Overview of Questionnaire Findings
The descriptive statistics for the 54 questions are presented in Appendix 4. What follows
is a list of six items with which there was most agreement and four items with which
there was least agreement.
Items with which there was most agreement were:
Q.11 I like it when people are honest with me (91%)
Q.16 I am usually trustworthy (89%)
Q. 31 I try to be fair and respectful to other people (86%)
Q.49 I am ambitious to do well in my exams (84%)
Q.1a My mother influences my attitude to life (83%)
Q. 4 I am usually truthful and honest (82%)
Items with which there was least agreement were:
Q.1i People in the media influence my attitude to life (23%)
Q.1h Community leaders influence my attitude to life (25%)
Q.13 I take part in school run community projects (28%)
Q.24 I help out in my community (30%)
Identification of influential background variables
In order to identify variables from the data set which were potentially influential over
responses to individual questions, an ordinal logistic regression model was estimated for
each item. All background data (demographic, study intentions, parental education,
school subject options and future study intentions and school) were tried for each item. In
all cases most variables proved to have no explanatory value. Only those which could not
be ignored without reducing the accuracy of the model were considered for presentation
below. In many cases, but by no means all, religion proved to be the most influential
variable.
The following graphs and tables demonstrate for a few questions some of the most
prominent variables from the ordinal regression analysis. The responses to the
questionnaire are collapsed into three responses (yes = agree and strongly agree. Unsure
= neither agree nor disagree. No = disagree and strongly disagree). The items are
presented in relation to the variable which is significant to it. For illustrative purposes,
qualitative data from phase one is included after each item.
33