Pages

Sunday, 13 December 2015

A Plea For Sophistication

As I pointed out several posts ago now, I have refrained from writing about feminism/gender issues on this blog. Part of the reason for this is that I have complex views on the matter and, in my experience, the internet does not deal with complexity in a mature manner. That is to say, it ignores the complexity and shoehorns whatever it reads into its preconceived notions, However, as I pointed out then, this blog doesn't have any readers so why does that matter? Well, this subject, feminism in particular, is a hot-button topic and I am convinced that a post on the matter would attract at least some attention. Yet, at the time I also pointed out that not having any readers (posting in a vacuum) isn't desirable because gender issues are just about the worse thing to discuss without a conversation... and, also, do you remember what the key thing about technical anonymity is? Time. I'll actually quote the bit where I discussed vaccum posting because it is well worth a read again:
I don't like the "in a vacuum" metaphor much... it just irritates me. However, discussing issues like gender in isolation are dangerous because it's something which teenagers tend to get as a gateway drug for politics more generally and, by implication, this means that people enter when their views are formulating. Now, look, I'm pretty sensible and, as you can see, when I write blogs I tend to try and characterise other views properly, but that doesn't mean that my view is the gospel truth and it doesn't mean that because I characterise other views as accurately as I can that they are, in fact, represented in their whole. This is huge. It's not a little irrelevance, By the nature of a blog, if someone beginning to cultivate their knowledge of feminisms, feministsms, the MRM and MRAs were to read my views on such but my views in isolation they're to, in all likelihood, be given a foundation from which they'll approach other views, but they won't get an idea of just how complex, nuanced and frequently heated and combative internet "discourse" on the matter is and I just don't think that's right. Thus, until you start posting comments, those views won't exist.
Wait, surely I am writing all this because I want to write something on a gender issue, despite there +
being no comments? Yeah, that's right. The reason why I have changed my mind is because what I am writing should point out why sophisticated or complex approaches are needed. The case study I will be using is the Gender Pay Gap (in NZ; are you really surprised?).

Now, firstly, it would help if one had an understanding of what the gender pay gap is. (You see, even if something sounds simple, with subjects like this, it is critically important to know what you are on about before you develop any conclusions.)
The gender pay gap is a high level indicator of the difference between women and men’s earnings. The gender pay gap compares the median hourly earnings of women and men in full and part-time work.
 In 2015, there was a gender pay gap of 11.8 percent.
Statistics New Zealand, in the course of explaining how that figure is determined, present the following:
The gender pay gap (or the gender wage gap) is a way to understand the differences in pay for males and females. It uses income received from jobs, rather than the total income available to males and females. Historically, the gender pay gap was used as a measure of fairness or pay equity – are males and females receiving a similar amount of money for doing the same job? 
As you can see, this is actually a fairly crude measure. Wait, don't take my word, here's Statistics New Zealand again:

We can measure a gender pay gap for either full-time or part-time workers separately. When we do this, we find that for full-time workers only, the gap is smaller but has a similar up-and-down pattern over time as the ‘total’ gender pay gap. For part-time workers, the gap reverses – women who work part time typically earn more (per hour) than men who work part time.

When we separate workers into full-time and part-time groups, we hope to remove the differences caused by the types of jobs that offer (or don’t offer) part-time hours. However, splitting workers into full-time and part-time work can change the balance of other factors that affect pay, such as age. For example, females working part time are more likely to be older than males working part-time.

Overall, we recommend using the median hourly pay across all workers, rather than using full-time or part-time workers separately.

That is saying that they would rather reduce the sophistication (by producing one statistic rather than two) because the increased sophistication of "full" and "part-time" measures would, in fact, be misleading (due to, for instance, part time males generally being younger than part time females). So, what that means is that median hourly pay is the best way to do things because, well, if we have those two statistics we've got a "lies, damned lies and statistics" situation, right? Well, not actually. After all, if you remember what Statistics New Zealand wrote a little bit earlier in that explanation...

