Skip to main content



SearchKids

Introduction

Become a Kidsteam kid

If you are interested in getting your child involved in kidsteam, please contact Dr. Allison Druin at allisond@umd.edu.

Our intergenerational design team is partnering with elementary school children (ages 5-10) and teachers from Yorktown Elementary School in Bowie, Maryland to develop a children’s digital library environment containing rich multimedia resources. We are developing visual interfaces that support young children in querying, browsing, and organizing multimedia information. Throughout our design experience, we hope to understand how new technologies can affect learning and collaboration in school environments.

Our digital collection of videos, text and images is being developed with the generous support of the National Science Foundation, the Discovery Channel, the Patuxent Wildlife Research Center, and the Baltimore Learning Community/Maryland Electronic Learning Community project (BLC/MELC), and a Technology Innovation Challenge Grant Project funded by the U.S. Department of Education.

Goals

  • Understand the unique needs of children in learning environments
  • Develop new visualization techniques for use in children’s digital libraries
  • Develop collaboration tools so children can search for information together
  • Create new presentation tools for digital libraries users
  • Understand the impact a digital library environment can have on children as researchers and learners
  • Create support materials and inservice training opportunities for teachers to develop engaging activities that make use of the digital library
  • Formulate and evaluate new research methods for developing children’s digital libraries technologies

Background

The majority of today’s digital libraries were not created for young children.  To access these libraries, children commonly need to negotiate complex text keyword searches. Few libraries offer visual browsing of information in addition to querying, so that young children can explore freely.  Children differ from adults in the way they choose to organize and present the information they have found.  We created SearchKids to support children's organization and presentation styles as well as to give them an opportunity to work together with other children.

Project

This human-centered research is concerned with four main areas of implementation: querying, browsing, organization and presentation.

Visual Query Specification

Today’s most common query interfaces rely on boolean text keyword searches.  Boolean searches use the words "and" or "or" to specify what the user wants to find.  For example, in traditional boolean searches to find webpages that feature both cats and dogs the user would type in the key words "cats and dogs".  To find cat webpages or dog webpages the user would type in the keywords "cats or dogs".  Boolean searches are  difficult for young children because they need  concrete representations of ideas. Typing is also difficult for young children (especially if they do not know how to read yet).  Therefore, all the searches in SearchKids can be performed by clicking on visual icons.

Initial Screen

This is where the search begins. Children can look for information about animals  by searching through taxonomical categories, looking in the appropriate  zoo house or searching for them geographically on the globe.

Search

In the search section children can find information about animals by choosing categories.  At the broadest level the categories are: what they eat, where they live, how they move, and how they are taxonomically categorized.  The results of the query are shown in the red blob next to the search helpers.  Children can zoom into this area to see the results more clearly.

Zoo

This is the zoo.  Here children can find animals by looking in different animal houses such as the reptile house the bird house and the farm.

World

Children can spin the globe in this area.  They can click on the area of the world they are interested in and SearchKids will zoom into that area.  Several animals native to the chosen region will be displayed over an image of that region.

Once children have found a specific animal they want to learn more about, they can click on that animal.  Clicking on a specific animal enables children to see more pictures, see other children's picture, watch video, hear audio and read text about their chosen animal.

Browsing

Children love the process of exploration.  In the course of our research we found that children want a way to visually explore information.  Therefore, we decided to make a library for children that enables them to visually explore information as part of a zoomable user interface (ZUI).  ZUIs are ideal for children because all the information is presented together and the user can zoom in and out of areas to get specific or general information as needed.  Our experience developing and testing systems that incorporate real-time continuous zooming as the primary navigation mechanism, has taught us that many users (adults and children alike) prefer zooming to non-zooming systems. We have found with the drawing/storytelling software KidPad, that children are thoroughly engaged when zooming through information. We theorize that this may be due in part to the visual context that is provided by zooming through space.  The act of zooming from one object to the next, makes visually explicit where users are going and where they have been. In traditional systems that don’t use zooming to navigate, different objects that are semantically related are linked visually by jumping from one object to the next (e.g., links on the Web).  One child said that using non-zooming interfaces is like "closing your eyes and when you open them you're in a new place."  The child then added, "Zooming lets you keep your eyes open.”

Information Organization and Presentation

Children enjoy the social experience of learning through collaboration, not just with distant collaborators, but with children sitting right next to them.  We call software that enables children to work together on the same computer at the same time "single display groupware" (SDG).  We consider co-located collaboration tools a priority in supporting children as researchers. Therefore, SearchKids enables two children with two mice to work simultaneously on the same computer

Design Process

Participants

We are developing SearchKids with two groups of children.  The first group includes our regular child design partners who have been involved in the design process from the initial brainstorming session.  We also work with about 100 seven to nine-year-old students at Yorktown Elementary School as "informants.”  Informants play an important part in the design process at various stages, but unlike the design partners they do not contribute continuously to the research project.

