Gender and STEM: no shift required
Gender and STEM: no shift required
Deborah Tatar

Overview
In the past thirty years, several waves of opportunity have come successively closer to realizing Papert's vision of a world in which children can self-actualize as owners and creators of technology. Each wave, starting with Logo, has had strengths and limitations and while some have had considerable reach (FIRST Lego League, for example), none have as of yet become fixtures of childhood.  Now, part of the opportunity that comes with a switch from a STEM to a SEAD perspective is the chance to build foundations for female---and more widespread male---participation in computing on a wide, humane platform in which the outside world is involving, inviting and discovering rather than persuading, cajoling and selling.  In particular, recent tools associated with the Maker or DIY ("Do It Yourself") movement have the potential to increase embodied, craft-oriented, performance-focused behavior.  Girls (and a range of boys) can now create inexpensive personalized objects that cause them to rub elbows with technology and technological thinking without having to first (or ever) label themselves as one of "them," the kind of person that actually likes technology. They can tinker, both with creations and identity. They can develop skills that will help them no matter what they go on to do, and their relationship to those skills can change over time.  The crucial opportunity, ironically, lies in the relative unimportance of the technology in defining the students' projects.

The Sewable Computing Opportunity
Although tools such as Leah Buechley's sewable electronic components ([4];http://web.media.mit.edu/~leah/) are new, the opportunities they present resonate with older successes. They have social and technological properties that have been to some extent lost with the rise of internet-based computing.  Additionally, the world of young people has traditionally included legitimate peripheral participation in activities that could be pursued in a more sophisticated fashion by adults.  We are interested in children's relationships to sewable computing, but these activities are compelling for adults in a way that Logo, for example, never was and was never thought to be [13].

Sewable electronic components consist of a familiar selection of sensors, actuators, and power components that can be sewn like buttons, snaps, or trim using conductive thread (https://www.sparkfun.com/categories/135).  LED lights, buzzers, buttons, toggle switches, light and motion sensors, batteries and the board are all equipped with eyelets (grommets) that serve the double function of allowing them to be attached to material and act as elements in a circuit.

Sewable computing is only one aspect of the larger Maker or DIY movement.  But sewable electronic components are particularly exciting when we think about women, when we think about digital divide issues, and when we think about STEM careers.

Women and Sewable Computing
A number of factors make sewable computing a likely venue for becoming fixtures of childhood.  The activities can themselves be social, just as knitting and quilting are often social acts. With even modest mastery, outcomes can be a personal expression on the part of the maker.  They can be worn.  Most importantly, they can be given. Gift giving is a foundation of civilization. Gifts do not have to attain perfection or win competitions to be valuable.

Sewable computing activity can spread easily. After attending a 2-hour sewable electronics workshop, I organized a novel 1-hour sewable computing activity for 95 local 7th grade girls during Virginia Tech's 15th Annual Women in Computing Day. Scaffolded by Computer Science volunteers, the girls sewed a simple circuit with a battery, button-switch and LED onto felt bangles, which were then further decorated with beads and embroidery thread.  There were audible sighs of disappointment when the end of the activity was announced.  We had to pry the groups out of the room (even with the materials to continue at home).   The activity involved a great deal of volunteer effort, but less than $12/student in materials.

Figure 1: 7th Grade girls and Virginia Tech students use sewable electronics at the Association for Women in Computing's 15th Annual Women in Computing Day, April 2012.

Pragmatic Access
Sewable electronic components permit many levels of participation in handicraft. The thresholds, in particular, are very low.  While some adult encouragement and guidance is required, the initial level is more comparable to that required for lanyard-making, knitting or embroidery than most interactions with electronics. The physical dexterity to sew with large needles and thick thread is in most cases attained by early elementary age.

Projects of moderate complexity can be funded with the kind of money "tweens" earn babysitting: LEDs, switches, and buttons are $1-$2; the conductive thread is about $.26/yard; a battery case is about $5.  A project can involve parallel and series circuits and remain under $20 including the cost of an inexpensive garment. A consideration is that, while the individual components are light in weight, the decorated garment has to be sturdy enough so that it will not stretch too much.  Stretching depends on both material and weave.  The lightest weight--and least expensive--t-shirts are poor candidates.  But the components are washable.

Participation may be possible through informal mechanisms including after-school clubs, libraries, Boys and Girl's Clubs, Girl Scouts, camps, community centers, homes, and through more formal academic middle or high school classes, such as  vocational training, art or computer science.

