The Classification of Multiple Intelligences of People with Epilepsy

Purpose: to identify the intelligence profile of epilepsy patients for job placement purposes using Hierarchical cluster analysis. The classification was based on Gardner’s eight intelligences musical, bodily-kinesthetic, math/logic, spatial, linguistic, interpersonal, intrapersonal and naturalist. The study also seeks to identify determinants of employability based on the identified clusters and classifications. Methods: A psychometric test, named Ability Test in Epilepsy (ATIE) was developed, validated and used to measure the level of eight intelligence types of epileptic patients. The subjects were randomly selected from the epilepsy out-patient at the Neurology Department, Kuala Lumpur General Hospital. Only 166 patients had met research pre-requirements and were included in the actual study. The data were collected through the administered test, the examination of the subjects’ medical reports and face to face interviews. The study subjects were then classified into eight types of intelligences based on the ATIE scores. In order to classify the intelligence of the epilepsy patients, Hierarchical Clustering technique was applied. Results: The focal group has higher mean scores in Musical, Verbal and Naturalistic items while the generalized group has higher mean scores in other five intelligences. Three cluster solution was found to be the most appropriate to represent the intelligences with the distance coefficient difference from d=119.8 to d=139.2. Spatial and kinesthetics have the smallest Euclidean distance, d=101.51 and interpersonal and music have the largest distance, d=224.03. From the hierarchical cluster analysis, three new clusters have been derived that represents Agility (musicalpersonal), Logical Thought (math/logic -personal-spatial), and Foundation Skill (combination of six skills). Conclusions: The designed test, ATIE had assisted the epilepsy patients to identify their strengths and levels of their competencies where the results of ATIE can also help the people with epilepsy (PWE) to enhance the chances of being employed. The eight intelligence types of the 166 epilepsy patients had been successfully clustered into three more meaningful clusters based on their intelligence scores and characteristics which are musical, math/ logic and mixture of six remainder intelligences that agrees with the Gardner’s MI theory. Keywords— Epilepsy; People with Epilepsy; Hierarchical Clustering Method; Multiple Intelligence. Siti Rahmah Awang is with Universiti Teknologi Malaysia, 81310 Skudai Rasimah Aripin is with Universiti Teknologi MARA, 40450 Shah Alam Md. Hanip Rafia is with Department of Neurology, General Hospital Kuala Lumpur, 50586 Kuala Lumpur Tahir Ahmad is with Universiti Teknologi Malaysia, 81310 Skudai.


I. INTRODUCTION
PILEPSY is known as one of the oldest illnesses in history and has affected millions of people for many centuries [1,2].Epilepsy in Malaysia is an area that still needs to be explored.The majority of people with epilepsy (PWE) in Malaysia prefer to be unidentified and fear to disclose their identity.The major problems faced by these people are job employment and stigma [11].
Human ability or human intelligence in the Malaysian society changes with time [13,14].Just like in other cultures, intelligence is regarded as an important asset in achieving one's dream especially in determining one's hierarchy in the Malaysian society [13,14].Even the IQ of people with epilepsy is generally no different from that of people without epilepsy [15] but, at times having seizures is associated with difficulties in learning, psychological wellbeing, and socialization which eventually can slacken their progress in education and employment [11,12].
In 1983, Howard Gardner introduced Multiple Intelligence Theory (MI) in his book, Frames Of Mind [3].He suggested that every human being possess a basic set of intelligences namely musical, kinesthetics, verbal, math/ logic, spatial, interpersonal, and intrapersonal.Then, in 1997 he added the eighth intelligence, naturalist or nature smart [4].Gardner [4] was also interested in two things (i) how it can be ensured that every human being develops to the fullest of his/her intellectual potential and (ii) how to ensure the use of those intelligences for positive ends rather than for destructive.Hence, the strength of each individual that has certain intelligence will enable the individual not only to pursue a suitable job, but also to produce the best in his or her job, so that he or she would get the best in life as well as in his or her career.Yet, logically each job does not focus on a single responsibility but it requires many kinds of intelligences to accomplish the task.Thus, it is important to nurture and polish all of the intelligences in order to be successful especially at workplace.For example, Armstrong [5] points out that the theory of multiple intelligences emphasizes the broad range of ways in which adults pursue their work in life, so it can help youngsters begin to develop vocational aspirations.This study served two general purposes: • To cluster the PWE's intelligence using Hierarchical Cluster Analysis (HCA).• To identify determinants of employability based on the identified clusters and classifications.

