SCHOOL AGE CHILDREN'S COGNITION IDENTIFICATION BY

MINING INTEGRATED COMPUTER GAMES DATA

Rob Whent

, Dragana Martinovic

, C. I. Ezeife

3∗

, Sabbir Ahmed

, Yanal Alahmad

and

Tamanna Mumu

OTEP Inc, 13300 Tecumseh Road East, Suite 366, ON N8N 4R8, Tecumseh, Ontario, Canada

Faculty of Education, University of Windsor, Windsor, Ontario, N9B 3P4, Canada

School of Computer Science, University of Windsor, Windsor, N9B 3P4, Ontario, Canada

Keywords:

Cognitive Skill Mapping, Computer Games, Database Integration, Database Mining.

Abstract:

National statistics conﬁrm that nowadays about 20% of school children have some type of mental health is-

sue, about 70% of adult mental health disorders originated in adolescence, while about 40% have unidentiﬁed

learning differences that affect their learning abilities. Starting early enough with proper screening of a child’s

cognitive skills is critical for improved learning and mental wellbeing. Unfortunately, assessments and treat-

ment can be costly, elusive or conﬂicting. This paper describes an approach to identifying child’s cognitive

skill level that is adopted by an online product called “Think2Learn” developed by OTEP Inc. (Online Train-

ing & Education Portal). OTEP uses ubiquitousness of the Internet and attractive features of online computer

games to give parents automated opportunity to screen and follow their children’s cognitive development.

OTEP presently uses a collection of approximately 100 video games for children to play and while doing so,

it records their score to continuously assess and monitor their cognitive strengths and weaknesses. The Web-

based tool for identifying cognitive skill level is developed as an integration or data warehouse of a number

of relevant data sources such as the cognitive skills categories data, games data, player inventory data and so

on. The integrated data are continuously mined, analyzed and queried for proper and quick assessment or

recommendations.

1 INTRODUCTION

Currently available digital technologies provide for

multiple representations (e.g., visual, text, and sound)

that draw youth as attractive, artistic and fast-paced.

There is a bodyof research that pointsto unique learn-

ing habits of young people. Among else, they pre-

fer short visual explanations, to receive information

quickly and process it rapidly, prefer multi-tasking

and non-linear access to information, have a low tol-

erance for lectures, prefer active rather than passive

learning, and are kinaesthetic, experiential, hands-on

learners who must be engaged with ﬁrst-person learn-

ing, games, simulations, and role-playing((Junco and

Mastrodicasa, 2007); (Oblinger and Oblinger, 2005);

(Tapscott, 2009)). The youth nowadays also rely

∗

This research was supported by the Natural Science and

Engineering Research Council (NSERC) of Canada under

an Operating grant (OGP-0194134) and Engage program.

heavily on communication technologies to access in-

formation and to carry out social and professional

interactions ((Veen and Vrakking, 2006); (Pletka,

2007)). At the same time, playing computerand video

games are lately being recognized as valid cognitive

activities as they affect a player’s capability to self-

regulate, make right decisions, and problem-solve.

Such environments are engaging but put a strain on

the cognitive load and attention span of the user. It

is notable that some parents deliberately avoid hav-

ing computers at home, while some restrict access for

their children out of fear that they will use comput-

ers to play games rather than for educational purposes

(Dance, 2003). How should educators respond? Is

there a way to convince parents that their children

may learn while they are playing computer games?,

that children may enhance their cognitive capabilities

while having fun on a computer?. Cognition refers to

the mental processes involved in gaining knowledge

and comprehension, including thinking, knowing, re-

495

Whent R., Martinovic D., Ezeife C., Ahmed S., Alahmad Y. and Mumu T..

SCHOOL AGE CHILDREN’S COGNITION IDENTIFICATION BY MINING INTEGRATED COMPUTER GAMES DATA.

DOI: 10.5220/0003917204950505

In Proceedings of the 4th International Conference on Computer Supported Education (SGoCSL-2012), pages 495-505

ISBN: 978-989-8565-07-5

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

membering, judging and problem-solving. These

are higher-level functions of the brain that encom-

pass language, imagination, perception and planning

skills.

According to the U.S. Surgeon General, national

statistics conﬁrm that nowadays about 20% of school

children have some type of mental health issue;

the Mental Health Commission of Canada (Canada,

2011) states that 70% of adult mental health dis-

orders originated in adolescence, while about 40%

have unidentiﬁed learning differences that affect their

learning abilities (e.g., the acquisition, retention, un-

derstanding, organization of information). Starting

early enough with proper screening of a child’s cog-

nitive skills is critical for improved learning and men-

tal wellbeing. Unfortunately, assessments and treat-

ment can be costly, elusive or conﬂicting. This paper

describes an approach to identifying a child’s cog-

nitive skill level that is adopted by an online prod-

uct called “Think-2-Learn” (Whent, 2012) created by

OTEP Inc. (Online Training & Education Portal).

