The data points on pay gap by gender and ethnicity are ones I’ve not seen before and it’s great that this dimension is visualised.
A surprising finding here - that doesn’t match what I would have expected - is how Engineering is compensated 90% in cash. I would have assumed equity is pretty heavy for software engineers, especially in senior and above positions.
I do wish there were more takeaways to what seems to be data confirming these:
1. Compensation is going up, and especially software engineering compensation.
2. Startups are starting to compete stronger on cash compensation (thanks to a strong funding environment, and the market).
3. Remote work is slowly eroding regional differences in the US (and, as a note, globally, as well).
Sadly, there’s no data on remote work here, nor is there anything on what % the market moved up the last year in various disciplines. At least for engineering, the past year has been a major jump upwards in compensation.
> A surprising finding here - that doesn’t match what I would have expected - is how Engineering is compensated 90% in cash. I would have assumed equity is pretty heavy for software engineers, especially in senior and above positions.
FAANG and similar jobs are equity heavy, but standard engineering employment is often more cash-heavy.
Working outside of FAANG, my base comp was significantly higher than FAANG base comp, but of course the total comp at FAANG would edge out my total non-FAANG comp.
An average from a bimodal distribution occludes more than it reveals. These days we all have a bit of a fetish for what appears to be data driven analysis even when it involves something like this where it makes things actively worse.
The hiding of the data’s distribution always makes me sad.
It is not costly to provide quintile or decile data, yet no one ever does when that is where the meat and pickles are. I assume most of the time, averages (and especially unspecified averages where you do not know if it is mean or median), are used to invoke emotions and land more clicks.
In most of these types of salary surveys I've seen they report a single average salary for a position. I wish they would break that down by years of experience -- probably have to pay for the "full report" to get that level of detail.
This is from Europe, but I think it's true in the US as well. Software engineers at large tech companies and venture funded startups get paid a lot, including equity. Software engineers at consultancies or doing internal development at businesses with several thousand employees are on a different salary scale.
> our gender gap data shows that companies are still struggling to offer their female and BIPOC workers equal pay for equal work
Does this data account for levels or years of experience? For example, if there are more junior-level women in the industry than senior-and-above level women, then of course you'd expect women to earn less, percentage-wise.
If it's not accounted for, then the quoted statement seems false.
Similarly, how come asian workers are left out in the following statement:
> With a broad stroke, male and White/Caucasian workers simply earn more than their counterparts.
It seems asian workers earn even more so it's weird how only white/caucasian is called out.
To be fair to the parent comment, you need to click “By Ethnicity” to see that Asians and White earn similar pay comment (at least on the mobile version), it does not appear without a user interaction. However the general conclusion that Caucasians earn more is viewable by default, so someone who skims the article is going to come to a very different conclusion than someone who read it carefully. And let’s be honest here, most people are probably just going to skim it (and even more honestly, I personally skimmed it and didn’t even realize it was addressed until I read this comment chain).
In the same spirit, I suppose one could say that it's weird that, in your first post, you didn't mention that it does ultimately say that on the website.
>I could be misunderstanding the point you are making.
I don't disagree with the broader point that you're making, I'm just pointing out that you're doing what you're critiquing as you critique it. Your first post says that they don't actually say what the data shows, and you go on to suggest that an improper presentation of data can cause people take away the wrong idea.
I'm saying that you choosing not to mention the fact that they do state that fact accurately merely a few sentences later also could cause people to take away the wrong idea.
I don't understand the point of this conversation.
> Your first post says that they don't actually say what the data shows
And I maintain that the summary omits and misrepresents what the data shows.
> I'm saying that you choosing not to mention the fact that they do state that fact accurately merely a few sentences later
It's behind a click. It actually doesn't show at all without a click. I missed it the first time I skimmed the article (thank you for pointing it out) and I'm sure other people will as well.
There are also pretty significant median age differences between ethnic populations and whites in the US (15 years IIRC). Unfortunately, you never see that controlled for.
I saw that section. It's weird to me to use tenure here instead of industry experience or level since tenure specifically means how long you've been at the same job.
In my experience in tech, staying at the same company (long-tenured) is the best way to ensure you are underpaid. They seem to allude to that with:
> The “Great Resignation” may present an opportunity for long-tenured, underpaid workers to find competitive pay elsewhere.