A. Subjects
In this study, 166 out-patients at the Neurology Department, Kuala Lumpur General Hospital (HKL) were selected to participate.The patients were classified into two groups according to the seizure type which are the partial and generalized [6].The selected patients were informed about the study objectives as well as the procedures and consent from the patients or the family/caretaker was obtained.
The selection of subjects in this study has undergone several stages before the final subjects were identified.The selection process started with the process of determining the eligible subjects through the patients' medical records obtained from the Neurology Department, HKL.These records were then examined in determining whether these patients were eligible to become samples of this study.If they did not fulfill the following criteria they would be excluded.
i. Age of 16 to 45 years.ii.
Presence of epilepsy confirmed by clinical history.iii.
Having at least one seizure in medical history.iv.
Absence of any other neurological disorder.v.
Capable of being independent and not registered with the Social Welfare Department as a mentally retarded or referred to psychiatrists for having abnormal neurological disorder.vi.
Have at least the experience of being employed.Patients with major mental illness and low mental ability (less than Malaysian standard six level) were also excluded since one of the purposes of this study was to promote the epilepsy people to the job market.Finally, only 166 patients met the criteria and were able to proceed with the study.This also includes the omission due to incomplete answers, doubtful answers, as well as the mental status of patients who are mentally retarded, psychotic or having other mental disorder.

B. Instruments
The main instrument in this study is a structured intellectual ability test, Ability Test in Epilepsy (ATIE) which has been developed based on Gardner's MI theory to measure the eight factors of multiple intelligences among the epilepsy patients.These eight factors are musical, bodily-kinesthetic, logical -math/logic, spatial, linguistic, interpersonal, intrapersonal and naturalist.In this test, a 5point Likert scoring system was employed whereby the score ranges from 1-not at all like me to 5 -definitely me.The research subjects were needed to respond to items that best describe their views, feelings and opinions towards their mental ability or level of intelligence.Based on the test scores, they were classified into eight types of intelligences.
The test was divided into nine main sections: a profile and eight intelligences as shown in the Table 1.

C. Statistical analysis
In order to fulfill the study objectives, Hierarchal cluster analysis (HCA) is applied.This method is widely applicable in research especially social science researches since it can be used to determine clusters of similar objects.HCA is one of the straightforward methods and can be done either agglomerative or divisive [7,8].

A. Patients' Intelligence Profiles
Table 2 shows the distribution of the subjects' intelligence type and their type of seizure.The generalized group has high scores in five intelligence skillsinterpersonal, intrapersonal, kinesthetic, logical and spatial where the interpersonal is the highest score among these five skills.However, the focal group has high scores in Music, Naturalist and Verbal skills where verbal score is the highest.

B. Proximity Matrix
The squared Euclidean distance [7,8] is the most popular resemblance coefficient used for quantitative attributes in clustering.These distances are illustrated in Proximity matrix table (see Table 3) which indicates all the diagonal entries are 0 since an intelligence does not differ from itself.
The smallest distance (d) was identified which presented the similarities of the two objects.Kinesthetics and visual spatial scores have the smallest distance, d = 101.51whereas intrapersonal and music scores have the largest distance, d = 224.03.
Fig. 1 illustrates the scenario brought by the proximity matrix visually.
Since kinesthetics and visual/spatial intelligence have the smallest distance, the two intelligences are the first to be combined as Cluster 1.  4 shows the agglomeration schedule.The column labeled coefficients is the Euclidean distance that represents similarity coefficient to form the clusters.The obtained value of the coefficients suggest that the 3-cluster solution is better than the 4-clusters since the distance coefficient in cluster 3 and cluster 4 drastically jumps from 119.8 to 139.2.

C. The Cluster Memberships
The merging process is also illustrated in the dendogram (Fig. 2) where it started with merging the Kinethetics and Spatial intelligence that based on the smallest rescaled distance; it then followed by {Kinethetics ∪ Spatial} and Verbal.
The process continued with {Kinesthetic ∪ Spatial ∪ Verbal} and Interpersonal and so on.Thus, the dendogram confirms the three cluster solution.