Cognitive abilities of individuals such as auditory, vi-

sual, sequential, conceptual, speed, and executive de-

termine their learning abilities in the main cognitive

categories such as basic reading, reading comprehen-

sion, math calculation, math reasoning, writing me-

chanics, writing content, oral expression, and listen-

ing comprehension. According to (Crouse, 2010),

student needs to achieve, for example, high score in

auditory domain, high score in conceptual and moder-

ate score on speed of processing in order to have abil-

ity for reading comprehension. Determining a child’s

cognitive abilities in various categories of cognition

and later identifying any improvement in these abil-

ities in a natural environment such as their perfor-

mances when playing relevant computer games is an

approach being adopted in this paper. This paper pro-

poses integrating into a data warehouse a number of

relevant data sources such as video games data, cog-

nitive data, student game play inventory and others, in

other to quicklymine and compare their performances

with the norms for cognitive achievements for differ-

ent children from various age, gender, mental, physi-

cal, ethnic, social and other background they may be-

long to. While this approach does not claim to be a

panacea for alleviating developmental difﬁculties in

children, it should be taken as one attempt that, in

conjunction with other preventive and treatment mea-

sures, may help children to improve their cognitive

skills and strategies. More research is needed to es-

tablish to what extent is this approach helpful and

in which cases; however our preliminary case stud-

ies and data mining method provide the ground work

for future studies.

Section 2 presents the related work including those

in computer gaming and cognition, data warehousing

and mining. Section 3 presents the OTEP solution ap-

proach including the data warehouse integration al-

gorithm, schema and example application. Section

4 discusses performance analysis of how games data

can be effective in identifying cognition while section

5 presents conclusions and future work.

2 RELATED WORK

2.1 Youth and Gaming

Cognitively, young generations are known to have a

shorter attention span and to need immediate answers

(Pedro, 2006). This may be a consequence of the ex-

tensive propagation of video games among youth. In

(Rideout et al., 2005), it is noted that in the U.S., 8-

to10-year-olds spend more than an hour a day play-

ing video games. Since the main features of many

video games are quick reactions and immediate feed-

back, the games may reinforce the children’s inclina-

tion towards fast, focused, and repetitive actions that

result in direct (task-oriented and instant-feedback-

based) and limited (fast, focused, repetitive) learning

in a short time. In (Pedro, 2006), it is concluded that

“nothing seems further from this than the students’

daily school experiences requiring longer attention

spans, engaging students in reﬂective activities, fo-

cusing intensely on one activity at a time, and involv-

ing properly written text” (p.11). The learning habits

of youngsters not only differ from traditional learning

habits, but they also demonstratecertain uniqueweak-

nesses. For example,nowadays studentsare strong vi-

sual learners (Berk, 2010), but usually weak in learn-

ing from text (Vaidhyanathan, 2008). In addition,

while such students prefer fast-paced environments,

their performance sometimes lacks depth and critical

thinking ((Oblinger and Hawkins, 2006); (Rockman

and Associates, 2004)).

2.2 Computer Games and Cognition

There exist conﬂicting reports on the social and cog-

nitive consequences of playing computer and video

games (Martinovic et al., 2011). In the interest of

space, we will focus on studies that relate playing

games to positive effects. There are recorded ben-

eﬁts emerging from playing computer games, such

as improving visual intelligence skills (which are es-

pecially relevant for subject areas in which manipu-

lating images on a screen may be particularly use-

ful, such as science and technology) (Subrahmanyam

CSEDU2012-4thInternationalConferenceonComputerSupportedEducation

496

et al., 2000). Computer-based games may enhance

hand-eye coordination, visual scanning, auditory dis-

crimination, and spatial skills (DeLisi and Wolford,

2002).

Moreover, emphasis on visual information pro-

cessing may be connected to a signiﬁcant increase

in average non-verbal scores in various psychological

tests across all groups tested (see (Greenﬁeld et al.,

1994), (Subrahmanyam et al., 2000)). A comparative

study of children aged 10 to 11 who played two dif-

ferent computergames, onewith strong visual content

and the other text-based, showed that playing the ﬁrst

game improved the children’s spatial performance,

while playing the second did not (Subrahmanyam and

Greenﬁeld, 1994). Repetitive game playing may in-

crease the young children’s working memory (Thorell

et al., 2009); mental rotation accuracy (DeLisi and

Wolford, 2002); and spatial rotation, iconic skills, and

visual attention (Subrahmanyam et al., 2001).

2.3 Computer Games and Learning

Playing the carefully and purposefully designed com-

puter games may positively affect learning among

children of wide range of ages. Early studies on

high school students’ use of educational software at

home showed signiﬁcantly higher scores on com-

puter literacy tests among students who played these

games (Subrahmanyam et al., 2000). Describing the

study done with pre-school children, (Freeman and

Somerindyke,2001), concludedthat the children ben-

eﬁted and thrived when given the opportunity to de-

velop skills through the use of computers. For these

authors, computers “are an ideal vehicle for learn-

ing in a social setting” (p. 206), and as such, have

a role in an active-learning, play-based curriculum.

Because playing computer games involves integration

of “touch, voice, music, video, still images, graphics,

and text” (IBM, 1991), children are faced with the en-

vironment geared to a variety of intelligences (e.g.,

linguistic, logical, spatial, kinaesthetic, musical), this

may particularly inﬂuence development of literacy

skills and ability to problem-solve. In (Gee, 2007), it

was argued that as “problem-solving spaces”, video

games can be excellent tools for learning, as they

provide an environment for recursive practices. It

is through these practices that learners can reach the

point of automaticity in a “non-boring” way. Players

see computer games as systems, rather than as collec-

tions of discrete skills. In addition, exploration and

successive challenges of increasing difﬁculty are pri-

mary principles of computer games. Beck and Wade

(as cited in (Bogost, 2007), p. 240), maintain that

the “videogamegeneration” (those born after 1970) is

uniquely positioned to use meta-cognitive skills ob-

tained through video game playing (e.g., reﬂecting

on immediate situation, analyzing choices and com-

paring odds, ﬁnding the right strategy). In (Bogost,

2007), Gee’s notion of computer games was taken

as “situated . .. and ... embodied learning” further,

stating that “videogames ... offer meaning and ex-

periences of particular worlds and particular relation-

ships” (p. 241, emphasis in original). In other words,

“the higher-order thinking skills still matter, but so

does the ninja” (p. 243). These statements are sup-

ported by the ﬁndings from the neurological research

that identify emotion as one important element of

learning (Zull, 2004). Another element of learning

is practice, and both are present in playing computer

games.