All this is good. However, taken as a whole, cost becomes more difficult.  At $22, the actual computer boards are expensive. And projects can become quite expensive indeed if the electronics are very elaborate or if the cost of the decorated item (shirt, scarf, jacket) is high. Furthermore, programming the boards requires access to a computer.  However, projects do not require ownership, just access.  Finished products stand alone.

Tying Sewable Electronics to STEM
Sewable components provide a practical, general, inexpensive opportunity for engagement with a wide range of creative activities, easily and comfortably organized, with low monetary or knowledge thresholds for participation and high potential for deep-seated widespread involvement. The creative opportunity can be situated within a wide range of extant formal and informal settings.

But there is also STEM learning potential.  Even the simplest sewable computing project has systematic elements.  The simplest projects still must involve sewing a circuit. As long as students are sewing circuits, they are engaging with an intrinsic, embodied connection to the physics of electricity and electronics.  Circuitry, resistance, power, and signal degradation all come with the territory.  Solving problems is a core scientific activity.

One focused STEM opportunity is to encourage girls to pursue computer science, a field that is markedly lower in female and minority participation than most.  There are two aspects to this: 1) create learning pathways 2) do not mess it up.

Create Learning Pathways
At a certain point, we hope the students' imaginations will become too complex to accomplish their projects simply by sewing. They will want to use the computer. In this scenario, their first encounters with the programming interface will be driven by their image of what they would like to accomplish.  The drives---gift giving, adornment, curiosity, self-expression, sociality---to create particular special items will motivate learning and exploration.
One challenge is to make those first encounters intriguing in the way of Logo: no threshold, no ceiling. Right now, several computer interfaces work with sewable computing components.  One lovely interface builds on a Scratch approach to teaching children to program, using visual building blocks ([2], Fig. 2). However, even more focused environments are needed that explicitly scaffold movement from physical representations to digital ones and then scaffold movement into more sophisticated programming efforts.  The issues here are not only teaching how the computer operates, but awakening the question of why.  Why are some important elements (LED's) represented in the program, but others (conductive thread) not?

Figure 2: Scratch-like interface to Arduino programming, developed by ModKit [2].

B. Don't Mess It Up
The potential exists and is exciting.  But the opportunity can easily be lost or preserved only for the most privileged.  There are three primary ways to curtail it with the best of intentions: over-meddling, over-marketing, and over-measuring.

i. over-meddling
Over-meddling is putting undue focus on the novel elements in the situation, on what we believe that we are changing, and insufficient emphasis on the relationship between old and new elements that allows the new to succeed.  For example, suppose that we want to create sewable computing clubs for women.  The temptation is to focus on understanding the sewable computing components and related activities.  But the transformative success of the endeavor rests every bit as much on creating or finding the settings and understanding how they work with our new activities.

From a personal perspective, the steps to take to encourage women's participation can seem very simple and direct, "show them what it is".  Yet, unhappily, the history of direct efforts to create positive social change is strewn with disappointment.

The great social psychologist Kurt Lewin, and his students left us with two ground rules that we would be wise to remember [8, 9].

The first is that the situation that you go into has its own strengths.  It is far easier to undermine the strengths of the current situation than it is to build comparably strong new ones.  Thus, all the resources that were poured into the Cambridge-Somerville experiment failed to compensate for the ways that those resources undermined the role of local groups and churches in providing support [8,  p. 189]. It is for this reason that the mantra that we today in public discourse that the public school system in America today "could not be worse" is so very destructive. It is a light claim to make that shows a frightening lack of imagination and ignorance of history.  To become a fixture of childhood, sewable computing must be pursued in small, locally sensible steps.

A second key concept is that certain highly influential components of a social system (called channel factors) may appear insignificant from the outside. This concept is embodied in the power of the turtle at the bottom of the stack in Dr. Suess' Yertle the Turtle, or canonized in the idea of the pebble that diverts the stream.    In my own work with the large-scale investigation of a highly successful mathematics intervention that involved technology [14], a curriculum that featured soccer examples was a big hit with the white and Hispanic populations of Texas.  But, forewarned, we knew that urban African-American children would need a different context.  Getting the examples right was not a sufficient condition for success, but it was a necessary one.

One opportunity with sewable components is best explained by contrast.  FIRST Lego League builds on the pre-existing popularity and familiarity of Legos and a certain way of engaging with Legos practiced primarily by young boys.  The movement into the institutional structure of a league depends on the congruence of that kind of play and the movement of children at a certain age from free play into rule-bound, team-oriented activities that often involve competition and reward.   This works well. There are lots of benefits to participation and many girls like it too.

