Through the Glass Wall

The Somerville Community Computing Center is softly carpeted and brightly lit. Signs on the wall ask, "What is the Internet anyway?" There are diagrams of modems and servers and monitors. The computers are named - Calvin and Hobbes, Mantle and DiMaggio... The shell of an old Macintosh has been converted into a donation box; you slip bills through the disk drive. Over the deep electric hum of the room a group of experts are clustered around a computer screen, thoughtfully debating:

"Send that one with the blue nose over the top bridge."

"Oh, no, he can't! He got blown back."

"So maybe the ones with the pink shoes can go over the top bridge."

"Good, she made it! Now, how about this one with the red nose and pink shoes?"

"Which bridge?"

The problem? How can this group of students get their Zoombinis - little creatures with different feet, noses, eyes and hair - through a series of puzzles to safety in Zoombiniville?

To find the answers, they must use newly learned skills in set theory or informed trial and error. In this puzzle, the challenge is to get the Zoombinis across some very particular bridges that only let those with certain characteristics cross safely.

For Jan Mokros, Andee Rubin, Megan Murray, and the researchers on the project Through the Glass Wall, there are larger puzzles to solve. The burgeoning field of "edutainment" software is producing numerous computer games, some of which purport to promote math learning for kids. But little research has been done to determine if and how these games can support mathematical learning outside of the classroom.

Games for Girls?

Computer games can be an important entree into the field of technology for some children. Kids who play computer games at home tend to think that they know a lot about computers and are less daunted by using them in school. Unfortunately, the majority of children in this category are boys. Boys buy and play a huge percentage of the computer games on the market. Although girls do play computer games, many games have male characters, use boys' sports as a metaphor, or include some sort of violence; these games are not engaging for girls. According to Andee Rubin, "...for girls, the computer's screen seems to be a kind of glass wall. They are allowed to glimpse its worlds from a distance, but are not invited inside."

The Glass Wall project is investigating the interaction of computer games, mathematics learning, and gender by examining a few different questions: How do children learn math from computer games? What characteristics of games and game-playing contexts contribute to or distract from the possibility of learning? What patterns are there in how girls and boys play and learn from computer games? How could games be designed that would be at least as attractive to girls as to boys? To find the answers, the researchers went to the store shelves to find games that were strong in math and appealing to both genders.

They discovered that in recent years marketing executives have become aware of girls as an untapped market in the computer game industry. The result has been an effort to create games especially for girls. Unfortunately, many of these games are based on stereotypically female interests, such as shopping, make-up, and dating. A new piece of Barbie-based software, in which girls help Barbie design outfits, sold over half a million copies in its first several months on the market. Another piece of "pink" software even comes with a tube of lipstick. Not surprisingly, the survey turned up no software aimed specifically at girls that also required significant mathematical thinking. Given this, the project focused on finding mathematically challenging games that were as "gender-neutral" as possible. This meant avoiding games with a lot of violence, only male main characters, and elements of time pressure and hand-eye coordination (as in video games), all of which seem to predispose a game to male users.

The Lincoln School

The project made its first observations last winter and spring at an after-school program for third, fourth, and fifth graders at the Lincoln School in Brookline. The researchers watched kids play a variety of games (Operation Neptune, Escape from Dimension Q, Math Blaster and a few others) and then conducted an in-depth observation of kids playing Logical Journey of the Zoombinis (designed at TERC by Chris Hancock and Scot Osterweil) over a period of several weeks. They wanted to determine just which characteristics of games and game contexts would be important to study. Some of the qualities they noticed were the structure of the game, the role of narrative form and challenge of the game, the ability to control levels of difficulty, and the roles of competition and cooperation that grew up around game play.

They noted one contrast between two ways that kids interacted with the Zoombinis game: some kids focused more on the puzzles, while others were more involved with the characters and narrative.

There were also contrasts in the role of cooperative game-playing. Zoombinis seemed to support collaborative play, because the puzzles were complex problems that involved strategic thinking and experimentation. Many students recognized that it was good to have a second mind working on the puzzle with them. Knowledge and strategy (how to send the Zoombinis across the right bridge) were shared within the group, with some students knowing particular strategies. In contrast, a game like Math Blaster was not as conducive to talk or to partner play. Because of its reliance on quick recall of facts and hand-eye coordination (giving quick solutions to changing math problems enables the player to shoot targets or navigate mazes), there was less collaboration and discussion.