2.4 Examples of Studies Involving

Game Playing

There are several recommendations coming out of

previous work that we took into account in this study.

For example, Ko (2002) analyzed 7-10 year old chil-

dren’s behaviour and cognitive development while

playing a game based on ’if-then’ rules. The author

used statistical methods to ﬁnd if the child used the

game rules or achieved the goal by chance only, and

analyzed children’s game performance in relation to

their individual differences. Ko concluded that the

success in playing depended on the child’s age and

planning skills; also, the children improved in using

the game clues and developed their planning strate-

gies through repeated game playing. Ko’s recommen-

dation is that educational software must be measured

against how children think and learn (e.g., the purpose

of the game, activities during the play, what and how

the child performs, and what concepts are learned).

Computer games of different types affect children

cognitive processes differently. (Pillay, 2002) investi-

gated the impact of two recreational computer games

on 14-16 year old children’s subsequent performance

in a computer based instructional tasks, by employ-

ing both quantitative (i.e., speed and accuracy) and

qualitative (i.e., cognitive strategies) measurements.

The author found that playing recreational computer

games may inﬂuence performance (e.g., time efﬁ-

ciency and correct problem solving) on subsequent

computer-based educational tasks. Pillay concluded

that although recreational and educationalgames may

have different goals, commonality in their structure is

what matters. By playing games of different types,

children develop a repertoire of cognitive schemas

that can help them later in performing learning tasks.

To conclude, (Lieberman et al., 2009) suggested that

SCHOOLAGECHILDREN'SCOGNITIONIDENTIFICATIONBYMININGINTEGRATEDCOMPUTERGAMES

DATA

497

more research is needed to understand cognitive and

other development of children when they play com-

puter games in order to support their healthy emo-

tional, cognitive and learning development. We con-

tinue our report by introducing the OTEP approach

and philosophy.

2.5 Data Warehousing and Mining

Approaches

A data warehouse is a historical, integrated, subject-

oriented database storing data from multiple data

sources in the one data warehouse schema (Han et al.,

2011). Construction of a data warehouse is done

through processes of schema and data integration of

different data sources which involve data cleaning

(Ezeife and Ohanekwu, 2005), data transformation

and loading with periodic refreshing. A popular data

warehouse schema approach is the star schema where

there is a central fact table having foreign key at-

tributes that include the main subjects of interest, the

integration attribute, the historical time attribute and

some non-foreign key aggregate measures of inter-

est. Other descriptive tables in the data warehouse

design using the star schema are dimension attributes

for describing the foreign key attributes in the fact ta-

ble. A measure such as score achieved during a game

by a child can be calculated from a multidimensional

model version of the data warehouse called the data

cube (Ezeife, 2001). Other online analytical process-

ing (OLAP) queries such as drill-down and roll-up

analysis can be performed on the data warehouse fact

table or cube. Also, data mining querying for ana-

lyzing the correlation between groups of data in the

warehouse such as association rule mining and clus-

tering can be done to answer required queries.

3 THE OTEP TECHNIQUE

OVERVIEW

3.1 The OTEP Model and Problem

Addressed

As shown in the OTEP model of Figure 1, one goal

of the model is to measure a child’s cognitive abili-

ties through its performances in repetitive playing of

a variety of games in different cognitive categories.

The model proposes to accomplish this goal by com-

paring the child’s performance in these games with

the performances of dynamically changing normal-

ized performances(termed norms) of other childrenin

similar comparison groups such as age, ethnic back-

ground, social background, learning or physical level,

etc. In order to gather the necessary historical, inte-

grated and correlated game playing scores and param-

eters of each child, this approach proposes using the

data warehouse approach to integrate structured data

sources. The structured data sources to be integrated

consist of game playing database, cognitive inventory

database, and other data sources such as learning in-

ventory database. The game playingdatabase can also

result from a continuous integration of various gam-

ing sites. From the integrated data warehouse, online

analytical processing and mining of data for compar-

ative purposes can be performed. For example, some

analysis can be accomplished using multidimensional

data cube views (Ezeife, 2001).

Figure 1: The OTEP Model.

Some Challenges to be Resolved

Some of the questions to be answered include:

1. The data warehouse techniques to be used for con-

tinuous integration of both database schemas and data

of different game sites as different attribute names

need to be merged as one name in the data warehouse.

Also, data cleaning needs to be done to resolve differ-

ences in units of data (e.g., when one database stores

a games parameter in inches and the other stores them

in meters).

2. Establishing the approach for computing and re-

vising the ’Norms’ data. For example, the Norms for

children who are 8 years old and who are playing a

game involving reading can be computed using the

average speed over a period of time. However, many

other ways for computing the Norms are to be inves-

tigated.

