Presenter: Adriana Moscatelli, CEO, Play Works Studio
In 2014 Play Works Studio received funding from the National Science Foundation to advance the design of a Robot Game for girls, and to conduct a psychological study of the game in collaboration with the University of Washington’s Institute for Learning and Brain Sciences. The results of the study were encouraging: playing with the Robot Game increased girls’ efficacy and motivation in STEM. Girls who played the Robot Game reported equal motivation to boys in terms of how fun programming is, how fun robots are, and how good they perceived themselves to be with robots. In the US not enough women choose careers in STEM fields, with computer science and engineering attracting the smallest percentages: 20% female undergraduates (NSF, 2012). Interventions such as programs, games, and toys that interest girls can provide early and positive experiences with STEM concepts and help level the playing field in STEM disciplines.
10. H1 2013 H2 2013 H1 2014 H2 2014 H1 2015 H2 2015
V1 prototype
user test
V2 prototype
University of
Washington
CS Ed Week
V3 prototype
1st study @
UW’s I-LABS
N=96
children
Play Works
incorporated
Paper
prototyping
1st user test
NSF SBIR Grant
Go to
Market
Department
of Education
SBIR grant
Studies & Summer
Camps
Boze Elementary
11. Research led by Allison Master and Andrew Meltzoff
Institute for Learning & Brain Sciences
University of Washington
12. Research Questions
What are young children’s gender stereotypes
about programming and robotics?
Can experience playing a robot-
programming game increase girls’
technology-related motivation (without
decreasing boys’)?
13. Participants
96 6-year-old children (48 girls, 48 boys)
Randomly assigned to one of three conditions
– “Robot” treatment group
– “Game” control group
– “Baseline” control group
14. Robot Group
Children chose a “pet” robot and used a smartphone
to program the robot to navigate an experimentally
specified spatial path
15.
16. Two Control Groups
“Game” control group: children played a storytelling
card game
“Baseline” control group: no games
17. Dependent Measures
STEM-Gender Stereotypes (4 items; 1–4 scale):
– Are boys or girls better at science, math, programming, and
robots?
Technology-related motivation (3 items; 1–6 scale):
– How fun is programming?
– How fun are robots?
– How good are you with robots?
18. STEM-Gender Stereotypes
All error bars are +/- s.e.
*p < .05, ***p < .001
1
2
3
4
Science Math Programming Robots
Whoisbetter?(1=girls,4=boys)
* ***
19. Programming Enjoyment
Condition main effect: p = .017
Gender main effect: p = .013
**p = .01, ***p < .001
1
2
3
4
5
6
Girls Boys
Howfunisprogramming?
(1=low,6=high)
Both control conditions
Robot condition
**
***
22. Conclusions
• Children held strong stereotypes associating
boys with robots and programming
• The 20-minute experience increased girls’
technology-related motivation
• It increased girls’ self-efficacy, and the
robot treatment eliminated the gender
gap in self-efficacy
23. Next Steps
• Classroom tests with 1st grade students
• Curriculum integration
– Math
– Reading/writing
– Life sciences
– 21st century skills
• Support for teachers (tools)
• Support for different game styles
– Puzzle/logic
– Strategy/RPG
– Simulation/sandbox
– Combat/arena
– Infiltration/tower defense
My personal background: I worked at the Stage 3 Lab back in 1999 when I was a graduate student at Carnegie Mellon University. I had the pleasure of meeting Randy Pausch then, he had a great influence in all of us.
I graduated an had “serious” jobs designing hardware and software products and user interfaces at places like Microsoft and Nokia. In 2008 I ended up in games again! I was lucky to end up at Wizards of the Coast, the creators of the games Magic the Gathering and Dungeons and Dragons. It was a great training ground to learn about how to design successful strategy games that have a lot of complexity and potential for educational value, particularly in STEM.
I also worked in the development of the Pokemon online trading card game. Unfortunately 98% of Magic players are men. I was wondering how to attract more women but I realized that it was very difficult because the culture of the game is hard to change. I saw that more women play Pokemon and that game me the encouragement to build a culture around a new game that attracts more girls!
Back in the mid 80s we had almost 38% women enrolled in Computer Science at the undergraduate level. Fast forward to 2012 and we dropped to 20%! Some schools are doing better, Harvey Mudd famously graduated 45% women from their undergraduate program in 2012. CMU is doing better and we are seeing an improvement but as an average, we still have more work to do. There is nothing about computer science and engineering that makes it a field for men. And what’s more, women in STEM related jobs make 33% more than women in non-STEM jobs. We need to attract more women into STEM fields!
When I think back about my childhood I remember playing with LEGO, back in the 70’s and 80’s LEGO was not gendered. We could ALL build whatever we imagined.
Fast forward to today and the toys and games industry pushes forward the message that if girls want to play LEGO, it should be around the themes of beauty and fashion. While it is not true that only boys and men play video games, the majority of the public associates video games and robots with boys. Robotic toys are very successful at attracting children into STEM disciplines and yet it is very difficult to engage girls in robotics events.
