Microsoft has once again fallen back on a discredited privacy defence to deflect questions on the proportion of its certified engineers who are women.
Last month, Microsoft told El Reg that it would try and work out how many women had qualfied as Microsoft Certified Systems Engineers.
The vendor has now clammed up on the issue, issuing a written statement citing data privacy concerns and refusing to answer further questions. But there should not be anything to be embarrassed about. It is indeed likely that a small proportion of the 1.7m certified Microsoft professionals in the world are women. This is usual for an industry in which around 20 per cent of IT workers are women and most of them are stuck in lowly jobs.
Debbie Ellen, a British academic who has been putting pressure on Microsoft over its gender policy, said Microsoft UK's previous diversity manager, Suzy Black, had told her she was looking at whether contracts with certified examiners Pearson Vue and Thomson Prometric could be renegotiated in order to oblige them to collect gender data.
It is understood that examiners and trainers do not collect gender data intentionally and that this makes it impossible to tell how many engineers are women. Microsoft cannot confirm whether contract negotiations are still on the table. It has previously said both that the data does and does not exist.
It since emerged, thanks to a Register reader, that the MCSE section of Microsoft.com requires .NET passport registration, which requires gender to be given. Does that mean the data might exist? Again, Microsoft remains schtum.
There are other indications that the data could exist. Microsoft sources revealed to InfoWorld Australia in 2001 that of 10,000 MCSEs in the country, just 326 were women.
Ram Dhaliwal, Microsoft UK manager for training and certification, said in a statement last week that knowing how many MCSEs are women was not as important as having employment policies that support them, such as "flexible working policies" and "good career opportunities".
Yet statistics about the numbers of women are necessary to formulate such policies and see if they are working, said Ellen. As the numbers of women working in IT have fallen proportionally by a quarter in the UK since 1997 the statistics suggest policy is not working.
Further, statistics can identify where the problems are. If, say, 10 women attend a certified training course, three take the exams and only one ends up taking a job as an engineer, it may be that the training is not working for them.
Ellen had sought Microsoft's gender data to see how successful standard training programmes were compared to a women-only scheme at getting women into the industry. Without the data, a proper assessment cannot be made, though it is increasingly thought that women fare better on single-sex training courses where they are not suppressed by overbearing, often sexist, men - the launch this school term of Computer Clubs for Girls is a case in point.
Dhaliwal also pointed to the work of Microsoft diversity manager Bronwyn Kunhardt in tackling the gender issue. While Kundhart has shown her commitment to the cause she is, by her own admission, charged mostly with addressing gender equality within Microsoft. The sub-contracted certified training regime and the army of certified engineers all fall through the net.