If my understanding is correct then that doesn't account for what I'm calling experience or level.
The average salary distribution is surprising to me. I find myself right in the middle of the distribution (~50th percentile), and yet I know for a fact that I have the highest salary out of everyone in my social circle, which includes several competent workers in the same field. I also believe other data suggests that my household income puts me at least in the top 10% of earners. How can these both be true at the same time? Do these data really represent the state of the general labor market?
This post shows an average in every region more than double that. It's not a representative sample. The charts are pretty but I don't think this analysis has any value.
says "People 15 years old and over beginning with March 1980, and people 14 years old and over as of March of the following year for previous years." seems it would be reduced by full-time students, retired people earning Social Security, etc.
> This post shows an average in every region more than double that.
says "this report reveals some of the rapidly changing dynamics in how companies of all sizes are compensating their employees." seems it only includes employees and thus would not include students, retired, etc.
If this is meant to be a representative sample of the US, the data doesn't pass the sniff test. They don't provide enough context to understand how else we should understand it.
> Insights from this report represent data collected from Sequoia’s Employee Experience Surveys (2017 – 2021), anonymized information and trends from our database, and Dataforest surveys. Dataforest is refreshed periodically with updates from new survey submissions.
The data here appear to come from salary surveys of working individuals. The US Census Bureau data appear to be reports of income from all sources for all persons, which appears to include children 14 and over, full-time students, disabled persons, retired persons, and other persons unlikely to be reporting a salary to Sequoia.
the Fed counts all people over the age of 14. using the same population, the median income is ~$36k [1] so over half of the population makes less than 2/3 of the mean personal income.
also, this report only covers a listed number of "departments" and leaves off many lower wage occupations. retail, child care, most unskilled labor, etc dont appear to be included in the Strategy, Sales, Design, Engineering, Leadership, etc.
They, in poor form, report the "average" without explicitly specifying if they are referring to the mean or the median. Your perception would be shared by the majority of people if it's the mean they are presenting (which is most likely the case) because of the extreme income distribution. Overall, I am not a fan of the data presentation here.
Indeed. Yet, this graph shows the distribution of averages. See the x-axis. I didn't want to be so blunt, but due to its shoddy presentation, I cannot trust what this page is showing.
(As mentioned) median, not mean, is the relevant statistic for income. The income distribution is exponential, so mean is irrelevant for individuals. One individual could account for 100% of the income, and the mean would be the same.
My female colleague used to hypothesize why men earned more in general.
She said that, if we take an average male and female that have no skills whatsoever, male can still be a coal miner or do a construction job or be a mover better.
This means male inherently has higher earning potential. It makes sense that male earns higher.
As a side note, I appreciate she talked about this at work openly. As a male, I am scared to discuss such a thing at work.
This doesn't make sense. A skilless women can always be a prostitute. And they earn a lot better than construction workers who aren't exactly skillless in the first place given they are capable of doing physical labour. Men are paid for their ability to work, while women's bodies have intrinsic value just for being a female. I know it's not the most politically correct example, but skillless women are clearly valued more than their counterpart men, reason why you don't see as many homeless women as men.
Not sure where this is coming from, or if it relates to the article at all really, but do you think it squares well with the section titled "Does the gender pay gap decrease among long tenured employees?"
I agree that it's often been awkward to talk about my compensation at work, though I continue to try if only to make such disparities more clear.
This would explain a difference in average or median pay across all jobs. But it doesn't explain the difference in industries and jobs unrelated to physical labor.
Those things show how prices/wages could be linked across industries through opportunity cost of the laborer. It doesn't suffice to explain how a sex pay gap could emerge. If hiring men were more expensive (because they have greater opportunity costs), then just don't hire them--hire the women, or whomever is willing to accept the wage you're willing pay for the particular labor services desired.
You wouldn't expect a supermarket to pay male cashiers more because the men could become miners.
But they might have to pay more for the people who haul heavy boxes around in the back, which might be male-dominated, leading to a pay gap for 'supermarket worker' if not for 'supermarket cashier'.
>It doesn't suffice to explain how a sex pay gap could emerge. If hiring men were more expensive (because they have greater opportunity costs), then just don't hire them--hire the women, or whomever is willing to accept the wage you're willing pay for the particular labor services desired.