If we want to understand the fairness of pay (do males and females get equal pay for equal work?) the hourly pay measure is the best. It allows us to compare male and female pay for a fixed amount of work (one hour). 
 In an ideal world, we would also match males and females on characteristics that influence pay, and see if there is any remaining difference. For example, we expect occupation and qualifications to affect pay. So we would compare the difference in pay for males and females within the same occupations, and holding the same qualifications.
 However, we don't do this analysis because it isn't possible to control for all factors that influence pay (and we don't measure all factors). We are also limited by our surveys' sample size.
This is an extraordinarily unsatisfactory explanation. Sure, you can't actually capture everything. However, you can, quite easily, get qualifications data, age data and occupation data (i.e. industry and position, e.g. manager). So, why wouldn't you do that? You don't get everything but you get more. Sure, maybe Statistics New Zealand is just making an unconvincing case that it is better to have something obviously flawed rather than something that looks like it could be the actual truth. I guess the real reason why they don't look at any of those things is because you rapidly end up with a lot of statistics and you need to start thinking about tables: 11.8% is nice and tidy. 11.8% is simple. Hmm, didn't Primo Levi have something to say about that?
This desire for simplification is justified, but the same does not always apply to simplification itself, which is a working hypothesis, useful as long as it is recognised as such and not mistaken for reality. 
The trouble is that the 11.8% figure is used as reality (look where Statistics New Zealand talks about "equal pay for equal work") whereas, in reality, things are quite different. For instance, you are unlikely to find anyone who agrees that a shop assistant who scans items and works the till should be paid as much as, say, a policeman or a rubbishman... both occupations (regardless of who fills them) of vital civic importance. Yet, I think you'll agree that these three occupations probably have different gender make-ups. In other words, could it be the case that the 11.8% figure reflects less equal pay for equal work and more industry and job differences? Well, we don't know the scale, but it definitely does on some level. From the Ministry for Women (which needs to be the Ministry for Gender Affairs) link:
occupational segregation  (the clustering of female and male workers in particular occupations e.g. nursing, and similarly at the industry level e.g. health care and social assistance). Female-dominated occupations tend to be lower paid than those dominated by men. Vertical segregation  is also a cause (where there are a higher proportion of men than women in senior better-paid positions). 
The use of nursing is interesting because teaching is the far more powerful example. After all, about 2% of early childhood educators are men. Let me repeat that, 2%. That's a travesty. Anyway, if you look at primary teaching in particular, I believe it is the case that most teachers (significantly more) are female yet most principals are male.* That is horizontal occupational segregation... and it arises largely because teaching is a so-called "female job" (these reasons are also complex). Now, principals don't necessarily have strictly teaching qualifications but you would expect that in an industry utterly dominated by women that there are similar numbers when you split the industry up on a hierarchical basis. There isn't. That there is that discrepancy is a big problem (which, for reference, is not solved by quotas). That is vertical occupational segregation. In other words, the 11.8% figure is, quite possibly, merely a proxy for a more disturbing phenomenon (i.e. occupational segregation)... although I only say it is more disturbing because no-one talks about it (instead they talk about how women are paid less: the non-sophisticated measure is, in fact, actively harmful). However, this is just one of several proposed explanations (generally only idiots propose one, they are all thought to be working simultaneously).