D. Clustering Variables
From the previous analysis, three clusters are formed.Table 5 shows the demographic background of each cluster.Cluster 1 is made up of mostly employed, female with secondary education.The skills that most of them have are intrapersonal and interpersonal.Cluster 2 is dominated by unemployed, generalized and single female who have secondary education.Those in this cluster are likely to have intrapersonal skill.The last cluster, Cluster 3, consists of employed, single, generalized female with secondary and college education.Most in this cluster possess spatial skill.
From the merging process Math/Logic and music skills are actually unique since when a three-solution cluster is formed, those two skills formed their own cluster.Several studies have shown that musicians have distinctively different brains [9,10].

E. Mean Scores
Table 6 shows the descriptive statistics of the three cluster solution.Looking at the mean column in the table, Cluster 3 has the highest mean in all the eight intelligences where the interpersonal and intrapersonal contribute to the highest mean among all, which are 38.23 and 37.92 respectively.

F. Characteristics of Cluster 3
The majority of subjects in this cluster is Generalized seizure, female adults between 30 to 40 years old and employed by the private sectors with high scores in all eight intelligences.Many of them are interpersonal and spatial intelligence type.

IV. CONCLUSIONS
The cluster analysis resulted in three clusters as follows: 1. Cluster 1 namely Agility Cluster represents musicalpersonal skill that exhibited by the employed patients with either generalized or focal epilepsy.
2. Cluster 2 namely Foundation Skill Cluster consists of general cognitive skill that featured by the unemployed people with generalized epilepsy.(Kinesthetics, Visual, Verbal, Interpersonal, Intrapersonal and Naturalist).
3. Cluster 3 namely Logical Thought Cluster portrays the mathematical-personal-spatial skill that characterized by employed, good educational background with either generalized or focal epilepsy.
The first group consists of majority subjects from the study where most of them were still single and many of them were employed.The group composes of a mixture of job roles such as sales promoters, waiters, government support staffs, teachers, a lecturer, an engineer and an assistant director.
The second group acquires all the skills but is not outstanding in any particular one.Those in this cluster are only superficial in all skills where majority of them was International Conference on Management, Economics and Social Sciences (ICMESS'2011) Bangkok Dec., 2011 unemployed.Thus, it can be concluded that this group was multi-skilled but the scores were not high.Thus, to ensure that a person is employed, he/she has to be "really" good multi skills.
The third group comprised of people with higher qualification and employed.Many of them had better job role than the first group.They were executives, officers and businessmen.Therefore, in order to be employed, there are a few prerequisites that need to be considered.
If a person acquired only one skill without knowing how to apply it in his work and daily life, that skill would be ineffective.Spatial skill or visual thinking is an extra skill that is able to help a person to see and recognize objects, faces and details and this ability is needed by engineers, designers and even successful businessmen.Thus, the findings support Gardner's view that there are productive interactions between math /logic and spatial intelligences in areas like engineering and architecture.Besides that, Gardner also pointed out that many scientists and mathematicians are attracted to music which amplifies ideas in music, visual arts and mathematics [4].
As a conclusion, even though members in Cluster 2 have several skills but the scores are indistinctive compared to other clusters.Thus, this group is less likely to be employed and faces stiff competition in the job market.--------+---------+---------+---------+---------  Among the three clusters, Cluster 3 comprises of highly employable members with the combination of the math/ logic, personal and spatial skills.The additional of spatial skill to the other skills can be considered as a 'big plus' to increase the probability of being employed and having higher role at the workplace.For example, arithmetic is often associated with the left brain but for high level mathematics and problem solving, visual-spatial reasoning International Conference on Management, Economics and Social Sciences (ICMESS'2011) Bangkok Dec., 2011

Fig 1 .
Fig 1. Clusters from the Merging Steps in the Proximity Matrix Dendrogram using Average Linkage (Between Groups) -15 included respondents' characteristics like gender, ethnics, age, education level and employment.Items 16-18 addressed the information of respondents' illness such as types of epilepsy, onset age and number of antiepileptic drug (AED).The other eight sections explored the perceptions of the PWE's on their intelligence strengths and weaknesses.

TABLE II DISTRIBUTION
BY INTELLIGENCE AND SEIZURE TYPE *G= Generalized; F=Focal C. The Number of Cluster Table International Conference on Management, Economics and Social Sciences (ICMESS'2011) Bangkok Dec., 2011 International Conference on Management, Economics and Social Sciences (ICMESS'2011) Bangkok Dec., 2011