Yes exactly. It would only explain a pay gap at the low end of the pay scale. It wouldn't explain it in the middle or high end.
It's an interesting speculation but the main reasons are child birth (women out of work for longer) and personality differences (agreeableness is known to negatively predict income irrespective of gender) explaining the within-occupation delta and different career interests (more nurses less engineers) explaining the broader delta.
Unfortunately social differences cannot be controled for differences in socialization. People instead reach for essentialism as an explanatory model. Essentialism is a rationalization not an explanation that better falls under the umbrella of reifications than as real and true knowledge.
When I graduated with my degree in electronics, I found the software industry paid much more than the electronics industry, so I entered the software industry.
If the electronics industry wants to hire me, they would have to pay higher salaries. So the salaries in one industry (software) can drive up salaries in another (electronics)
Admittedly, nobody's leaving programming to become a miner - but plenty of jobs pay less than mining, so it could explain a pay gap in some parts of the salary range. Especially in countries like Australia with big resource extraction industries.
Yes, a salary raise in industry A will lead to a salary raise in industry B. Industry B has no option to stop hiring people altogether, so they have to hire at the higher wage level.
But a gender gap in industry A will not lead to an gender gap in industry B. Industry B will just hire the cheaper gender until industry B has no gender gap.
Yes, it improves salaries. But it will improve salaries for both men and women at non-physically demanding jobs.
At non-physically demanding jobs, men and women are competing with each other fairly, so if the salaries demanded by one group go up, the employers will hire the other group until the salaries are equal.
Definitely interesting to look at the Pay Gap charts. But it might be interesting to see what the pay gap is after looking at hourly wage.
I often hear this argument that men work longer hours. Given how much data they have presented, will definitely be worth looking at that dimension too.
Edit: In S/W that might not matter at all, but some business functions hours might matter. More data the merrier.
Kindest interpretation is that the 'and' there is a restrictive and subsets to the male and White.
Alternative is that Asians are a minority in the US, being 6% of Americans, while White people are 60%, and so it is more interesting for talking about this larger group.
I was surprised by gender pay gap that follows tenure. Very interesting. Would have expected that total pay gap exist since proportion of women in junior engineers is higher than in senior engineers.
Interesting stuff for future perhaps: foreign vs. domestic; and breaking out Chinese + Indian at least for Asian.
It’s unclear what tenure means. Usually it means time at a company not time in industry. But performance ratings also matter and the type of company matters. It would be best to see this normalized by years in industry, company size (revenue or market cap), and ideally with performance ratings accounted for in some way.
For a company in the west, it would be nearly suicide, to discriminate women. So based on this assumption, I think there are other reasons for the 'pay gap'. Actually to really measure it, you should create a sample of companies and go from company to company analyzing any 'pay gaps' including all the factors, like reduced work time. Would not wonder if the graph flips after analyzing it this way.
It would be great if that conversation around diversify started including original social class diversity as well: if your tech companies are hiring more diverse minority candidates, but they all had lawyer/upper class parents, did you fix anything? Probably not.
The data is apparently from their own employee surveys and includes "a minimum participation standard of two hundred companies but can often grow to include thousands of companies."
Hard to know how much you can trust the results with a small and biased sample like this.
One good thing about data from Sequoia is that they handle payroll for a lot of companies. They know exactly what companies are paying each of their employees, so it's a really good data set from that perspective. It's not self-reported by employees, for example.
Sequoia designers, this is beautiful but it would be great not to rewrite my scroll functionality. Why in the world would I want implied momentum in my scrollwheel? I already know how to use my mouse and am happy with it.
"Please don't complain about tangential annoyances—things like article or website formats, name collisions, or back-button breakage. They're too common to be interesting."
Scroll based UI stuff is one of the most frustrating thing of the web. Leave my scrolling mechanism alone! Web designers should realize the difference between landing page design vs article design. People clicking on your article link means that, they want to mostly skim the article and see the graphs. Why would you want to bother them with stupid design choices?
Wayyyy back in the day, people thought they could make smooth scrolling "better," e.g. nicer easing, more responsive, etc. There is, however, no actual benefit to this, no.
In this case it's crazy choppy for me. My firefox updates my scroll position on this page at like 10FPS on a recent ultrabook. So definitely not a success story for "better smooth scrolling".