But there are also many people and communities who do not like to participate in such structured, team-based, goal-oriented activities and who reject overt competition.   I have sometimes written about acompetitive activities, such as jump-roping [16].  Jump roping can be pursued in a competitive way or a non-competitive way.  But the difference is in the player not the game. Very often, and with good reason, some men and many women prefer to show their competence in the form of helping behavior.  CompuGirls ([15], http://sst.clas.asu.edu/about/compugirls) builds on the particular strength of this value in Native American communities.

Rather than (or in addition to) giving people external rewards, we need to create or find the sewable computing equivalent of this situation.  Busy hands, individual projects and talk can be very fulfilling without the need for a permanent institutional superstructure. Rather than aggrandizing the team or the institution, gift-giving is about the thought and the relationship.  Adornment is about pleasure.

The hope would be that, if we can refrain from over-meddling by building on existing structures or creating ones with emotional resonance, we can use human sociality to educate a wide enough range of people to allow sewable computing to become a fixture of childhood.

ii. over-marketing
"One lie undoes a thousand truths"---East African Proverb
We start from the idea that we would like more women to engage in STEM fields.  Why?  Some reasons have to do with the women themselves.  It seems to modern American society wrong or unfair if women do not participate in equal numbers in elite vocations.  Other reasons have to do with a perception of the benefits of involving women in STEM activities.  For example, NCWIT promotes on its brochures research showing that mixed-gender teams work better than single-gender teams.  The idea is that women should be involved in computing because they are needed.

Notice that, in a brute force way, the desire to involve women equally stems from the belief that girls are essentially the same as boys, while the desire to persuade them that they are needed entails the idea that women are somehow essentially different, that a women's perspective is a special contribution.  So what are women, and why are they wanted?  This is very confusing, even if one considers women as a coherent group.  It is even more confusing if one considers the range of women and girls, their hopes, dreams, and prospects. It is yet more confusing when one considers the range of high-school, college and work-place environments that they might encounter.  The situation is confusing.

Over marketing describes a complex of persuasive behaviors that can be seen as producing a desired outcome. We want women to go into STEM fields, and so we try to persuade them using the tools at hand.  This is deeply problematic for two reasons.  First, programs based on simple persuasive tactics are vulnerable to several kinds of deceit and, second, such programs may engage in practices that undermine existing interest.

The epistemic basis for work on persuasion is the question of how people end up doing things that they would not otherwise want to do (for example, Asch's classic line length study [1], the Milgram's shock experiment [11] and the Stanford Prison Experiment [17]). These concerns overlapped with and migrated into the sales and advertising techniques that surround us, nicely summarized in Cialdini [5].

But pursuing gender and ethnic diversity in STEM fields is not the same project as persuading people to buy a car.  We must ask ourselves whether we care about the sheep or the shearing.  We know that we can get people to do things that make them deeply unhappy. I like to presume that, fundamentally, we don't want to do that.  A better plan is to find or develop precisely those women or minorities that could be genuinely engaged with STEM fields, and that might not otherwise consider STEM fields.  For one thing, there is no particular reason to believe that the women that might succumb to an offer made with the right persuasive overlay will be the women that most likely to prosper in STEM fields.

We should learn from one of the most successful educational enterprises of all: the enterprise whereby middle class (white) toddlers learn to love reading by being read to.  The child appreciates the ball-of-wax in which s/he is held, talked with, and entertained with world knowledge and pictures. They learn that we value reading (and lots of other things) because it is what we do with them.  Unless some other problem intervenes (dyslexia, for example), the child that has been read to walks into school ready to learn to read.

Part of the significance of DIY movement is that it is interesting enough for us to do, and for us to do with them, without reference to the future we hope for.  Furthermore, the truth of the child's experience with sewable components does not put us in the position of over reaching.  When we seek to persuade, we are liable to offer too much, substituting extrinsic motivation for more powerful and enduring intrinsic motivation [8].  And sometimes in our efforts to persuade, we lie.  We lead women to believe that they will be valued and treated well in situations that are not set up to value them.  We are tempted to deny that STEM fields are difficult. But they can be quite difficult intellectually, emotionally, and pragmatically.  At their best, the work is absorbing and challenging, just as medicine and many other worthwhile endeavors.  Difficulty, seen through this perspective, may even be a selling point for people who are inclined to love an area.

