“You have to know the past to understand the present.” Carl Sagan
Whilst I was reading the Global Gender Gap Report 2018 section on the lack of women in AI I began to think about women and tech generally. My own experience of tech has been limited to running an online retail business and through doing that I quickly saw that there were not many women on the web development side of things. I strongly encouraged the young women who worked with me to go on coding courses. I then remembered that the girls’ school I went to in the early seventies had taught computer science and we learned the basics of binary coding. When I mention this to people today there is surprise and praise for the school that seemed so progressive for the time. But now that I have researched the topic I realise that it wasn’t that it was progressive to teach girls computer science. Until the mid-sixties computing programming was considered a female occupation. What was unusual was that my school could afford a computer at that time not that it taught computer science to girls. Many more girls studied computer science then than now. Today only 7% of UK computer science A-level courses are female. Just half of the girls that study IT & Tech subjects at school go into a job in the same field. Only 17% of people who are working in the tech industry in the UK are women. It is a similar situation in the US. In 1984 women made up 37% of those doing computer science degrees. Today it is 11%.
A glance at the big tech companies shows that their gender makeup does not reflect their users in any way. This is a cause for concern and they are increasingly being scrutinised and many are trying to improve the representation of women in their workforce.
Percentage of tech roles by women 2018
The lack of young women going into the industry will have a societal long term adverse impact, including the perpetuation of the gender pay gap as the power of tech grows. Apart from missing out on broad pool of talent, technology is being developed for solutions across many fields without the input of women at the outset. The feminist analysis of science (e.g. Harding 1986) has highlighted the need for women to not only do the science but to make the decisions on where science should focus its efforts.
Adding women into the picture at a later date is not mainstreaming gender equality. We are in the process of repeating the same mistakes of the past (in science) allowing mainly men to determine where a new technology is applied which will impact the lives of both men and women. Who knows if the problems the big social media companies are now having could have been lessened by having more women in the top teams at the outset?
I have talked to several women who have worked in Silicon Valley and they all say that the culture is macho and difficult to work in. Unlike more established industries tech does not have a history of excluding women and therefore has not had a long time frame for the development of a masculine culture (in this sense I mean one that suits men more than women). We have very few rational explanations for why so few women are in tech apart from ‘male culture’ and sometimes citing women’s lack of science qualifications and/or the essentialist fall-back position of women being less suited to that kind of work. Women have made inroads into almost all professions which were formally barred to them in the past for social, educational and cultural reasons. However in computing women’s representation proportionally is smaller now that it was forty years ago. This decline goes against the trend of increased female representation in almost every other work sector, including science sectors. How has this happened?
Programming is women’s work
We are familiar with work becoming feminised as women enter a field and sadly often the status of the work declines. But not so much the other way. However computer programming started off as women’s work and was made masculine – a process of masculinisation through the intentional development of a professional status and masculine culture. It is necessary to understand this specific and conscious gendering of the computing industry – how it was made masculine- in order to understand the current gender imbalance today.
At around the same time as the Gender Gap report was published I saw the obituary of Grace Roman (1925 – 2018) who was instrumental in the development of the Hubble telescope and I also noticed a social media post about the American Evelyn Berezin ( 1925 – 19 ) who designed the first word processor. So women were in there at the beginning on the frontline.
It didn’t take much digging about to find stories of many women who had been at the forefront of the computer industry but whose influence and input was left out of history until recently. (Tessa Dunlop’s book The Bletchley Girls and the 2016 film Hidden Figures which shows the story of how three black women were key mathematicians at NASA during the Space Race years in the 1960’s both brought women’s forgotten efforts into the public mind).
A female mathematician, Ada Lovelace, is widely recognised now as history’s first computer programmer when in 1843 she published an algorithm to calculate a sequence of Bernoulli (a Swiss mathematician famous for his pioneering work in probability and statistics) numbers intended to be carried out by Charles Babbage’s Analytical Engine. She was followed by many highly influential women. A wonderful article giving the brief stories of seven early women pioneers of programming (coding) can be found here along with the names of many other influential women.
These women were ground breaking computer scientists specifically in language development. But many other women worked specifically in the development of computer systems. “For decades, the work of many of these female computing innovators has been overlooked. However, every time we use our computers, phones or tablets, we are using tools that may not exist only for them” says Marie Hicks, author of Programmed Inequality.
Indeed all the early computer programmers were women. Women were the largest trained technical workforce of the computing industry during the Second World War and on through to the mid-sixties both in the UK and the US.
Women had made up the majority of the first programmers during the Second World War; Early on war efforts like the Colossus in the UK (codebreaking ) at Bletchley Park , 8000 of the 10000 employees were women. Women were recruited into these projects in large part because of the shortage of masculine labour during the war and also because much of the work was considered monotonous and therefore more suited to women’s ‘skills’. The work, particularly as part of the war effort, attracted many educated women who had little hope of gaining any other fulfilling work.