3. Data structuring, integration and mapping of the

cognitive inventory to various levels of Norms and

performances as well as to learning inventory.

4. Determining types of analysis and reporting that

are needed to indicate a child’s cognitive abilities and

improvement from the integrated data. Multidimen-

sional online analytical data cube approach is one

technique to be explored.

5. Establishing ﬁrmer correlations between playing

visual games for example, to learning skills.

6. What does it mean when a ﬁrst level of a game

is played in comparison to when a second, third or

higher level is played with regards to performance in

cognitive abilities?

CSEDU2012-4thInternationalConferenceonComputerSupportedEducation

498

Table 1: General areas of cognitive processing and related abilities (based on Crouse, 2007).

Cognitive Sub-categories

Category

Visual seeing differences between things, remembering visual

processing details, ﬁlling in missing parts in pictures, remembering

characteristics, visual-motor coordination, visualization and

imagination, organization of their room, desk, artistic skills

Auditory hearing differences between sounds/voices, remembering speciﬁc

Processing words or numbers, remembering general sound patterns, under-

standing when they miss some sounds, blending parts of words together, music

Sequential Short-term memory for details, long-term retrieval of details, ﬁne-

Rational motor coordination, ﬁnding the words you want to say or write,

organization of your thoughts and materials, writing mechanics (spelling,

punctuation), reading speed/sounding out new words, attention to

details, putting words and thoughts in order

Conceptual memory for general themes or ideas, reasoning spatial awareness,

Abstract general knowledge, inferential thinking, estimation/approximation,

conceptual understanding, creativity/inventiveness, reading

comprehension, use of context rhythm, music, art

Processing short-term memory (with time pressure), long-term retrieval

Speed (with time pressure), talking speed, word-ﬁnding, writing speed,

reading speed, attention reasoning (with time pressure),

general response speed

Executive ability to stay focused on tasks, plan and anticipate, organize

Functioning thoughts and materials, follow-through and complete tasks,

cope with unstructured situations, cope with changes in routine, regulate emotions

Table 2: Correlation between areas of cognitive processing and student achievement.

Cognitive Auditory Visual Sequential Concept Speed Executive

skill

Basic high moderate high high

Reading

Reading high high moderate

Comprehension

Math high moderate moderate high

calculation

Math high high moderate moderate

Reasoning

Writing high high high high high

Mechanics

Writing high high moderate

Content

Oral high moderate moderate

expression

Listening high moderate

Comprehension

7. How do we compare a single person’s record and

compare it to hundreds of people’s game records.

3.2 The Cognitive Categories

The cognitive categories being addressed or targeted

are based on six general areas of cognitive process-

ing: Visual processing; Auditory processing; Sequen-

tial/rational processing; Conceptual/abstract process-

ing; Processing speed; and Executive Functioning

that involve abilities listed in Table 1.

SCHOOLAGECHILDREN'SCOGNITIONIDENTIFICATIONBYMININGINTEGRATEDCOMPUTERGAMES

DATA

499

3.3 The Video Games Categories and

Mapping

The current collection of games has about one hun-

dred games in ten categories encompassed by the

six general areas of cognitive processing of Table 1.

These ten cognitive categories are: 1. Visual Process-

ing, 2. Processing Speed, 3. Reasoning, 4. Self Di-

rection, 5. Self Regulation, 6. Auditory Processi ng,

7. Language Processing, 8. Motor Functioning, 9.

Memory/Learning, and 10. Flexibility. Each of these

main game categories has sub categories of games

they contain. For example, the visual processing cat-

egory has games that are classiﬁed as one of the four

types of i) puzzle games (e.g., number balls, the miss-

ing jigsaw and ﬁnd the pair), ii) action games (e.g.,

fruit collection, mouse and cat, christmas gifts), iii)

sports games (e.g., free kick, freestyle soccer, aster-

oids) and iv) educational games (e.g., word search

II, whack a difference). Some games may belong

to more than one main game category. For exam-

ple, motor functioning games consist of some types

of action and sports games that may include some

games of those types in the visual processing cate-

gory. The games inventory continues to expand and

each of these games has a web link where they can be

played. The games results are classiﬁed as one of the

following ﬁve outcomes: 1) extremely below average,

2) below average, 3) average, 4) above average, or 5)

extremely above average.

3.4 Theoretical Correlation between

Cognitive Processing and Student

Achievement

According to (Crouse, 2010), a student needs to

achieve, for example, high score in auditory domain,

high in conceptual and moderate score on speed of

processing in order to have the ability for reading

comprehension. Similarly, one can determine stu-

dent’s strengths or weaknesses in the domains of

mathematics calculation or oral expression, and sim-

ilar (see Table 2). Table 2 presents the relationship

between cognitive processing and student achieve-

ments. From Table 2, assessing a child’s math cal-

culation skills or abilities as being strong can be done

with tests that would ascert the child’s visual skills as

high, conceptual skills as high, processing and execu-

tive skills as moderate. The OTEP approach is to use

a natural testing environment consisting of computer

games where these skills such as auditory, visual, se-

quential, conceptual, speed and executive are directly

measured through the child’s performance in repeti-

tive playing of games in various categories.

3.5 The OTEP Data Warehouse

Integration Approach

The goal of this system is to use the online games to

screen or assess children’s cognitive skills and later

suggest a learning plan that would be most suitable

for their learning success. While this paper describes

gathering and integrating the relevant data from (1)

video games data, (2) cognitive skills and mapping

data and to obtain a data warehouse schema called

OTEP GamesDW, in the future, other data sources

will be integrated including the learning achievement

data and third party data. The current schemas of the

games data source and the cognitive data source with

the integrated data warehouseare provided in this sec-

tion.

