|a.||a few symbols were incomprehensible to many of the children => find new symbols.|
|b.||most younger children did not use alphanumeric representations => possible tailoring of the game to grade level.|
|c.||large numbers of symbol choices were confusing and unnecessary, => symbol libraries should be moderate sized and possibly tailored by grade level.|
|d.||during the search for a symbol replacement children often got caught up in examining images and forgot the word that they were trying to replace. => keep message to be coded and choices from the symbol library visible on the same screen during coding.|
|e.||organization of symbols by category was very confusing to younger children. Organization by words was helpful to them.|
|f.||older children were able to, and very much enjoyed, making up their own messages to be coded. After having become familiar with the symbols in the symbol library, they readily created messages for which appropriate symbols were available.|
|g.||younger children required text to speech capability but had no problem playing the games otherwise.|
One of the most important findings from the evaluation was that careful instruction and/or other scaffolding is required in order for the children to understand and play the games. Children were initially clearly confused with the objective and procedure for many of the games. It often took one or two rounds of tentative play before they became comfortable and involved in the games. Often the children who seemed most confused initially were the ones who became most intensely involved later. The difficulty that children had grasping the game may have had something to do with our initial instructions. However, all children needed at least one walk through of each game and the number of rounds required decreased with age. We believe that the structure of the activity was unfamiliar to them and that the degree of confusion was related to their level of symbolic processing development. The observation that, after a small number of sample sessions, all of the children could play most of the games independently (assuming textual material was read to non-readers) and enjoyed them very much suggests a developmental readiness across the entire age range.
These findings have important implications for how the computer software should be designed. It is clear that the provision of instruction within the context of the games will be a critical factor that determines the success of the program. Since, the games are intended to be "self-teaching" it was important to find ways to describe game play to the children. We found that some forms of instruction are much more effective than others. In particular, we found that walking through an example of game play was far better than verbal directions for how to play. Clearly it is important to provide demo walk throughs of game play and scaffolding so that games are presented in a sequence of gradually increasing complexity and abstraction.
Although the Rebus game was enjoyed by all children regardless of age or level of development, we found that developmental readiness was an important determinant of the children's level of interest for some of the games. Younger children were particularly fascinated with a mathematics coding game that resembled the game of Battleship. Older children, who tend to have more exposure to symbolic language, were more readily able to utilize the coding models and had less trouble with more abstract games and games that required greater sophistication of geometric reasoning. Younger children needed more trials and more systematic instruction. However, once they grasped the coding model used in a game, they were easily able to use what they learned and apply it to new situations. In fact, we were surprised that even children who did not yet read could play the games if an adult read textual messages.
We found that the KidCode games improve children's skills with particular representations used commonly in mathematics and they seem to improve the children's competence with symbolic processing. While all children who played the games in sequence were able to understand and play each game independently after being guided through one or two examples, children had a great deal of trouble when the games were not presented in sequence and games with higher levels of coding abstraction were played without the benefit of experience with earlier games.
It was our further aim that the games develop children's conscious understanding of symbolic representation and confidence in their ability to translate any representation to retrieve it's underlying meaning. We believe that experience with various kinds of representations will give children a foundation so that they are not intimidated when presented with a complex looking mathematical