Hughes: Just to put more color on it, the experience you’re talking about in my case was that I went from a school where having an Afro was normal (maybe 30% black, all the white kids there saw kids with Afros all the time) to a school where now it was notable, because the white and Asian kids just didn’t know anyone with an Afro. So they would just touch it all the time. And the intention was totally benign, but it was just so fucking annoying and it built up and up, and I felt like I would have to be in the position of a dick to tell them to stop. So it just became this problem for me. And the POC conference taught me to view this as a microaggression and, actually, as on the continuum of racism with slavery and lynching on one end and this kind of thing on the other. Framing it that way is very powerful and it imputes an ill motive on the people doing it which just, in fact, wasn’t there. You asked what is the right approach to take to these kinds of situations. Well, the wrong approach is to impugn the motives of people doing things that actually have benign motives.

The wrong approach is to ignore intentions because intentions matter. On the other hand, how I dealt with it by simply just ignoring it until it built up and I got upset enough to cry to my parents about it—well, that’s not really good either, right? I think a wise adult in that situation, or a wiser child, would have said, “I’ve got to set boundaries with these kids. I’ve got to tell them, ‘I know you want to touch my hair, but it’s not cool to do that. It makes me feel like shit. It ruins the preparation I’ve done. And I know you don’t mean anything by it, but I’m asking you to stop.’” And if they persist after that point, well, then you get into a situation where maybe it’s bullying. But the way to walk the line is to respect the concept of intention.

8 References