Working with the students at Yorktown as informants made it possible for us to observe a large diverse population of children, while only minimally disrupting their busy school day.  We learned how to modify our digital library to make it more useable for children with different backgrounds and learning styles. This was the first occasion where we, as a team, worked with children in both the informant and design partner capacities as part of a large-scale research study.  This study also marks the first time we worked with teachers as design partners over an extended period of time.

Child and Teacher Design Partners

We began our research with a "low-tech prototyping” session.  Before the teachers or children looked at any other digital libraries, we brainstormed without consideration to previous work. We felt that this would promote a feeling that "anything is possible."  The team split into three groups,  each consisting of two or three children, one teacher, and one or two university researchers. Each group was asked to sketch a digital library of the future that contained all of the information they ever wanted know about animals.  Each group used low-tech prototyping materials such as paper, clay, glue and string. From this brainstorming session, three low-tech prototypes were developed that generated ideas for digital libraries (e.g., the interface did not have to look like a book, the interface should be specific to the content area—in our case animals, the interface should use graphical representations as queries).

Following this experience, the team used and critiqued various children’s digital libraries that contained information about animals.  Children and adults took notes while other children explored the software.  Everyone wrote notes about what they liked and didn't like about the software.  Two main criticisms emerged from these notes.

  1. There was no motivation to search.  We wanted a purpose for searching and we wanted something to do with the information once it was found.
  2. The animated characters that told the children how to search were annoying.

At the beginning of our "sticky note session," the adults on the team were  baffled by numerous sticky notes with comments such as, "It doesn’t do anything", "I was bored at looking" and "Nothing happens" (Researcher notes, November 1999).  With these notes children were explaining that it was not enough to search for things.  They wanted to use the information they found to make something.  The children observed that the one application that allowed them to "Do something" with animal images was annoying because an animated character kept telling them what to do.  Later the adults compared their own notes, and found they had made the same observations as the children.

The team then spent a few  design sessions brainstorming and drawing in their journals.  From this experience, critical ideas crystallized.  One idea was the metaphor of searching by going on a journey.  An eight-year old design partner explained; "Finding things is like going on a trip, so you should go with friends” (Researcher notes, December 1999).  She thought that these friends shouldn’t be "pushy” like the animated characters we saw.  Instead, characters should give kids a reason for wanting to find things.

Another idea that emerged was that the notion of looking for animal parts that represented search categories. Therefore, instead of a typing the question "What animals move by using their legs?” children said that they preferred to click on a picture of animal legs.

Other ideas had to do with the questions that the children wanted answered about animals. These questions were: (1) what do they eat, (2) how do they move, (3) where do they live, (4) what animal family are they part of.

The teachers we worked with were also interested in different search areas.  One teacher pointed out that in the early grades children learn about animals in different contexts.  They might learn about "pets at home” or "farm animals.” This teacher also said that older children learn about animals while studying specific geographic areas.  For example, children might learn about kangaroos when studying Australia.

The team decided that we needed several ways to search for animals, so children at different grade levels could take advantage of the library. The teachers pointed out that there are big differences between the needs of second and third grade teachers even though their students are only one year apart in age.

Putting Our Ideas to Work

Next we began designing an "interactive sketch. "To create this sketch we used KidPad, a zoomable authoring tool for children.  Our team artist began sketching with KidPad, and as she sketched the team refined its ideas. We clarified the role of the search kids (the characters that help children on their journey to find information).  Search kids do not tell children what to do.  Rather, they remind the children of the query they are forming.  The search kids "hold" the search criteria a child wanted to use. If a child searches for mammals a search kid will hold an icon with pictures of mammals on it.   

The notion of "doing something” with the search results also began to take shape. Since the team was already helping to develop KidPad, it made sense to link the two programs so children could use the animal media they found to create stories or art.  Ultimately, this meant building our first interactive prototype on top of the KidPad architecture. 

At this time we also decided that we needed three different areas to look for animals. These three areas are a zoo (with a farm house, a pet house, a bird house, and more), the globe, and the search area.

As the first functional prototype was being developed by our technical team, we continued to refine the SearchKids interface by using paper representations of the search criteria and people to represent the kids. We also created, in consultation with our team biologist, a Microsoft Access database with metadata on animal images contributed by our content partners.  When our first interactive prototype was sophisticated enough to be usable by people outside of our design team, it was brought back into the school to be used with our informants. Fifty children who had not previously explored the paper prototype were asked to offer feedback on the interactive prototype.