If you are wondering how this affects the pay gap, imagine the following world. We have 100 people. 50 of them work in a job being paid $10 an hour and 50 of them work in a job being paid $5 an hour. Thus, the overall median hourly pay rate is $7.50. Now, suppose that exactly half of the world are male and half female, and they are equally split across the two. That is, 25 men earn $5 and 25 earn $10. Now, the medians are still $7.50 and the pay gap is 0%. What happens, though, if 35 men earn $10 and 15 earn $5? Well, then we have a male median hourly pay rate of $10 (and a mean of $8.50, this being dragged down by the few $5 people) versus a female one of $5. Suddenly, we have a pay gap of 50%... ((10-5)/10)*100. It just so happens that in the real world that majority female jobs tend to also be ones that don't pay as well.
different patterns of participation in the paid workforce, principally because women spend a greater proportion of their time on unpaid and caring work than men. Women spend less overall time in the workforce than men (a combination of time outside the workforce and part-time work). The accessibility of well-paid part-time or otherwise flexible work arrangements is one issue: part-time work is paid less than full-time work on average. When women return to the paid workforce from career breaks, they may not be able to access flexible work at the level of role they previously held. 
This is basically talking about the impact of work experience and training. Imagine, for instance, that you are an air traffic controller in 1992 but stop that job for over a decade to look after a succession of kids and return to work in 2006 once the oldest hits 14 and can babysit the others until you get home. You are basically probably going to have to start over... maybe even having to get a new qualification. Compared to someone who started work at the same time as you and didn't take all that time off, you are at a massive disadvantage. Most people wouldn't take that much time off (I don't think) and even this person probably worked a bunch of part-time jobs during the school hours maybe, but you get the idea. You also, unless you're a complete moron, are probably aware that taking time out of a career to look after children disproportionately affects women. As a bloke that pisses me off: there's something deeply wrong with a society that assumes any given female will, at some point, be a mum but any given male will not, in fact, be a dad. Thus, you see, again, that gender expectations and roles can affect income without going to the obvious level of paying Brooke $10 more than Kim simply because Brooke's a bloke. But you see, again, that this just gets ignored when you work with simplistic measures like 11.8% which kinda imply the Brooke/Kim thing.

unconscious bias  (stereotypical views about gender that can negatively influence decisions about recruitment and career progression of women in the workforce). 

This is more promoting Brooke, without realising it, because he's a dude and Kim isn't. It is one reason for the aforementioned vertical occupational segregation. Indeed, if Brooke and Kim altered their names to masculine ones, they are probably more likely to get their CVs pushed into the "read more closely" pile, or the like, than with their feminine looking names.

Previous New Zealand research (from 2000) found that most of the gender pay gap (between 40 and 80 percent) could be ‘explained’ by differences in four variables: differences in occupation and industry of employment, differences in the amount of work experience between women and men, and women’s qualifications relative to men. The remainder was ‘unexplained’, which is commonly thought to include some level of discrimination that works against women.
However, the ‘explained’ portion of the gender pay gap can also be influenced by societal expectations of women, including that women will be the primary care-givers in families, and the appropriateness of different types of work for women and men.
Hullo... I actually quite like this link, except for the " 'explain' " bull. What the hell is that about? I don't know but it needs to be changed. I should also point out the influence of qualifications could well change over the next thirty years as the era where women were less qualified becomes less influential in these measure and the more equal and female advantage eras become more influential. That's a guess, though.

So, what have we seen? Well, I have tried to show that by using simple measures like 11.8% when dealing with topics like this, one conceals the true, and deeply problematic, issues. I have also tried to point out that because of that, no-one talks about those issues and, thus, the 11.8% figure really accomplishes nothing. I have also tried to show that it is easy enough to develop more sophisticated, but less convenient, measures which should give better explanations.

I will leave this with two excellent slides from a presentation I found online credited to Susan Doughty of Ernst and Young (I imagine she also presented this). The third slide I don't like because it shows, in action, what I was complaining about above. That is, people really do take figures like 11.8% and use it to say things which suggests that any woman in any role should expect to earn that much less. That's not true, as we shall see with my manager example below.



