Analysis of Gender Bias in 2008 Salary Survey
By Scott Sehlhorst
[Originally posted at Tyner Blain blog]
A related analysis showed the distribution of total compensation. This article takes a more detailed look at gender differences in compensation.
Salary Distribution by Product Manager Sex
When splitting the data into compensation for women and men, we see the following:
The graph shows two things - more men responded than women, and the men that responded to the survey reported higher total compensation than the women. We should not (yet) jump to the conclusion that men statistically tend to earn more than women in product management roles. It may be true, but we have to acknowledge that other factors might influence the data.
Product Manager Compensation and Age
One trend that “feels true” is that older people earn more than younger people. As long as people enter the workforce at roughly the same time, and as long as raises are larger than increases in starting salary, this will hold true. The following chart shows the reported ages of the product managers that responded to the survey:
This distribution shows that there is a wide range of ages among the population of product managers that reported total compensation data for 2008. Perhaps a higher percentage of the older respondents are male (and with higher salaries that correlate to their ages), explaining the disparity from the first graph.
The next diagram shows product manager compensation versus age. The diagram also shows the sex of each respondent - red “O” symbols for women, and blue “X” symbols for men. We’ve also applied linear regressions to the data - showing the overall trends in compensation versus age. The blue trend lines are for male respondents, the red trend lines are for female respondents.
The trend lines that are drawn show the linear regression as well as the +/- 95% lines. You can see validation that total compensation does tend to climb with the age of the product managers. The trend lines show that compensation for female product managers is lower than that of male product managers, as a function of the product manager’s age.
Since the trends are so visibly different, we can conclude that even if more older males responded to the survey, it does not really matter - the trend lines show a disparity across the reported age range.
Product Manager Compensation and Experience
While a trend of “the older you are, the more you make” feels accurate (and the data also suggests it), a trend of “the more experience you have, the more you make” also seems logical. When all other things are equal, a more experienced product manager is likely to be more effective than a less experienced product manager. What does the data show?
An older product manager is also likely to have more experience. The following diagram shows age versus experience as reported by product managers in Pragmatic Marketing’s 2008 survey.
The blue X symbols represent male responses, and the red O symbols depict female responses. There appear to be more responses from experienced product managers than inexperienced product managers - both for men and women. The following diagram shows the frequency of responses by reported experience.
There are definitely more responses from more experienced product managers. And there are disproportionately more male responses among the more experienced product managers. If compensation correlates strongly to experience, that could explain the distribution variance within the overall population.
Product Manager Compensation Versus Experience
The previous two graphs give us insight into the ages and experience levels of both male and female responses. The following diagram shows the reported total product manager compensation by years of experience.
There is a visible trend that shows average compensation increasing with years of experience. The grey “+” symbols represent each response (total compensation) at each experience level. There are also green, red, and blue lines showing connecting the average total compensation at each experience level (for everyone, women, and men, respectively).
The disparity seems to be the most pronounced in the 11-15 years of experience range. Zooming in on that data, we see the following:
There is a clear difference in the compensation levels for male and female product managers having 11-15 years of experience. At other experience levels, the disparity is reduced or reversed. There is also a significant disparity for respondents that reported zero years of experience, albeit with far fewer respondents.
Conclusion
I don’t believe that we can extrapolate from this data to reach crisp conclusions, but we can definitely acknowledge a disparity in the compensation levels reported by respondents to the 2008 survey. Generalization is dangerous, when looking at sampled data - especially when the respondents self-select for participation. Sometimes, that’s the best information we have, however - and it seems reasonable to form suspicions from the data. I suspect that the data does indicate inequality in the compensation of men and women in the product management role. The data does also indicate that product manager compensation increases correlate both with product manager age and experience.
Scott Sehlhorst has been helping companies achieve Software Product Success for more than a decade. Scott consults as a business architect, business analyst, and product manager. He has also worked as a technical consultant, developer, project manager, and program manager. Scott has managed teams from 5 to 50 persons, and has delivered millions of dollars in value to his customers. Contact Scott at scott@tynerblain.com, or join in on the Tyner Blain blog.
age bias article
Oh, and regardless of the gaps, I will use this to show my boss I'm underpaid :-)
breakout by age
Take a look at the "Product Manager Compensation by Reported Age" graph. The thick red and blue lines give you the 'average by age for female and male." Those are linear regressions, which removes the noise that you would get by looking at each reported age individually.
I did a similar analysis with 2007 data (but the data was noticeably different). You can get to it by clicking through to the original article on the Tyner Blain blog, then following the links back to previous analyses. I'm not sure if it is kosher for me to include the direct link here.
Gender Bias In Earlier Survey
Skillset / College Degrees
Level of technical acumen - good enough?
Degrees
Statistical Significance
As the data size continues to grow, more of these dimensional analyses can be performed.
Compensation components
I'd like to see in this or future studies the breakout of compensation to areas such as:
- Base salary
- Bonus
- Profit sharing
- Stock options and option types
- Bonus tied to product / product-line profit and revenues
- Other compensation components (beyond the little perks or benefits)
Thanks,
DAO
Other Slices
How would you compare $100K salary plus $10K bonus with $90K salary plus $20K bonus with $110K total compensation? Which is better and why? If you could help me understand the benefit of having the additional information, it would help me provide more valuable analyses in the future.
Thanks!
Recessionary Advantage
Food for thought...











Figures
*Shaking fist at sky* Pay me MORE you sexist jerks!