After the war, in the early years of the development of computers men worked with machines, the mainframes, the hardware; women programmed them, the software. Back then programming had no cachet. It certainly wasn’t seen as inspiring work. Every company that made or used computers hired women to programme them. The excitement and initially the money was in the mainframes, men’s work. Before computing became electronic, women were tapped as the ideal workers for what was considered rote, mundane calculation work. The skills required were considered ‘female’ dextrous, attention to detail, repetitive work.
Even though at a higher level which women were able to gain entry to, this work required advanced mathematics knowledge – everything from spheroidal trigonometry, to calculus, to accounting, to codebreaking – it was perceived as nonintellectual and, because women were doing it, largely unimportant. These women were actually called ‘computers’: before a computer was a machine, it was a job classification, like ‘accountant’ or ‘baker’. When electro-mechanical and then electronic computers began to automate this work, women carried over: operating, programming, troubleshooting, and even assembling the hardware of these new machines.
The gendering of work and skills
We tend to think of skills as objective and neutral whereas in reality they are socially constructed and require a continual input of cultural resources to maintain their status (or lack of it). What does or doesn’t constitute skilled labour?
“Social judgements often determine which job related capabilities are seen as necessary and which become optimal or even invisible” says Janet Abbate in Recoding Gender .. and this includes how skills are gendered for as Judy Wacjman said in 1991 “ the classification of women’s jobs as unskilled and men’s jobs as skilled frequently bears little relation to the actual amount of training or ability required for them. Skills are saturated with gender bias”
Women’s emotional labour was taken for granted as ‘natural’ and hidden until Arlie Hochschild revealed and named it as such in her book The Managed Heart. Several years later this was turned into emotional intelligence in management jargon, a recognised skill and one that men could learn. It is a now a ‘skill’ applauded and valued in men but arguably still seen as natural and expected of women. Ascribing lower value to women’s work combined with occupational segregation are arguably the strongest cause of the gender pay gap.
How the tech industry was made masculine
It was not until the mid-sixties that perception started changing both here and in the US. Computers, instead of being seen as intimidating behemoths that were only good for technical work, were now becoming integrated into every part of the work of government and industry. As their great power became more apparent, low-status women workers were no longer seen as appropriate for this type of work, even though they had all the technical skills to do the jobs.
The computerisation of the nation advanced, with computers becoming more important to the functioning of government and industry, and all the while this growing feminised machine underclass formed the infrastructure for many critical activities, from taxation, to scientific research, to economic forecasting. In the Civil Service women programmed and operated computers that ran the new VAT and ‘pay-as-you-earn’ taxation systems that enabled the welfare state and the NHS to be funded. Women also wrote software for prestige cutting-edge, international projects like the Concorde. Within government and the public sector, these women formed an underclass of highly technically trained and operationally critical technology workers – dubbed “the machine grades” within the civil service. As computers grew in power, having an underclass of highly-trained technical workers who were not aligned with the interests and goals of management was a potentially disastrous situation. At the time it was highly unusual for women to get to management grades. So Hicks argues that the British government began a decade-long effort to remove women workers from these newly-important technical positions and replace them with management-minded young men: people who could go from the boardroom to the machine room and manage computers as easily as people.
“Most businesses went out of their way to ensure that gendered hierarchies always put male managers on top, and the supposedly meritocratic civil service itself was the worst offender, carefully re categorising women and men workers into different job categories even when they were doing the same work. Irrespective of skills, job categories ensured that more men remained in positions of power while most women remained in functional roles, regardless of how they might have shown themselves to be capable.” (Hicks 2018)
As well as structural controls like grading and segregation, the professionalization of the industry and the development of an idealised and male computer worker added to further barriers to women’s entry in the same way as happened in the US.
The US computer industry at the outset also had an ambiguous gender identity, starting out requiring feminine skills and was also deliberately transformed into a masculine and high status profession. Women were very much a part of the early development of computers in the US. Six women were recruited to assist with the development and operation of the University of Pennsylvania’s ENIAC machine (Electronic Numerical Integrator and Computer) which was one of the first and most famous early electronic computers. These women were the first programmers (although they were not called that then) and all went on to have influential careers. Despite being instrumental in its development they were subsequently eliminated from the historical record until records of them were retrieved in recent years.
After the war there was an explosion in computers and in what they could do and companies were unable to recruit enough staff. An industry doubling in size every year could not afford to discriminate against women and besides the ‘qualifications’ required were still being negotiated. This was a similar situation in the UK. There had been no time to create an industry/sector/profession with boundaries to keep women out as existed in other science and management fields. People assumed that because women were doing this work, and the apparatus looked like a giant switchboard, that it was simple, concealing the reality that this work was highly technical and complex. This illusion lasted for many years. As in the UK these positions were kept low grade despite the fact that the operators were becoming more and more skilled in programming design as well as operation.