The Games Data Source

The database schema for the games data source to be

integrated holds data descriptors for each of the 100

or more games that a child can play. Information on

the games include the type of game it is. For exam-

ple, a game called “Building Blocks” belongs to the

game category type of “Puzzle” with a description of

the type of activity involved as “Scanning”. For this

game, the maximum allowed number of replays is 3

and the web link where this game can be found is in-

cluded as well as the override options that this game

can be set on. Other tables that are presented, con-

tain games instance, and records of each user’s play

of each game. The games data source schema is rep-

resented through the following eight table schema.:

1. Usertable(userid, useridc, userlogon, usermail,

timecreated);

2. Gamestable(gameid, gametag, gamepath, swfpath,

confpath, gametype, title, typedesc, maxplay, link,

options);

3. GamesCategorytable(catid, cattag);

4. Gamesidentiﬁcation(catid, gameid);

5. GamesConﬁguration(conﬁd, gameid, conﬁgmatch,

timestamp conﬁg, conﬁguration);

6. GameInstance(usergamesid, userid, catid, insttag,

timecreated inst, timecompleted inst, gameslist);

7. UserGameInstance(usergamesid, gameid, time-

played, timecompleted, gamecompleted, numﬁn-

ished, numtries); and

8. UserGamePlayLog(usergamesid, gameid, times-

tamp, conﬁd, status, completed, playresult);

The Cognitive Data Source

The current cognitive categorization has each game

belonging to one of the following ten main cate-

gories: 1. Visual Processing, 2. Processing Speed,

CSEDU2012-4thInternationalConferenceonComputerSupportedEducation

500

3. Reasoning, 4. Self Direction, 5. Self Regulation,

6. Auditory Processing, 7. Language Processing,

8. Motor Functioning, 9. Memory/Learning, and

10. Flexibility. Each of these main cognitive

game categories (i.e., Visual Processing, Processing

Speed, Reasoning, Self Direction, Self Regulation,

Auditory Processing, Language Processing, Motor

Functioning, Memory/Learning, and Flexibility) is

then mapped to one of the currently deﬁned eight

categories. The sub categories are deﬁned based

on the main category. For the ten main categories,

the corresponding two to eight subcategories are

provided next to them below:

1. Visual Processing has Saccadic tests, visual array,

feature detection, scanning, symbol processing,

visual perception, spatial perception, and non-verbal

solving.

2. Processing Speed has speed-efﬁciency, verbal

output, and written output.

3. Reasoning has verbal reasoning, complex problem

solving, non-verbal reasoning, and social component.

4. Self-Direction has planning, strategy generation,

organization, prioritizing, initiation/activation, and

task monitoring.

5. Self-Regulation has emotional, motor output

(writing), self-monitoring, attention, behavior, and

cognitive inhibition.

6. Auditory Processing has conduction, perception,

processing, direction following, and comprehension.

7. Language Processing has language concepts,

verbal logic test, reading, verbal expression, word

ﬁnding, and spelling.

8. Motor Functioning has speed, dexterity, static

steadiness, motor planning, and visual motor pro-

cessing.

9. Memory/Learning has verbal, non-verbal, spatial,

and knowledge aspect.

10. Flexibility has working memory, and set shifting.

The following database schema presents the cog-

nitive data source, which describes the cognition lev-

els and the connection to game play instances.

1. UserCogMeasure(cogid, cogsubid, usergamesid,

userid, usercatid, normcogid, measure);

2. GamesCogMap(gameid, cogid, cogsubid);

3. CognitionCat(cogid, cogdescr);

4. CognitionTable(cogid, cogsubid, cogsubdesc);

5. GamesCogNorm(gameid, age, sex, highscore,

mediumscore, lowscore);

The Data Warehouse Schema

Using the Star schema which consists of a central

main fact table called FactTable, the data warehouse

contains an integration of the relevant data from the

Games data source and the Cognitive data source that

will enable continuous answering of needed queries.

One such needed query is “Given the game play data

of a child, what is the cognition language process-

ing skill of this child (an in particular, her word ﬁnd-

ing abilities) in comparison with the normal cogni-

tion level for this game and for this child’s age, eth-

nic, and other group the child may belong to?”. In

addition to the FactTable, the data warehouse star

schema also contains dimension tables (e.g. Users-

Dim) , which are descriptor tables for the foreign key

attributes (e.g., userid) in the FactTable. The follow-

ing schema is similar to the current data warehouse.

1. FactTable(userid, gamid, gameseq, gameDB,

gamelevelid,catid, normcogid,cogid,cogsubid, time-

m, coglevel, gamescore, duration, tries);

2. GameCategory-Dim(catid, catname);

3. Users-Dim(userid, gender, age, userlogon, user-

mail);

4. GamesDB-Dim(gameDB, weblink);

5. GamesLevel-Dim(gamelevelid, gameid, levelid);

6. Games-Dim(gameid, title, maxplay, gametag,

gamepath, gametype, link, cogid, cogsubid);

7. Cognition-Dim(cogid, cogsubid, cogdesc);

8. CogCategory-Dim(cogid,cogdescr);

9. CogGameMap-Dim(gameid,cogid,cogsubid);

10. GameCogNorm(gameid,age,sex, highScore,

mediumScore, lowScore);

3.6 The Data Integration Algorithm

The data integration algorithm used to integrate the

various relevant data sources into a single historical,

subject-oriented, data warehouse which is refreshed

periodically and mined to obtain changing normal

data and user data is provided as algorithm 1. The

current implemenationdetails ofthe system are as fol-

lows: the data warehouse and source databases are

stored in mySQL database management system, the

data integration algrorithms are implemented using

mySQL stored procedures and triggers. The interface

for all the databases including the data warehouse is

implemented using PHP, JavaScript and jQuery tools.