The most reliable way to avoid over-marketing is to focus not on persuasion but on voluntary involvement and pleasure. Papert's vision of the child as bricoleur, or tinkerer, in the domain of computing is one of the child finding order and playing with it [13].  Systematicity and patterning have pleasures and attractions for many people outside their formal treatment in mathematics and science. Although, eventually, all computer scientists have to learn system properties, understand formal languages and engage in top-down thinking, their discovery of top-down thinking may arise from bottom-up, embodied exploration.

iii. Over measuring
Over measuring is utilizing the forms of quantitative experimentation and inferential statistics without a sufficiently firm grasp of the phenomena and circumstances being measured. We provide limited resources to support the promotion of STEM careers, and even more limited resources to support arts or crafts.  Naturally, funders want to know whether their investments are worthwhile.  So too do the researchers who are spending their lives in well-meant endeavors.   But the simple question, "can we create fun, nurturing environments for girls in which they are exposed to electronics and computer science?", is self-evident.  Of course we can, with enough money and good will.  The harder questions---whether providing such circumstances and whether particulars of such circumstances lead more women to choose STEM careers---are part of complex epidemiological problems.  We know something about indicators, but all sewable computing or any program can really do is invite participation, not determine it.  And these issues do not really need measurement, just documentation.

When thinking about particular questions, such as whether women exposed to sewable computation will go into Computer Science, it is important to remember the scale of change involved; we would have a lot of female computer science majors at Virginia Tech if .1% of the graduating seniors from public schools in Virginia enrolled in the program, as opposed to the .016% that actually do. To detect this scale of change against noise is, at best, extraordinarily difficult and expensive.  To compare the goodness of one program against the goodness of another is impossible.

Unhappily, we can mess up the project of sewable computing by over measuring.   The demand to measure is not neutral.  The terms of the measurement become the terms of concern in describing the situation, especially in poor or needy circumstances. They push researchers and practitioners towards over meddling and over marketing.  They may encourage adults to push girls too quickly or too hard towards the technology, turning that into a power struggle rather than an invitation.

In all cases, we must think, as a society, about what is most important to us and how we intend to get there.  Perhaps less emphasis on measurement would risk some waste.  However, it might also allow a more sophisticated descriptive phenomenology to arise and contribute to participation by the neediest young women.

Suggested Actions
Goal: To promote sewable computing knowledge and practices in a way that will allow them to become fixtures of childhood, and thereby to lay the groundwork for increased mastery of STEM skills and increased participation in electrical engineering and computer science professions for women.

Approach: Pursue a variety of deliberately distributed activities that support widespread local ownership of sufficient knowledge and development of local taste cultures and communities.

Stakeholders: Funders interested in increased involvement by women in STEM careers.  These include NSF, CSTA, CRA-W, ISTE, and NCWIT. Google and Microsoft have been known to fund education work with a CS focus.   The High-Low Tech lab at MIT, which derives an income from the sale of sewable computing items, may also have an interest.

Roadblocks:
i. accessing the people who might become part of an enduring community.
ii. helping implementers refrain from seeing the "real" purposes of sewable computing activities as teaching STEM.
iii. developing and deploying sufficiently simple, fine-tuned computing environments and materials.

New Opportunities:
1) Provide seed money for many small sewable computing efforts, housed in a variety of public and private spaces.   Think of this as loosely analogous to micro-financing.  Training and materials will be made available to individuals or small groups with passion and local knowledge.  They will not pay back the money, but instead undertake to pass along their learning to the communities they strengthen and create.   Gather low-stakes, phenomenological reports.
2) Training through cascading, small, inexpensive workshops.  One starting place would be workshops attached to conferences that highly trained people will already be attending, such as AERA, sigCSE, IDC and the Grace Hopper Celebration of Women in Computing.
3) Create even simpler computing environments, more focused on the electronic underpinnings, for the transition into computing.
4) Create and build upon existing simple resources that teach about electronics and problem solving in electronics (such as resistance, capacitance, properties of the thread).  Leah Beuchley's group has a number of tutorials (http://hlt.media.mit.edu/?p=1283). These are, however, oriented towards a lovely but specific taste culture, and a relatively high capacity to learn through written instruction.
5) Create one or more simple inexpensive, low-production value magazines for students to publish the projects, modeled on the Lego magazine and possibly partner with Beuchley's group at MIT or Scholastic. Content will consist of new technological items, pictures of kids with their creations, plans for designs and computer programs.  (Maker magazine serves an older, richer, up-market population; Lilypond (http://lilypond.media.mit.edu/) is more polished and assumes a high degree of internet connectivity. )
6) In all of the endeavors, the first effort should be to support existing local taste cultures.   Latina girls in Texas do not necessarily care about the same kinds of projects as New YoRicans.

