A few weeks ago the release of the UPA 2005 Salary Survey triggered a cross-blog discussion about salary differences between men and women in the user experience field.

Summing up: I was the author of the report (producing it is part of my duties as UPA VP and Director of Members and Sponsors). My writeup documented a salary differential between male and female survey respondents: median salary for males was 80K; for women, 72K. (Average salaries were 82.8K for men and 74.3K for women.)

In a rare instance of Jakob Nielsen lagging the discussion and not the other way around, Jakob posted some salary info of his own here.

In a comment posted to Webword I pledged to follow up on this issue with some additional analysis. I intentionally made that public commitment. I figured I’d be more likely to actually follow up if I said it in a public forum.

Thankfully, my UPA Board buddy Lyle Kantrovich helped out with these analyses. So give him props (and visit his blog).

So here’s what Lyle and I found:

The gender gap in our field narrowed slightly from 2000 to 2005. Lyle had access to the 2000 survey data. When he looked at the averages, he found they were 80.0K for men and 70.8K for women, and the gender differential was 9.2K. In 2005, the gender differential was 8.6K. Not a huge difference, but don’t discount it.

The gender differential in our field is less than those in other fields. Lyle converted the differential to a percentage. Turns out that women in the UCD field earn 90% of what the men earn. (In 2000, the percentage was 88%.) Well, according to the US Bureau of Labor Statistics, in 1999 women in the US workforce as a whole earned 77% of what men earned.

Interestingly, that paragon of research accuracy USA Today (spurious infographics, anyone?) also reported on a US Census Bureau report in 2005 showing that the pay differential had backslid to 75.5% by 2003.

Finally, Lyle dug into the gender differential for similar professional occupations (data courtesy of the US Census Bureau). The 2005 gender differential for men and women in management, professional, and related occupations was 73%. In computer and mathematical occupations, the differential was 86%.

Not to be outdone, I decided to see if my SPSS Fu was still strong. (In graduate school I used SPSS so much I occasionally dreamed in SPSS syntax language.) Stick with me here as I describe my thinking and how I tested it:

In the original report we demonstrated that there was a difference in salary between men and women. However, the data also show an effect for experience: as you would expect, the more experienced the practitioner, the higher on average their salary.

Having come from a family with a lot of working women, I know that women often leave the workforce to make and raise babies. (And they often return to the workforce when the little yard monsters are old enough to go to school.) So I became curious: was there a difference in experience between men and women in our sample? If there was, it would suggest that gender was not (or was not the sole factor) driving the gender differential.

It turns out that we had captured experience data in ranges, not actual values. That is, the survey asked people how long they had worked in the field, and offered them these choices: 0-1 years, 2-4 years, 5-7 years, 8-10 years, etc. I decided that before I went to the trouble of recoding these categories into something that could give me an average, I would first look at the distributions for each gender group. If there was a big hump toward the left side for women, that would suggest that women had fewer years of experience.

Because the number of men and women differed, I converted the number of respondents in each experience category into a percentage of the total number for each gender group. Then I graphed them.

As it turned out, the distributions didn’t look all that different. There are relatively more women than men in the mid-career categories (8-10 years, 11-15 years, and 16-20 years), but more men on the high end (21+ years). Certainly nothing to suggest that experience differences were driving the gender differential.

Next, I hypothesized that more women than men worked less than 40 hours per week. My wife, also a usability professional (and way smarter than me), has availed herself of the 32-hour work week offered by some companies. I figured other women in our sample might’ve done the same.

Alas, no luck. The data only indicated whether people were full-time employees with a company, or were solo practitioners. I had no continuous measure of hours worked per week.

So where does that leave us? The data suggest that yes, Virginia, there is a gender differential in our field. The silver lining (if it can be called that) is that the disparity seems to be less than the average (as measured in the US by the Census Bureau and the BLS), and less than the disparity in other professional occupations.

Any questions about this “bloganalysis?” Give me a shout.

This entry was posted on Friday, May 19th, 2006 at 3:38 pm and is filed under Professional, Resources. You can follow any responses to this entry through the RSS 2.0 feed. You can skip to the end and leave a response. Pinging is currently not allowed.