The OLAP interface for multi-dimensional querying

and data cube views is implemented using JAVA, JSP,

Mondrian OLAP engine and MDX language for data

cube querying.

3.7 Using MultiDimensional Data Cube

Views for Queries

From the historical data in the data warehouse Fact-

Table, a data cube can be used to compute normal

cognitive measures and other results. For example,

the average score of a speciﬁc game such as ‘Number

SCHOOLAGECHILDREN'SCOGNITIONIDENTIFICATIONBYMININGINTEGRATEDCOMPUTERGAMES

DATA

501

Algorithm 1: The OTEP GamesDW Game Play Integration

Algorithm.

Algorithm OTEP GamesDW()

Input: New GamePlay DB (GDB), Cognitive DB (CDB)

Data Warehouse (DWH)

Output: Clean updated Data Warehouse (UDWH)

Variables:

BEGIN

UDWH = DWH

FOR each new record in GDB.UserGamePlayLog

BEGIN

EXTRACT and CLEAN

(userid, gameid, gameDB, gamelevelid,

tries, score, timereq, completed)

FROM GDB.UserGamePlayLog

IF userid NOT IN UDWH.Users-Dim

THEN

EXTRACT and CLEAN User

information FROM DS.usertable

INSERT into UDWH.Users-Dim

END

IF gameid NOT IN UDWH.Games-Dim

THEN

EXTRACT and CLEAN Games

information including

FROM GDS.gametable

INSERT INTO UDWH.Games-Dim

END

IF gameid is not IN

UDWH.CognitiveMain-Dim

THEN

EXTRACT and CLEAN Cognitive

Categories of gameid FROM CDB

INSERT INTO UDWH.CognitiveMain-Dim

END

isVALID = DataValidatonProcess

(GamePlayData)

IF (GDB.UserGamePlayLog.completed

= ’Y’ AND isVALID = ’TRUE’)

BEGIN

INSERT

(userid, gameid, gameDB,

gamelevelid, tries, score,

timereq, completed )

INTO UDWH.FactTable

END

Balls’ for different ages of children between 3 and 8

years, during several attempts of 1st, 2nd, 3rd and 4

tries. This is accomplished with an SQL query on the

fact table data that computes the average gamescore

group by user age and group by tries. The sample

data cube view for this query is given as Figure 2.

3.8 An Example Application of our

Olap Approach for Queries

The algorithm 1 provides a way to keep updating a

data warehousethat is built as an integration of a num-

Figure 2: Example compute norm score cube.

ber of relevant data sources such as games data and

cognitive data. The data warehouse now has the cen-

tral fact table and the adjoining dimension tables and

multi-dimensional or data cube querying (also called

online analytical processing) can be used to answer

a lot of needed comparative queries. Thus, when a

child plays a game, a relevant record from the game

play session that is integrated into the data warehouse

has the schema

GameplayRecord (usergamesid, gameid, timestamp,

completed, value);

An example of such stored record is:

(26, 28, 2010-11-09 02:04:48, Y, {“totaltime”:253.496,

“levels”:({“time”:253.496, “success”:0, “score”:140,

“level”:1}), “totalscore”:140, “status”:3})

The sample data warehouse fact table schema after

integration and a sample record from this schema are

given below.

FactTable(userid, gamid, gameseq, gameDB,

gamelevelid, catid, normcogid, cogid, cogsubid,

time-m, coglevel, gamescore, duration, tries);

(33, 28, 20101109020448, ’think2learn’, 1, 1, 5, 4,

20, ’2010-11-09 02:09:48’, 3, 140, 253.496, 4)

To answer a query similar to query 1 in section

3.1, given below: “What are the norms (cognitive

norms and game achievement norms) for children

who are 8 years old and who have difﬁculty reading

for a reading game?”. This query may be answered

with an SQL instruction on the data warehouse fact

table joined with its dimension tables given below:

SELECT

u.age,u.gender,g.cogdesc,avg(c.coglevel),

avg(c.score)

FROM cleanfact c,users dim u,gamecog dim g

WHERE c.user id=u.user id and

c.game id=g.game id and g.cogid=100 and u.age=8

GROUP by u.age,u.gender;

Sample results from the above query is given below.

(Age, Gender, Cognitive Type, Average Cognition,

Average Score)

(8, F, Visual Processing, 2, 1980.00);

CSEDU2012-4thInternationalConferenceonComputerSupportedEducation

502

(8, M, Visual Processing, 3, 2155.71);

The above query has shown only a slice of the data

cube for children of age = 8. To display the entire

norms for children of all ages, we would just remove

the restriction condition in the “where” clause for the

speciﬁc age = 8.

4 PERFORMANCE STUDY OR

ANALYSIS

In this section we provide readers with an example of

user performanceanalysisbased on a sample (N=106)

of randomlyselected test cases stored in our database.