II) Goal: develop a better phenomenology of how women enter into STEM careers, especially computer science.  Utilize epidemiological as well as qualitative modeling.  Conduct longitudinal research on the development of interest in STEM careers.

Acknowledgements
Special thanks to Laura Trutoiu who first got me interested in sewable components; Raquell Holmes who promotes the relationship between STEM and performance; Carol Strohecker at Center for Design Innovation at the University of North Carolina; Ben Knapp, Liesl Baum, Barbara Ryder and many students from the Institute for Creativity, the Arts and Technology (ICAT) and Computer Science at VT.  Also, thanks to my mother and step-mother who taught me how to sew, knit and crochet and my father who taught me how to solder, without demanding any particular outcomes.

References

[1] Asch, S. (1955) Opinions and Social Pressures, Scientific American 11:32.
[2] Baafi, M.  (2012) Drag-and-Drop Arduino Programming, MAKE, 25 p. 52
[3] Bransford, J. and Schwartz, D. (1998) Rethinking Transfer: A simple proposal with multiple implications.  Review of Research in Education (24) p. 61-100.
[4] Buechley, L. and Perner-Wilson, H. 2013. Crafting Technology: Reimagining the Processes, Materials, and Cultures of Electronics, Forthcoming in the Journal ACM Transactions on Computer-Human Interacation (ToCHI).
[5] Cialdini, R. (2006) Influence: the Science of Persuasion.  New York: Harper Collins.
[6] Jepson, B.  (2010) Modkit: a graphical programming environment for Arduino, MAKE, Retrieved from: http://blog.makezine.com/2010/09/24/modkit-a-graphical-programming-en/ on November 2, 2012
[7] Kutznetsov , S., Trutoiu, L., Kute, C., Howley, I., Siewiorek, D. & Paulos, E. (2011) Breaking Boundaries: Mentoring with Wearable Computing.  CHI 2011, p. 2957-2966.
[8] Lepper, M. R., Greene, D. and Nisbett, R. E. Undermining children's intrinsic interest with extrinsic rewards: A test of the overjustification hypothesis. Journal of Personality and Social Psychology, 28 (1973), 129-137.
[9] Lewin, K. Field theory in social science; selected theoretical papers. Harper and Row, City, 1951.
[10] Lovell, E. Getting Hands On with Soft Circuits: A Facilitator's Guide.  Retrieved from: http://lilypond.media.mit.edu/  November 2, 2012.
[11] Milgram, S. (1974) Obedience to Authority: An experimental view.  New York: Harper Collins.
[12] Nisbett, R. and Ross, L. (1991) The Person and the Situation. In (T. Nadelhoffer, E. Nahmias, & S. Nichols, Eds.) Moral Psychology: Historical and Contemporary Readings.  Malden MA: Wiley-Blackwell.
[13] Papert, S. Mindstorms: Children, Computers and Powerful Ideas. Basic Books, New York, 1980.
[14] Roschelle, J., Tatar, D., Shechtman, N., Knudsen, J. (2008) The role of scaling up research in designing for and evaluating robustness. Educational Studies in Mathematics, Special Issue on Democratizing Access to Mathematics through Technology: Issues of Design, Theory and Implementation— In Memory of James Kaput (S. Hegedus and R. Lesh, Eds.). 68(2), p. 149-170.
[15] Scott, K., Aist, G. & Hood, D. W. (2009) CompuGirls: Designing a Culturally Relevant Technology Program, Educational Technology, v49 n6 p34-39 Nov-Dec 2009
[16] Tatar, D., Lin, S. and Lee, J.S. Playground Games and the Dissemination of Control in Computing and Learning (2008). In (DiGiano, C., Goldman, S. & Chorost, M., Eds.) Educating Learning Technology Designers. Mahwah, New Jersey: Lawrence Erlbaum Associates.   p. 230-257
[17] Zimbardo P. (1973) On the Ethics of Intervention in Human Psychological Research: with special reference to the Stanford Prison Experiment.  Cognition 2(2), 243-256.