Right, but uh... The vast majority of positions Glassdoor cares about are salaried positions which don't track hours. They're exempt from overtime, as well.
It also seems to me that that signal would show up weakly in the data because it wouldn't vary the immediate effect on apples-to-apples salary, only on promotions and possibly referrals.
So I think your complaint is specious and the inclusion of a totally unrelated data point from a radically different dataset is either very misguided or profoundly disingenuous.
> The vast majority of positions Glassdoor cares about are salaried positions which don't track hours
Yes. And this makes any kind of data analysis a lot harder.
On paper I work the same number of hours as my Japanese coworkers. In reality they work more than double.
It's the same way on-call for many companies is just rolled into salary.
I've managed projects as a senior dev while the senior dev sitting next to me was purely technical. Etc etc.
The truth is two senior devs sitting right next to each other can have wildly different actual job duties.
I suspect glassdoor has taken at least some of those differences in work duties vs stated title into account.
> because it wouldn't vary the immediate effect on apples-to-apples salary,
Why wouldn't it?
Its not uncommon to stay at a certain level/title for 3-5 years.
You can't just ignore hours worked because "salary".
I come from a strong union background where you would get overtime or be able to take the time off later and where there were fairly rigid job duties. Where interviews and scoring candidates is all done in a very standardised fashion.
And oddly enough after adjusting for these explicit factors we had a very small unexplained gender pay gap. Much smaller than average for the tech industry.
Perhaps it's in our interest to make these factors explicit rather than implicit?
Because hours worked doesn't affect salary. What's more, you keep assuming that there aren't counter-examples in this dataset at all. It's just a given to you: "I read on Forbes that across all Americans women work slightly less per week" and brought that to this conversation.
But in this industry long hours are the norm, single people without family duties are the majority, and many compensation models very much incentivize long hours.
> You can't just ignore hours worked because "salary".
Unless we're talking consulting, that's exactly what Salary lets you do.
> Much smaller than average for the tech industry.
The average pay bracket difference is indeed smaller. However, it's still large enough for concern, and the other issues I listed around work culture are problems that disadvantage women more than men in modern society.
> Perhaps it's in our interest to make these factors explicit rather than implicit?
I am always in favor of transparency and more analysis but I will not let ANY interpretation of the specific data we have on hand be constantly forced off on the grounds that we could collect better data.
Saying, "Women get paid less, let's make sure that's not happening" is a call to action, not an accusation leveled at anyone specifically. If women in your organization aren't being paid less, congrats. Job done.
Let's be clear: You're effectively trying to stifle and derail the entire conversation on the grounds that it might not be happening everywhere, or that statistics might not be what popular media claims they are. That's counterproductive and frankly unwelcome. If that's your take, your take stifles the discussion here.
We can simultaneous recognize that tech is a more complicated compensation environment and also that women are reporting under-compensation. We're all adults capable of multi-tasking.
> My point is: Work longer -> slightly more productive -> slightly higher pay.
And I'm telling you this is not a foregone conclusion.
> If the conversation starts from an inconclusive or flimsy premise then it should be derailed.
No. It should include both the initial position and the specific cases that inspired it and then move to a broader case. This is not customer service, these are people's lives.
> It is far too easy to make bad decisions based on flawed studies.
Are you suggesting that the decision to pay women equally, carefully audit corporate promotions, and firmly and directly punish racist and sexist harassment are actually open for debate? Is there an outcome where we might say, "Oh no, actually it's correct to pay women less for equal work?"
> The first step is to work out if it's a real issue we could be, or should be, solving.
How many individual women need to risk their careers explaining unconscionable behavior by their managers and employers HR departments before you're satisfied anyone anywhere is allowed to have this conversation? You should put that number out there.
> But you have to actually make a case.
You may not realize you're doing this, but you're actually interfering with everyone's ability to make a case by derailing every conversation and making it all about you and your (higher than any other sociological or scientific field) standards for allowing discussion.
Look at you. You're here because marketing copy for an event triggered you. You're so angry you're willing to argue that maybe pay shouldn't be equal after all.
I was talking about the specifics of a single study on comparing the tech industry with other industries on the wage gap (and I guess more broadly at the fact that people misuse these studies).
You are broadening the conversation to look at sexism in broader society.
Let's stick to the specifics here.
> making it all about you and your (higher than any other sociological or scientific field) standards for allowing discussion.
Nonsense. Most rigorous wage gap studies control for a lot more than the GlassDoor study.
The particular data (working hours) that I'm saying should have been included is straightforward to measure and collect.
> You're so angry you're willing to argue that maybe pay shouldn't be equal after all.
A swing and a miss. It's like you aren't even trying to read what I'm writing.
> You are broadening the conversation to look at sexism in broader society.
No. No. No sheepmullet. You've been broadening the scope from unrelated articles to Japanese work culture whenever it suits you. It's too late for you to go, "HEY I WAS RIGHT IN THIS LIMITED CONTEXT."
Unbelievable.
> Nonsense. Most rigorous wage gap studies control for a lot more than the GlassDoor study.
Yes, because the include many more hourly, tipped and commission-based jobs. The Glassdoor dataset has almost none of those.
> The particular data (working hours) that I'm saying should have been included is straightforward to measure and collect.
Salaried jobs don't collect the data, and that's the majority of Glassdoor's data.
> A swing and a miss. It's like you aren't even trying to read what I'm writing.
Given your prior attestation you are not broadening scope despite mentioning Japan's dangerously broken work culture as a pole star, I'm wondering if YOU are reading what you're writing.
And given the moderation, you're not doing such a swell job convincing people here.
But by not controlling for it you reduce the studies explanative powers.
You don't get to just ignore it because it is convenient.
It doesn't even pass the "more likely than not" standard of evidence.