Our analysis takes into account the following vari-

ables: (a) demographics information, such as user’s

age and gender; (b) qualitative information, such as

type of game as given by the production company,

two major cognitive attributes of the game as deter-

mined by an OTEP psychologist, and two cognitive

subcategories (4 in total); and (c) quantitative param-

eters of user performance, such as, score achieved

in each game, number of trials for each game level,

and time needed to complete each level. After a user

plays a number of games from one cognitive cluster,

his or her average score for this cognitive category is

compared to the normative score of the comparative

group of users (in terms of age and gender). The user

is then ranked according to the following schema:

if the score deviates up to one standard deviation

(1SD) from the normative score (mean = M), the user

demonstrates the ‘average range of skill’ for the cog-

nitive category; if the score deviates more than 1SD-

2SD from the normative score, the user demonstrates

the ‘moderate relative’ strength or weakness (depend-

ing on whether the user’s performance is above or be-

low the average score); similarly, ‘signiﬁcant relative’

strength or weakness is recorded if the user deviates

2SD-3SD from the average. The idea is to recom-

mend action based on the child’s current performance:

if the child has scores such that point to his/her weak-

ness in the area of Conceptual reasoning (see Table

2, this may affect Math Calculation, Math Reasoning

and Writing Content). So, our hypothesis is that if the

system suggests to the child more games in the area of

Conceptual Reasoning, the child will eventually de-

velop cognitive strengths relevant for his/her school

achievement. Recall that standard deviation (SD) is

a measure of how spread out numbers (e.g., n game

scores X

.. . X

) are from the average or mean (M)

and it is it is the square root of the Variance (V). The

formulas for computing the mean M, variance V and

standard deviation SD are respectively: M =

∑

i=1

V =

∑

i=1

−M)

SD = 1/n

∑

i=1

− M)

Standard Deviation provides a way to know what

is normal range of data (usually those within 1SD

or within 68% from the mean), and what is moder-

ate performance (usually those within 1SD-2SD or

within 68.1% − 95% from the mean), and those that

are extra large (strength) or extra small (weakness)

(usually those within 2SD-3SD or within 95.1% −

99% from the mean).

4.1 Performance Study Example 1

Our queryshowed that in the sample of test data, there

were 28 girls and 78 boys, with average age of 13.5

years. There was no statistical difference between

boys’ and girls’ general performance (score or time

needed to complete the game(s)). Say that an 8-year

old girl (id = 95), played a game titled “Stationery”,

which was by the vendor described as Type = “spa-

tial” game, and by the psychologist as Cognitive Cat-

egory 1 = “visual processing”, with Subcategory 1=

“visual perception”. The girl achieved Score = 24185

points on this game, for which it took her 42 min-

utes to complete. Overall, this child played 5 games

in the visual processing category and obtained mean

game score of M = 6937 points (SD = 994.4 points),

and played for mean time of M = 80 min (SD = 50.3

minutes). The question is how this child compares

with other girls or children of similar age who played

games in the same category. After performing a one

sample t-test, there was no statistically signiﬁcant dif-

ference between the score of this child and other chil-

dren who played similar games (t = .810, df = 58,

p > .4). There was no statistically signiﬁcant differ-

ence between the time it took this child to play the

visual processing games and it took other children to

do so (t = 1.09, df = 59, p > .2) However, when this

child’s average score on visual processing games was

compared to scores of other 8-year olds, the statisti-

cally signiﬁcant difference was established (t = -3.30,

df = 10, p < .01). According to Table 1, the child who

is lackingin the visual processing maybe at risk in the

mathematics calculation and reasoning, and in writ-

ing mechanics. Remediation could be recommended

in form of playing more games in the related category

and following the child’s improvement accordingly.

5 CONCLUSIONS

In this paper, we reported on our current work on the

the online product called “Think2Learn” developed

by OTEP Inc. (Online Training & Education Portal).

SCHOOLAGECHILDREN'SCOGNITIONIDENTIFICATIONBYMININGINTEGRATEDCOMPUTERGAMES

DATA

503

OTEP presently contains about 100 video games for

testing purposes, it records players’ scores to continu-

ously assess and monitor their cognitive strengths and

weaknesses with regards to the main cognitive cate-

gories. The Web based tool for identifying cognitive

skill level is developed as an integration or data ware-

house of a number of relevant data sources such as the

cognitive skills categories data, games data, player in-

ventory data and so on. The integrated data are con-

tinuously mined, analyzed and queried for proper and

quick assessment or recommendations.

We are presently working on: (a) increasing num-

ber of games; (b) increasing a reliability of catego-

rization of the games by achieving an agreement be-

tween 2-3 scorers; (c) developing a formula that will

incorporate the features of the game (including dif-

ferentiating the impact of different cognitive sub cat-

egories), the number of trials, the scores achieved and

the time spent playing. Once we have established col-

laboration with some schools, the system will asso-

ciate with each child a unique identiﬁer tracking the

children’s cognitive development, proposing remedi-

ation, increasing validity and reliability of our ap-

proach.

REFERENCES

Berk, R. (2010). How do you leverage the latest technolo-

gies, including Web 2.0 tools, in your classroom? .

International Journal of Technology in Teaching and

Learning, 6(1):1–13. .

Bogost, I. (2007). Persuasive Games: The Expressive

Power of Videogames. The MIT Press.

