Document Type


Date of Degree


Degree Name

PhD (Doctor of Philosophy)

Degree In


First Advisor

McCarthy, Ann Marie

First Committee Member

Delaney, Connie W

Second Committee Member

Street, William Nick

Third Committee Member

Lu, Der F

Fourth Committee Member

Kleiber, Charmaine

Fifth Committee Member

Vilhjalmsson, Runar


Icelandic children can expect to live a long and healthy life and have the right to the highest possible standard of health. Despite this, as in other Western countries, the prevalence of psychosocial complaints and long term conditions in Icelandic children is growing and they are struggling with increased levels of preventable health conditions.

The purposes of this cross sectional, secondary analysis were to perform a psychometric evaluation on the instrument School-Children Health Promotion; to describe self-reported health promotion behavior of 10-12 year old Icelandic school children, and to predict novel and potentially useful patterns of health promotion behavior of 10-12 year old Icelandic school children using data mining methods. Existing data from 480 10-12 year old Icelandic school children and 911 parents were analyzed.

Analysis of the instrument School-Children Health Promotion indicates that it is, in general, a valid and reliable instrument for measuring health promotion behavior of 10-12 year old Icelandic children. Five factors emerged from the 21 item instrument, which were labeled: "Positive Thinking." "Diet and Sleep Pattern," "Seek Psycho-social Support," "Coping Behavior," and "Health Habits." The results indicated that girls use more positive health promotion behavior than boys; however, differences in health promotion behavior between 5th and 6th grade students were not obvious. The results of data mining analyses, using the classifiers decision tree (J48) and logistic regression (Logistic) to predict health promotion behavior, showed better performance with the subsets of the five factors and the overall instrument than with the full dataset of 199 items. For the subsets, the logistic regression models performed better than the decision trees with AUC ranging from 0.71 to 0.80. The strongest predictors of health promotion behaviors were validation and caring in friendship, intimate disclosure between friends, and quality of life.

Results of this secondary analysis indicate that friendship is of vital importance with regards to health promotion behavior. Therefore, further studies on the effect friendship has on health promotion behavior of Icelandic children in the 10-12 year old age group are clearly needed.


health promotion, preadolescents, Iceland, KDD, data mining


xv, 285 pages


Includes bibliographical references (pages 187-201).


Copyright 2008 Brynja Orlygsdottir

Included in

Nursing Commons