What this one is saying is that what we saw above can get tracked back way before people actually enter the workforce for real. In other words, the sorts of subjects that pupils take and the propensities of genders to go to certain subject groupings (e.g. there were a handful of girls in my calc class in year thirteen and similar numbers of boys in my history one... although there were substantially more people altogether in the history one... but there were three versus two calc and history classes so I don't know about how this extended to the entire school) plays an important role. If you look at the end of the presentation Doughty has written "choice isn't always what it seems". Given that this is a well made slide show I can't be sure, but I think what that will have been based on is an argument similar to following (remember, the below consists of my words, an imagined example of what may have been said).
One might very well suggest that the reason why the pay gap exists is that women don't choose to work in high paying fields like, for instance, engineering. However, the reality is that the choice to work in that field happens way down the track, probably about age 15. Why? Well, most year eleven pupils will, at that age, be choosing their subjects for year twelve... when most schools choose to split sciences up into physics, chemistry and biology. In other words, if you don't make a decision then to study physics, and then a bit later, calculus, you are probably closing the door to engineering forever. 
 The thing is that young women, in schools, are surrounded by female teachers but a lot of the male teachers that they do encounter are in maths and science classrooms. Likewise, female pupils will generally be less keen to study those options because they are aware of the gendered nature of those subjects: STEM is for boys. This perception, and perhaps their expectation that they will be a relative rarity, means that they don't have full agency in choosing to not study physics and calc... they are subject to societal pressures and influences which discourage them from the tracks that will allow them to enter fields with higher pay. Indeed, one might say that young girls see so many female teachers that they begin to think that teaching is what women should do with their careers.
I have no idea if that is what Doughty said, but if it was anything even remotely similar to that you can see how a choice may not appear to be as clear cut as you think it is.




This one has some more explanations and shows the different kinds of pay gaps. To my mind, it is the above one that is most interesting. I would like to see what sort of figures Doughty would've talked about for this slide in particular because you don't generally find people talking about anything except the 11.8% type ones (for reference, 11.8% roughly corresponds to the bottom one here... just for the entire economy, not just one organisation).

That's the 2014 figure if you are a little confused. Anyway, I promised an example of why this is not really worth the screen it appears on? Hmm, updating that metaphor is a bit of a fix. What we are going to see is going to a) look purely at managers (because I am lazy) and b) will further break things up to look at how age influences this too. Also, we are sticking with 2015 figures. Data from Statistics New Zealand (and I am pretty sure it is the same source as that they used to generate the 11.8% figure but I don't get that when I use those numbers so maybe I misread things).


MalesFemale Pay Gap Percentage
15-241718.11+6.5
25-6433.1329.08-12.2
65+2828.77+2.75
Overall30.2128.77-4.8

So, this is saying that if you are female, 15-24 years old and you are a manager you can expect to earn 6.5% more than your male equivalents. However, if you make things less sophisticated and just look at managers more generally, that becomes 4.8% less. Interestingly, older females also do better but that number could well be non statistically significant, I don't know. If you dig a bit deeper you will find that there are substantially more males at every stage, which probably explains why there is the reversal when there are also way, way more in the middle range age group. Why these gaps exist is more interesting. I personally hypothesise (i.e. make the following educated guesses) that ideas about teenage boys mean that they while they are managers they get less responsibility compared to their female peers (probably related to the types of industry they are in). With the over 65s it is probably an artefact of age... i.e. these are the women who rose to become managers despite the most disadvantages so, quite plausibly, they bring a little bit more than their male peers (they may also be scarce). The previous explanations work for the 25-64 year olds.

Basically, as this case study shows, there are ways of creating more sophisticated measurements, but this means you have more measurements... and while broad strokes are antithesis to democracy, granular analysis is antithesis to the electorate. Ah... people: what would be if we weren't human?

*That data is now quite old, from 2004, so I don't know how much it can actually say about the situation now... it was the best I could find quickly. Notice that it says 3% of female teachers are principals versus 8% of males. That's not quite what I am talking about (although it does mean that you are more likely to become a principal if you are male, for whatever reason) although it does allow us to determine the following: .11*27 + .03*73 = total number of principals. That ultimately means that we have 42.4% of principals being female, which isn't really similar to 73%, is it? That's what we were talking about. Although there are a few problems with this figure as that 73% and 27% percentages are based on all those teacher types and I was just trying to talk about primary and secondary. Again, best I could find.

No comments:

Post a Comment