Canada, M. (2011). Child and youth development.

http://www.mentalhealthcommission.ca/English/Pages/

ChildandYouth.aspx.

Crouse, S. (2010). The Cognitive Processing Inventory

(CPI). LDinfo Publishing.

Dance, F. (2003). The digital divide. In Strate, L., Jacobson,

R. J., and Gibson, S., editors, Communication and cy-

berspace: Social interaction in an electro nic environ-

ment. Hampton Press, Inc., Cresskill, New Jersey.

DeLisi, R. and Wolford, J. (2002). Improving children’s

mental rotation accuracy with computer game playing.

Journal of Genetic Psychology, 163(3):272–282.

Ezeife, C. (2001). Selecting and Materializing Horizon-

tally Partitioned Warehouse Views. Journal of Data

& Knowledge Engineering, 4(1). Elsevier Sciences.

Ezeife, C. and Ohanekwu, T. (2005). Use of smart tokens

in cleaning integrated warehouse data. International

Journal of Data Warehousing and Mining (IJDWM),

1(2):1–22.

Freeman, N. and Somerindyke, J. (2001). Social play at the

computer: Preschoolers scaffold and support peers’

computer competence. Information Technology in

Childhood Education Annual, 2001(1):203–213.

Gee, J. (2007). Good Video Games and Good Learning:

Collected Essays on Video Games,Learning, and Lit-

eracy. New York: Peter Lang.

Greenﬁeld, P., deWinstanley, P., Kilpatrick, H., and Kaye,

D. (1994). Action video games and informal educa-

tion: Effects on strategies for dividing visual atten-

tion. Journal of Applied Developmental Psychology,

15(1):105–123.

Han, J., Kamber, M., and Pei, J. (August 2011). Data Min-

ing: Concepts and Techniques . Morgan Kaufmann

Publishers.

IBM (1991). Literacy development. (ibm’s writing to

read, and principles of adult literacy system multime-

dia/educational software). T H E Journal (Technologic

al Horizons in Education).

Junco, R. and Mastrodicasa, J. (2007). Connecting to the

net.generation: What higher education professionals

need to know about today's students. Student Affairs

Administrators in Higher Education (NASPA), Wash-

ington, DC.

Lieberman, D., Fisk, M., and Biely, E. (2009). Digital

games for young children ages three to six: From re-

search to design. Computers in The Schools, 26(4).

Martinovic, D., Freiman, V., and Karadag, Z. (2011). Child

and youth development beyond age 6 - transitions to

digitally literate adults. Unpublished Report. Ministry

of Child and Youth Services.

Oblinger, D. and Hawkins, B. (2006). The myth about stu-

dent competency. EDUCAUSE Review, 41(2). http://

www.educause.edu/apps/er/erm06/erm0627.asp.

Oblinger, D. and Oblinger, J. (2005). Educating the

net generation. educause. net.educause.edu/ir/ibrary/

pdf/pub7101.pdf.

Pedro, F. (2006). The new millennium learners: Chal-

lenging our views on ict and learning. http://

www.oecd.org/dataoecd/1/1/38358359.pdf.

Pillay, H. (2002). An investigation of cognitive processes

engaged in by re creational computer game players:

Implications for skills for the future. Journal of Re-

search on Technology in Education, 34(3):336–350.

Pletka, B. (2007). Educating the net generation. Santa

Monica Press, Santa Monica, CA.

Rideout, V., Roberts, D., and Foehr, U. (2005). Generation

M: Media in th e lives of 8-18 year-olds. The Henry J.

Kaiser Family Foundation.

Rockman, I. and Associates (2004). Integrating information

literacy into the higher education curriculum: Practi-

cal models for transformation. Jossey-Bass, San Fran-

cisco, CA.

Subrahmanyam, K. and Greenﬁeld, P. (1994). Effect of

video game practice on spatial skills in girls and

boys. Journal of Applied Developmental Psychology,

15(1):13–32.

Subrahmanyam, K., Greenﬁeld, P., Kraut, R., and E.Gross

(2001). The impact of computer use on children’s and

adolescents’ development. Journal of Applied Devel-

opmental Psychology, 22(1):7–30.

Subrahmanyam, K., Kraut, R., Greenﬁeld, P., and Gross, E.

(2000). The impact of home computer use and chil-

dren’s activities and development. The Future of Chil-

CSEDU2012-4thInternationalConferenceonComputerSupportedEducation

504

dren: Children and Computer Technology, 10(2):123–

14.

Tapscott, D. (2009). Grown up digital: How the net gen-

eration is changing your world. McGraw Hill, New

York, NY.

Thorell, L., Lindqvist, S., Nutley, S., Bohlin, G., and Kling-

berg, T. (2009). Training and transfer effects of exec-

utive functions in preschool children. Developmental

Science, 12(1):106–113. .

Vaidhyanathan, S. (2008). Generational myth: Not all

young people are tech-savvy. The Chronicle of Higher

Education, 55(4). B7.

Veen, W. and Vrakking, B. (2006). Homo Zappiens, Grow-

ing up in a Digital Age. Network Continuum Educa-

tion, London.

Whent, R. (2012). Otep think-2-learn: A private online pre-

screening solution for children. http://www.think-2-

learn.com/.

Zull, J. (2004). The art of changing the brain. Teaching for

Meaning, 62(1). .

SCHOOLAGECHILDREN'SCOGNITIONIDENTIFICATIONBYMININGINTEGRATEDCOMPUTERGAMES

DATA

505