Another important focus that grew out of the Lincoln school work was on "engagement" and "persistence." Students stayed actively engaged with Zoombinis over a several week period. Other games had less longevity. If the length of time that a child will engage with a game is directly related to the amount of learning that takes place, it is important to examine what aspects of computer games encourage this kind of persistence and engagement. Researchers began to study this issue in relation to gender differences and ways to assess learning takes place at a computer. (For more information on this study and its findings, see What Kind of Educational Computer Games Would Girls Like?.)

Summer in Somerville

The next phase of Though the Glass Wall was conducted this summer at the Somerville Community Computing Center (SCCC). The group presented eight computer math games to a diverse group of upper elementary students. These kids enthusiastically came in once a week to play games and give their opinions to Glass Wall researchers. During the first two sessions, each of the eight games was introduced. The students then chose which games they wanted to play. Whenever they finished a session with a game, they took out their journals, gave the game a ranking from 1 (highest) to 5 (lowest), and explained their score. Researchers also kept track of how long each student played each game.

After six weeks of observation and game playing, the students were each asked to rank the games from their favorite to least favorite. They also used screen dumps and magic markers to create poster ads for their favorite game. These rankings and posters became the centerpieces of group discussions and individual interviews. Researchers asked the students how they made decisions about their rankings. Which game was hardest? Which were appropriately challenging? Which game was easiest? Did any of these games have a really good story? Were there any particularly good or bad characters? Why was this game your favorite? What would you say to make your friend interested in this game? Did you learn anything from this game? The responses to these questions and the recorded discussions will help the researchers to further evaluate how and why students chose, liked, and persisted with particular computer games.

The researchers then presented two ceiling-high graphs showing how many times each student played each game and what rating each game received; one for the boys' data and one for the girls'. They also presented the results of students' rankings of the game from 1 (favorite) to 8 (least favorite) and the total number of minutes that each game was played. The visual presentation of these data helped the group think about which games were the overall favorites, which were the girls' favorites, and which were the boys' favorites. The students observed and wrestled with the disparity between the real and expected outcomes, the difference between average high scores and game popularity, and discrepancies between girls' and boys' choices and ratings.

For example, in their journals the boys only rated games as 1's and 2's, very high. However, the girls rated games between 1 and 5, and even added half numbers, like 3.5, when whole numbers seemed inaccurate. Both the students and researchers had ideas about why this might be (Megan and Andee are the researchers):

Megan: ...Someone looked at [the graphs] at work... and said well maybe the girls are pickier about games, they just are more particular about which games they like and don't like. But then somebody else said well maybe you [researchers] brought lots of boy kinds of games, like games that boys like...
John: Boys rated more on the bright side of games, girls rated more on the on the negative side of the games....
[later]
Andee: Tianli, you think boys...
Tianli: Just didn't think about it.
Andee: So you think boys just rated them quickly?...
Zak: ...If I ever played Math Workshop, then bang!, I'd be down at 5... The boys played the games they would like in first and second place but the rest they don't play.
Andee: They just didn't bother?
Zak: Yes.

The above discussion suggests many different ways to view just one interesting facet of the data. Perhaps girls are more particular about being exact in their ratings, while the boys are quicker to rate a game positively. Perhaps the games themselves were all male-oriented, so the boys loved them, while the girls had mixed impressions. Maybe the boys had enough previous experience with games that they knew which ones they liked and only played those. Or maybe the boys just "look on the bright side of games."

The students at the Lincoln School and the SCCC have provided only a small sample for the Glass Wall project. Even so, says Megan, the findings seem to coincide with other current research into gender, technology, and learning. Their studies have also given Glass Wall researchers a vocabulary for discussing the intricate relationships between learning, play, and gender, and are pointing the way towards the large questions yet to be answered. What characteristics of computer games and social environment encourage persistence, engagement, and learning for both girls and boys? How do we assess mathematics learning that takes place at a computer terminal and in a home environment? How can these findings translate into the development of computer games that might invite girls through the glass wall of the computer screen, and who will be responsible for the development of such games? These are the questions the Glass Wall team will be investigating as they continue study the relationships among computer games, mathematical learning, and gender.

Through the Glass Wall: Computer Games for Mathematical Empowerment. PI, Jan Mokros. PD, Andee Rubin.