As time moved on, academic Nathan Ensmenger suggests that men began to recognise the importance of programming in the new age of information economy and edged into the profession. In order to elevate the importance of their work the first generation of male programmers began crafting a professional identity that effectively excluded women. By the late 1960’s already the tide had turned and companies sought out men not women as the work grew in status. Despite the tide turning large numbers of women relatively continued to study computer science until the mid-eighties. In the US women doing computer science at university peaked at 37% in 1984 and since then has steadily declined and is now only 11%.
Academic Nathan Ensmenger in his chapter ‘Making Programming Masculine’ in the book Gender Codes says “computer programming was gradually and deliberately transformed into a high status, scientific and masculine discipline. Men from established fields like physics mathematics and electrical engineering making the leap to a new field that had no history, no barriers to entry.” In the 1960’s the job title of computer programmer was switched into software engineer and in the process this embodied a more masculine vision of programming. Engineering was even more male then than it is today with approximately 1% of engineers being female in the 1960’s. Programming became coding, the former associated with women as they had done most of it but the latter now associated with men.
Some commentators have attributed the decline in women doing coding in the US with the advent of computer video games which appealed to men. “When Women Stopped Coding”. The popular myth of the ‘accidental’ role of the male almost adolescent computer nerd at home in the development of computer industry is still fuelled by popular culture. However Ensmenger contends that this ‘type’ of masculinity was present decades before the personal computer. Early associations of programming ability with maths puzzles and chess playing were reinforced through dubiously designed aptitude tests and personality profiles and then embodied into the hiring practices of the commercial computer industry. These became more a masculine ideal as time went on and as more men became programmers that fed into the stereotype that we have today. Studies were done to find out the attributes of the ideal programmer. What emerged was a male figure who was very good at problem solving and not very good at working with people (many of the women had fitted this description in the studies too!). The establishment of this male ‘nerdy’ rather antisocial stereotype who is brilliant but not very communicative continues to feed into hiring patterns today.
This re positioning of computer work was done on both sides of the Atlantic. As men entered the new computer sector, a lot of effort was made to develop it into a profession, including the development of formal programmes in academic computer science, the creation of societies, professional certificates and professional journals. The purpose of professionalization is to increase the work’s social status, improve opportunities for advancement and better pay. A profession also implied managerial authority. All of this like the professionalization of other occupations taking place in the sixties had the effect of marginalising and or/excluding women for whom a public life was still very restricted at that time e.g. like an emphasis on a four year degree course at a time when women were underrepresented at college whereas before aptitude tests allowed non college women to enter the field.
The professionalization of computer programming also involved the work being segmented into low status and high status levels. Segmenting work leads to segregation of workers and into a hierarchy of status and pay. Women who were still present in the computer workforce increasingly found themselves at the lower grades and stuck there, like in the UK, unable to cross into the managerial ranks. So while computer work had been open to anyone who could pass an aptitude test over the course of two decades 60’ – 80’s the development of a more professional status brought this to an end.
Another barrier women cite today as excluding them to the tech industry is the long hours culture. In the early years when computers were large and expensive, shift work was done to utilise the twenty four hour day and very often a nocturnal culture developed. In some companies women were often precluded from working in the night for safety reasons, hence erecting a structural barrier. Even after technical requirements for this night work disappeared the working through the night culture persisted and in some cases celebrated.
So in the space of only three decades the UK and the US computer industry shifted from low status work to a high status profession and in the process from a sector with a reasonably high representation of women to one where they are scarce. Women who were once welcomed into the sector now find themselves marginalised or excluded by a hostile culture and anti-social work practices. A combination of professionalization, and segmentation leading to segregation and the development of a masculinised work culture ensured that women despite having been on the front line of coding at its inception were demoted and eventually excluded. The development of an idealised ‘male’ version of a computer programmer was embodied into hiring practices as men took over the computer industry and made it their own.
Today many of the formal barriers are no longer there but neither are the women! However informal barriers in the form of culture can be just as powerful in their exclusion and /or marginalisation of women in the workplace and this includes the culture of long hours.( Baggulely 1991, Witz 1992, Rutherford 2011)
What this short history shows is that the excluding culture we associate with the tech industry was not there at the beginning in the 1940’s and fifties. Like all cultures it had to be created and recreated over the course of decades, replicated through the development of practices and institutions. For cultures to change they require resources and it is always the dominant group will create a culture that suits them best. Therein lies the paradox at the heart of all our diversity programmes, which these days focus on culture change. They do not work if men don’t back them. The positive from this story is that it is possible to change cultures over time and in those early years despite the social climate for women being much more prohibitive than today women were able to carve out interesting careers for themselves.
In 2017 a Google engineer was sacked for saying in a leaked memo that one of the reasons that there was so much disparity between male and female tech engineers could be explained by genetic differences, thus perpetuating and repeating stereotypes which even this short article proves to be historically incorrect. Maybe instead of just sacking the 28 year old man the company could have educated the whole workforce on the historical, social and cultural reasons as to why there are now so few women in those roles in Silicon Valley. And take it from there.