Evaluation of our Search Hierarchy

Soon after these design sessions, three adult members of our team began working with 50 children at Yorktown Elementary School (YES). In working with these children, we realized how little we know about how children search for animals and how complex their queries could become. Therefore, we conducted an empirical study with 104 children at the school to see how they searched using the categories our lab team developed.  We created a set of hierarchically-nested envelopes based on the four categories our child design partners developed (habitat, food, movement, and animal taxonomy). The children in the school were asked to search for pictures of specific animals by looking in the envelopes. 

By observing the student’s behavior in this situation, we learned that boys and girls search differently. Generally, the boys dumped all the envelopes on the floor (with little thought of putting things back) and searched for the animal they were supposed to find. Girls tended to be more careful in their search style, but they often appeared to be more interested in browsing the pictures as opposed to finding the animal for which they were searching. These observations made it clear that the application should support both structured searching and browsing as equally valid methods of accessing information.

We also learned that children were able to efficiently and accurately find the animal or group of animals they were looking for, using the hierarchy we developed.  

Evaluation of Collaboration

After our evaluation of the search hierarchy we continued to refine the SearchKids interface for another year.  Then in the spring of 2001, we began a study to see how children collaborate using SearchKids.  The study was done with 98 second and third grade students at YES.  The children were put into same gender pairs and allowed to explore SearchKids.  They were then asked to locate as many animals as they could from a list.  The children were videotaped and the first five and last five minutes of their conversations were coded.  The conversations were coded to find out what the children were discussing.  

There were two modes of collaboration in this study. In the independent collaboration mode, if one child clicked on an icon it was enough to interact with or select that icon. In the confirmation collaboration mode, both users had to click on an icon in order to interact with it.

We found that there was no significant difference in how the independent and confirmation collaboration pairs performed.  However, there was a difference in their styles of collaboration.  The independent users tended to talk about the task they were completing.  For example they might say something like "Let's find the butterfly next."  Whereas, the confirmation collaboration users talked a lot about how to navigate the software.  A confirmation collaboration user might be likely to say "Click here."

We learned from this study that different types of collaboration can be used to teach children different things.  For example, if a teacher wanted children to learn about a specific content area, she might ask the children to use confirmation collaboration.  If she wanted children to learn to work better together, she might ask them to use the independent collaboration condition.  We decided to make both types of collaboration available with SearchKids in order to support different kinds of learning. 

Current Study

We are currently working with 133 second and third grade children at YES to see how SearchKids compares to traditional text keyword interfaces.  For this study we worked with one child at a time.  Each child played with a version of SearchKids as well as a traditional style interface much like Yahooligans.  Both interfaces accessed the same database.

p1 
 Traditional text keyword interface

During the study, a researcher read to the children from a list of animals and it was their job to find as many of the animals on the list as they could in 15 minutes.  The children had two minutes to find each animal.  If they could not find the animal in two minutes the researcher would read the next item on the list and the child would look for that item.  The children were given a demonstration of how to use the software immediately before they began the test.  After the children used both of the conditions, they answered a few opinion questions about each interface.

The preliminary results from this study show that SearchKids and the traditional text interface are equally effective in helping children access information about animals.  We are also finding that children prefer to use SearchKids over the text based interface.

Potential Applications

We are expanding our research on digital libraries by developing the International Children's Digital Library in conjunction with Archives.org and with additional support from the National Science Foundation, the Library of Congress, and the Markle Foundation.  For more information please visit International Children's Digital Library website.  

Recently we have developed book readers to be used with our International Children's Digital Library software.  To see videos of these book readers, follow these links: comic strip reader, spiral reader.

Papers

Videos

 

 

 


Tech Reports
Video Reports
Annual Symposium

News
Seminars + Events
Calendar
HCIL Seminar Series
Annual Symposium
Events Archives
Awards
Job Openings
For the Press
HCIL Overview
Collaborators + Sponsors
Collaborating Groups + People
Academic Visitors
Industrial Visitors
Active Sponsors
Become a Sponsor
Our Lighter Side
Join our Mailing List
Contact Us
Visit Us
Faculty/ Staff/ Admin
Students
Ph.D. Alumni
Past Members
Research Areas
Communities
Design Process
Digital Libraries
Education
Physical Devices
Public Access
Visualization
Research Histories
Faculty Listed by Research
Project Highlights
Project Screenshots
Online Tech Reports
Video Reports
Books
Products
Presentations
Studying HCI
Graduate Studies in HCI
Visiting Scholars
Class Websites