Interestingly, the development and evolution of the product itself mirrors the development and evolution of the teaching process within the product. Headsprout uses a recombinant model (Johnson & Layng, 1992; Lee & Anderson, 2001) in which component letters and sounds are taught and then combined to form a composite, which then becomes a component of a word. Words then become components of sentences and sentences of stories. This unique model is also applied to the instructional framework of Headsprout Reading Basics' episode sequence. Instructional techniques (components) are recombined (into composites) to form the pedagogical foundation of each episode. This entire process is analogous to chemistry: atoms of various elements (always themselves the same) come together in various combinations to yield different molecules. This model also extends to the engineering process. Because the modular computer code underlying component instructional techniques can be reused with different content or with different graphics, Headsprout achieves great efficiencies in product development. The modular structure also enables Headsprout to promptly take advantage of the observations and data gleaned from user testing; if necessary, changes can be incorporated across several episodes with relatively little effort.
In all of these cases, some of the instructional components are no longer used over the course of the program because learners "learn how to learn." In other words, they become more efficient in acquiring new knowledge and can learn new things in different, more sophisticated ways. This is yet another instance of Headsprout's ability to evolve along with the learner. These early skills will, of course, continue to be practiced and applied as implicit aspects of other activities. For example, practice with letter/sound combinations is reinforced as learners move on to deciphering whole words and sentences.
One of the crucial ingredients in Headsprout's success-perhaps transcending the pedagogical, developmental, and technological innovations described above-is delivery over the Internet. Although the Internet is Headsprout's distribution system and not its defining technology, it is central to the effectiveness of the product. The Internet model enables Headsprout to make changes and improvements to the product at any time, without forcing expensive and/or inconvenient updates on consumers. The program can adapt to every learner's behavior because all learning data is up-loaded to Headsprout servers for analysis. More importantly, the Internet makes every non-reading, Internet-connected child in the world (i.e., millions of kids) a potential Headsprout beneficiary.
The end-user experience is uniformly smooth, which is especially crucial when the user is a young child. Headsprout's code is engineered for streaming, allowing Headsprout Reading Basics to run seamlessly, even over 56K modems. Learners spend very little time waiting for program segments to begin because successive segments are downloading in the background as they work. In the event that they do have to wait, any one of thirty interactive screens appears to help pass the time; they can scroll or click to make entertaining changes in the graphics. Also important are the Macromedia Flash animations that drive the entire program and entertain and motivate children as they learn. There are dozens of cute characters featured in Headsprout's episodes; all of them have been created and animated in-house, drawing upon Headsprout staffers' graphical, vocal and instrumental talents to enliven them with motion and sound. In marrying "state of the science" learning theory, Hollywood entertainment, and Silicon Valley technology, Headsprout has created an Internet technology that seamlessly provides effective reading instruction.
The scientific basis for our claims comes first from a thorough grounding in measurement. We measure all opportunities to interact with our program and all learner interactions on a real-time basis for all learners via the Internet. For each individual learner, we know:
Because Headsprout Reading is broken down into tactics and strategies targeting specific skills, our data inform us about the effectiveness of each instructional strategy across all children and for every individual learner. The data are used to modify the instructional sequence to meet the individual needs of our learners, as well as to continually shape and redesign the overall efficacy of the program. Headsprout knows immediately what works and what doesn't, and can continually make revisions and test them against the established procedure, selecting the procedure that produces the best result.
All learner performance data are automatically collected and entered into a database allowing us to look at individual performance as well conglomerates of these data. For example, we analyze the number and rate of opportunities to interact, responses and corrects per episode, response fluency, episode and segment durations, and errors patterns to evaluate learner outcomes. We measure accuracy on literal comprehension tasks and the recombination of taught skills into news skills to obtain indicators of more generative reading skills. Each activity or teaching routine in the program has a designated mastery criterion, which, with exception of the practice exercises, is always 100% accuracy over a consecutive sequence of trials. Failure to reach mastery results in the learner being provided a corrective teaching sequence specific to the particular error pattern. The student is then returned to the mastery sequence, and must demonstrate mastery prior to exiting the activity. Learners are required to meet more challenging mastery criteria as they progress through the program. The program is designed to maintain a high level of correct performance as these criteria progressively change during the program.
An additional source of data comes from the videos made of each learner in our laboratory interacting with the program. By way of a split screen, these tapes show both the learner and the display screen as seen by the learner. These allow us to measure and evaluate both learner affect and specifics regarding eye gaze, attention, reactions to the cartoon sequences.
Figure 1 depicts the growth of fundamental reading skills as a function of progress through Headsprout Reading Basics. Headsprout administers computer-based pretests and posttests with our on-site user test children for each episode presented. These pretests and posttests assess the specific skills targeted for each episode. We have consistently found steady increases in letter/sound reading ability where previously there was little or none.
Figure 1. Outcomes as a function of lessons completed.
Figure 2 depicts the ratio of correct to total interactions for over 1000 beta (field-testing) participants. The relatively high correct percentages demonstrate the error reduction and cumulative shaping nature of our program--where new skills are built upon previous skills and there is not an abrupt learning curve for our learners. Correct responding criteria are maintained even as the program requirements become more demanding. The high percentages also reflect our rigorous accuracy criterion that is met on each segment before a learner completes an episode. Each episode reinforces some prior learning and introduces new sounds, skills, or strategies necessary for ultimate reading success.
Figure 2. Average percent correct per Headsprout Reading Basics episode. Data are for over 1000 learners.
Figure 3 shows the amount of time it takes our learners, on average, to complete an episode and meet the instructional criteria for exiting segments and episodes. These data assist in arranging the logical flow of instructional sequences while not taxing our learner's ability to be motivated and successful. On average, Episodes take 20 minutes to complete and range from just under 10 minutes to slightly over 30 minutes. The data were collected from over 1000 beta test learners.
Figure 3. Average duration per Headsprout Reading Basics episode. Data are for over 1000 learners.
Opportunities, responses and corrects are also monitored. Figure 4 displays the ratios between opportunities (each time an learner interaction is available), responses (actual learner interactions, where the child is making a response) and corrects allow us to evaluate the effectiveness of episodes and segments on a response-by-response basis. It is this fine-grained analysis of individual learner performance (from which this graph is complied) that gives us confidence in replicating these outcomes for over 1000 learners.
Figure 4. Mean number of response opportunities (bars), total responses (filled circle), and correct responses (open circles) per Headsprout Reading Basics episode. Data are for over 1000 learners.
| Aging Gracefully | Autism | Behavioral Safety |
Book Reviews |
Commentaries |
Education |
| Everyday Life |
Parenting |
Organizational Behavior Management | Pets & Animals |
Verbal Behavior | Virtual Community |
Copyright
©1997-2009 by the Cambridge Center for Behavioral Studies.
All rights reserved.
Feedback or questions about the
Cambridge Center for
Behavioral Studies or our website?
Contact our webmaster or
our Executive Director, Dr. Philip N. Chase.
The
Cambridge Center for Behavioral Studies Publication Office
is located at the following address:
336 Baker Avenue
Concord, Massachusetts U.S.A. 01742-2107
Telephone: (978)
369-CCBS (2227)
Facsimile: (978) 369-8584
Visit other sites through the Behavior Analysis Webring:
[
Previous 5 Sites
|
Previous
|
Next
|
Next 5 Sites
|
Random Site
|
List Sites ]