Friday, December 4, 2009

A Scientific Comment

The Lateral Thoughts article in the November Physics World (http://physicsworld.com/cws/article/indepth/40829) is a description by Rick Trebino, a professor of physics at the Georgia Institute of Technology in Atlanta, of the difficulties he experienced getting a comment on a paper with errors in it published. There is a longer version at http://www.scribd.com/doc/18773744/How-to-Publish-a-Scientific-Comment-in-1-2-3-Easy-Steps. And people wonder why women leave science to engage in more meaningful occupations!

My own theory on how to improve the reviewing process is that journals should employ people whose job is to go through reviewer’s comments removing the ad hominem remarks, unfounded speculation and unsupported opinions before they are transmitted to the author.

Thursday, November 26, 2009

The I-Word

The reviled word of the year in academic circles is, surely, impact. Is it really such a bad thing?

Firstly the Research Councils now require applicants for funding to assess the impact of the proposed research. The Research Councils invest £28 billion in research annually [RCUK Report, Excellence with Impact, http://www.rcuk.ac.uk/aboutrcuk/publications/corporate/excellenceimpact.htm]. It is reasonable to ask what benefit the UK derives from this investment. According to the guidance for applicants for funding [https://je-s.rcuk.ac.uk/jesHandbook/jesHelp.aspx?m=s&s=1202], impact is the demonstrable contribution that excellent research makes to society and the economy and can include fostering economic performance, increasing the effectiveness of public policy and enhancing the quality of life. In the impact summary applicants are asked to specify who will benefit from their research, how they will benefit, and what will be done to ensure that they have opportunity to benefit. Most of the Research Councils also require an Impact Plan that describes what the researchers will do to enhance the impact of the research. A diverse range of activities could be included in this plan, depending on what type of research is proposed. The Science and Technology Facilities Council guidance, for example, includes suggestions such as publications and publicity materials summarising the main outcomes in a way that beneficiaries will be able to understand and use or developing resources for schools or teachers.

Secondly, the current proposals for the Research Excellence Framework (REF) [http://www.hefce.ac.uk/pubs/hefce/2009/09_38/] that will replace the Research Assessment Exercise (RAE) are that research excellence, measured by peer review informed by citation analysis, will count for 60%, impact, assessed by an impact statement and case studies will count for 25% and research environment, assessed under the headings resourcing, management and engagement, will count for 15%. The draft ‘common menu’ of impact indicators includes

  • Employment of post-doctoral researchers in industry or spin-out companies.
  • Participation on public policy/advisory committees.
  • Changes to public attitudes to science.
  • Audience/participation levels at public engagement activities.
  • Positive reviews or participant feed back on public engagement activities.

These proposals could be very positive for women in science. My impression is that women often are concerned about the broader impact of their research. Also, explicit recognition of engagement activities is a welcome development, both in the funding process, where applicants have to assess the resource implications of their impact plan, and through the REF. In principle, these changes could encourage departments to take a broader view in their selection criteria for appointments and promotion. In addition, explicitly mentioning employment of post-doctoral workers as a criterion should encourage principal investigators to support the career development of their post-docs.

There are areas of concern. For example, will ‘hard’ measures of impact, such as research income from industry, outweigh ‘soft’ measures such as those relating to public engagement activities? Nevertheless, including an impact element both in criteria for funding and in criteria for assessment could help drive science departments to become more inclusive.

Sunday, November 15, 2009

Why gender diversity?

Why is it important to have more women in science? (Note: I am using science inclusively, that is, I imply ‘and engineering, technology, mathematics and medicine’.)

Broadly there are two approaches: justice and the business case.

1. Justice

In the preface to SET Fair: A Report on Women in Science, Engineering and Technology (SET Fair (2002)), Baroness Greenfield wrote

‘I have a strong personal conviction that any individual should be able to work in science to the extent, and at the level, appropriate for their personal choices, and commensurate with their abilities, but without gender as a constraining factor.’

I agree. A system that, for example, makes it much more likely that a woman with a baby or small child will be squeezed out of science than a man with similar responsibilities is simply unjust. This may appear to be an uncontroversial statement but one nevertheless hears it said that women ought to choose between having a family and having a fulfilling career commensurate with their abilities.

2. The Business Case

The SET Fair report lists four business reasons for greater gender diversity in science:

  1. Competiveness – the UK needs the best people engaged in SET
  2. Return on Investment – we do not want to lose expensively trained individuals
  3. Benefit to Science – a more diverse workforce brings new perspectives and priorities
  4. Missed markets and skills – 50% of customers are women so companies could be missing out on potential markets.

I will examine some of these benefits more closely in subsequent posts, for example, what does ‘best’ mean? However, let’s look at this just as a business case. To whom to the benefits accrue? Firstly, to the UK as a whole: there are compelling demographic arguments that the scientific workforce will need to be drawn from a more diverse group than is currently the case in order to meet demand. Also, the loss of expensively trained individuals is a loss to the country as a whole. Secondly, technology-based companies, both because wider participation would widen the talent pool and through the new perspectives and potential new markets that greater participation of women in the scientific workforce would bring.

What happens, however, at the level of a research group? Suppose a principal investigator has a project with two years funding and hires a woman who subsequently wants to take six months maternity leave. In the UK most research sponsors will pay for the maternity leave, but they won’t necessarily pay for cover, even if a suitably qualified person can be found, or to extend the contact to make up the lost time. Even if the research sponsor will pay for cover or a contract extension, the research programme has been disrupted, with the risk that the principal investigator’s ability to attract continued funding has been compromised. Or consider the case of a university department. If they have a woman member of staff who wants to take maternity leave can they recover her maternity pay? Is it all of it or just the statutory component? What if they need to employ teaching cover? How will they ensure supervision for her research students? If departments have to meet these costs from their own resources then departments with more women taking maternity leave are at a financial disadvantage compared to other departments. So at the level of the research group or university department we have actual quantifiable costs and perceived risks of employing women, but vague and intangible benefits mostly accruing to other people. (I suspect, for example, that most principal investigators are more concerned about whether their own ideas are funded than with the benefits to science of incorporating a more diverse range of perspectives and priorities.)

I have used examples from the university sector but I suggest that similar considerations apply whenever short term financial performance or short term productivity have high priority.

If business case arguments are to be effective then the incentives at the level of research group and department have to be aligned with the goals from which the UK as a whole will benefit.