Monday, December 12, 2011

2010 ChemCensus is out: not much good news.

Credit: Chemical and Engineering News
The results of the ACS ChemCensus that was held in 2010 is out in this week's issue of Chemical and Engineering News in the writeup by David Hanson.

It's going to take a while to wade through all the data, but here are the important details (to me, anyway):

  • Reported salaries were all down, with the exception of median B.S. chemist salaries. 
  • As we've known for a while, the unemployment rate of all respondents in March 2010 was 3.8%, which is 0.1% lower than 2009. That 2009 was a 20-year high for the ACS Salary Survey. 
  • The "U6-like" number for the ACS Salary Survey data (including part-time, post-doc and unemployed) is the highest in ten years, at 11.9%. Hanson notes the other side of this statistic, the lowest full-time employment of chemists since at least 2010. I'll note that I predicted a U6-like number at 12% way back in 2009. Good to know I'm only a year late. 
  • Unemployment is negatively correlated with education, with B.S. chemists at 5.1% unemployment and Ph.D. chemists at 3.2%. M.S. chemists (the legendary 'most employable') come in at 4.8%. 
  • The "ten year" rule still holds; among age cohorts, the 30-39 age group has the lowest unemployment at 2.5% and the 40-49 age group jumps to 3.6%. The bad news isn't over yet -- the 50-59 age group has the highest unemployment at 4.7%. This is lower than 2009's 5.0%, but not by much. 
  • I note that the "small company discount" is alive and well. Median Ph.D. salary for a company of 10 or fewer employees? $49,200. Median Ph.D. salary for a company of 11-20 employees? $86,200. That's close to a 50% discount. Wow. 
More soon. Readers, here's the article. What do you see? 

The Eka-Silicon caveat: The ACS Salary Survey and the ChemCensus have both had relatively low response rates from members, which limits the extrapolatability of the data. This year's response rate for the ChemCensus was 47%, which is much lower than the 80% that professional survey types like to see. A discussion of this can be found here. At the same time, ACS unemployment numbers for their members more-or-less track the BLS survey data for chemists (see linked graph). 

One more thing: All of the 40,480 responses were hand-entered by the ACS' Deparment of Research and Member Insights. Research associate Gareth Edwards (I believe) was responsible for much of that. This is a huge undertaking, and he should be commended for it. 


  1. @CJ - I assume the response rate of 47% is calculated from total chemists (USA) = 94,100 (USDOL stats page:

    If so, this survey doesn't account for foreign affiliates or any non-chemist members of ACS (journos, politicos, business, etc)?

    Also: Thanks, Gareth!

  2. From the article, looks like total US-based ACS members:

    "The 2010 survey is considered a census in that it is sent to ACS members who are in the workforce and have U.S. addresses. According to the ACS Department of Research & Member Insights, ACS sent questionnaires to 85,652 members who were under 70 years of age and were not retired, emeritus, or in a student member category. For the survey, ACS defines a chemist as one who works in any one of 15 chemical disciplines or has a chemistry degree and indicated business administration, law, computer science, or “other nonchemistry activities” as their work specialty. Those in the chemical workforce are chemists who hold full- or part-time jobs, are on postdocs or fellowships, or are unemployed but actively seeking employment.
    From the 2010 survey, ACS received 40,480 responses. The 47% response rate is higher than the 41% response rate for the previous ChemCensus, in 2005, but lower than the peak of 53% in 1995. It is also higher than the response rate for the last annual salary survey, in 2009, when ACS mailed 20,000 surveys and the response rate was 36%."

  3. Do they have a table where they separate the chaff errr.... those with “other nonchemistry activities”? I wonder how they influence the employment numbers, if at all.

  4. There's a pretty strong conflation problem here that I didn't talk about, in that the ChemCensus is supposed to be deeper and wider than the yearly ACS Salary Survey. It appears to me (though I haven't asked directly -- and they're pretty good about getting back to me) that they're drawing some of the Salary Survey data from the ChemCensus data.

    So I guess what I'm saying, A11:49, is that when the true ChemCensus report comes out, they'll probably have a breakout of that sort of thing.

  5. I've got a personal caveat! Sweet!

    Now for the troll-y bit:

    One more comment- that's amazing work by Gareth, but if someone actually entered 40k surveys by hand, there is another 5-10% (of the total) error in all those values.

    Double-data entry is crucial for the reason that everyone will make errors in 5-10% of all data, and *it is will not be random* since people tend to make the same types of errors over and over.

    Again, I applaud the ACS' efforts, but for the world's largest scientific organization to publish (even informally) such a non-robust survey is just embarrassing.

    Use the dues to field a professional survey ACS! Please!

    1. Very informative. I'll be making some juice in the near future. Passing this onto my friends. Thanks.


looks like Blogger doesn't work with anonymous comments from Chrome browsers at the moment - works in Microsoft Edge, or from Chrome with a Blogger account - sorry! CJ 3/21/20