And we know…
I spent time as a mentor in a LEGO league group. These events are organized by the non-profit organization First. They are very successful in teaching very important skills beyond an interest in engineering and science. An the groups of girls have been successful. But it is so hard to attract them! I worked with a team of middle school children which had about 18 children, only 2 of which were girls who attended regularly. It was very hard to engage them. When someone is in a group in which they see themselves as a minority, they don’t feel a sense of belonging. It is hard to engage them. And we know from the research that society perceives robots and vehicles as “male”. When I saw the teacher excited about being able to program the robots to “dance” I thought: I can build a platform that’s more accessible to girls. They don’t have to be vehicles and they can dance.
I started a company to design RoBees, a platform designed to teach girls and boys 6 to 10 the concepts of programming in a fun entertaining way. Most importantly it is designed for both girls and boys!
Here is a timeline of our progress. Today I will talk about the results from our study in collaboration with the Institute for Learning and Brain Sciences (I-LABS)
The study had 3 groups, those who played the robot game, those who played a storytelling game and those who played no game
96 first grade children total, both girls and boys
In the robot group condition, children spent 20 minutes playing a game in which they chose a “pet” robot and used a smartphone to program the robot to navigate an experimentally specified spatial path made out of hexagonal tiles, which could be laid out in many different spatial designs. Here you can see what the smartphone screen looked like; children used drag-and-drop programming to program the robot to move forward, turn left or right, and create loops to repeat instructions. The researcher first demonstrated how to program the robot to navigate four spatial paths, and then children themselves programmed the robot to navigate up to eight additional paths.
[Video 75D] So here’s a video of one of the participants programming her turtle robot. She has to count the tiles and program the robot to move correctly along this path, which is different from the paths she’s seen so far. It’s a little hard to hear, but she says the directions out loud to herself as she programs, so the robot will go forward, forward, turn left, and then go forward, forward again to get to the green goal tile, and she pushes the “run” button at the end to start the robot. And it’s hard to hear what she says at the end, which is, “Man, this is awesome!”
We wanted to have a control group that controlled for the experience of playing a fun game with the researcher, so in the first control condition, the “game” control group, children spent 20 minutes playing a storytelling card game, where they were given a series of cards with a person, object, or idea, and had to tell a story involving those cards. In the “baseline” control group, children did not play any games.
We had two main types of dependent measures. First, we measured STEM-gender stereotypes on a scale from 1-4 about whether boys or girls are better at science, math, programming, and robots. We defined programming for all the children as when you tell a computer or robot or phone what to do. We also measured technology-related motivation with three items; on a scale from 1 – 6: how fun is programming, how fun are robots; and how good are you with robots?
First, independently of condition, we examined children’s STEM-gender stereotypes. Higher numbers up to 4 indicate stereotypes that boys are better at each subject; lower numbers down to 1 indicate stereotypes that girls are better. The dotted line represents chance or neutral responses. You can see that both girls and boys reported significant stereotypes that boys were better than girls at programming and robots, but there were no significant effects for science and math at this age.
Now on to motivation. Here are the results of children’s programming enjoyment, where 6 indicates high and positive motivation and 1 represents low motivation. Preliminary analyses showed that the two control conditions (game and baseline) did not differ, so we collapsed these two conditions for analyses. The white bars show the results of both control groups, and the gray bars show the results of the robot group. There was a main effect of condition, with higher motivation in the robot group, which is the gray bars, and a main effect of gender, with boys overall higher than girls. [The interaction between condition and gender was not significant.] There was no difference for boys, who were already very high in motivation, but this 20-minute experience significantly increased girls’ technology-related motivation. Furthermore, after the robot treatment, girls became statistically indistinguishable from boys.
And here are the results for how much children enjoy robots, and the overall pattern is very similar. Again, there was a main effect of condition, with children in the robot condition reporting higher enjoyment, and a main effect of gender, with boys reporting higher enjoyment. [And again, the interaction is not significant.] Once again, the gender difference in the control conditions was significant, but not in the robot condition. However, in this case there was only a trend for condition to increase girls’ enjoyment.
And here are the results for children’s self-efficacy with robots, and again the overall pattern is very similar, with main effects of both condition and gender. And just as for programming enjoyment, we found that condition significantly increased girls’ self-efficacy, and the robot treatment eliminated the gender gap in self-efficacy.
These results suggest that, even though children held strong stereotypes associating boys with ability at robots and programming, providing positive experiences with technology to girls can increase their technology-related motivation. Well-designed experiences can increase their positive attitudes about programming and boost their self-confidence to equal that of boys. And this matters, because enjoyment of a field and a sense of self-efficacy can lead children to seek out further opportunities in that area, and build a well-developed sense of intrinsic interest in that area.
A sneak peak of our new characters and games. In addition to the main puzzles we have a number of games coming up that teach children how to create their own puzzles by playing engaging card and board games.