Independent evaluation of the SAPLER peer counselling programme

Impact of the programme on learners’ knowledge, attitudes and behaviours

To provide an overall comparison between learners from Sapler versus non-Sapler schools, four global scores summarizing the main themes in the questionnaire were used (see Appendix 2 for an explanation of the scores). 

It was found that learners from Sapler schools scored consistently better than learners from other schools on these four measures of safe-sex knowledge, safe-sex attitudes, sensible condom use and sexual risk behaviour. The difference between Sapler and other schools in attitudes was statistically significant while the difference in risk behaviour approached statistical significance.

Girls from Sapler schools showed consistently better scores on knowledge, attitudes, condom use and risk behaviour than girls from other schools, with the difference in attitudes being statistically significant.

1 T tables such as this one can be interpreted as follows: The values under "2-tailed T" and "St. Dev." are included to facilitate statistical cross-checking and can be ignored for  purposes of basic interpretation. "Other N" and "Sapler N" are the number of learners from non-Sapler and Sapler schools respectively for whom it was possible to compute ascore. 

"Other mean" and "Sapler mean" are the mean (i.e. averages) scores obtained by learners from each of the two types of school. "p-level" is an indication of whether the difference between the means are statistically significant - the smaller the p-level, the greater the probabllity that there is a real difference. A p smaller than 0.05 is considered to be significant, while a p smaller than 0.01 is considered to be highly significant.

Boys from Sapler schools showed consistently better scores on knowledge, attitudes and risk behaviour than boys from other schools, with the difference in attitudes statistically significant. However, they had slightly worse scores on the condom-use scale.

Thus global indicators tentatively suggest that the Sapler programme does teach learners more accurate facts about safe sex behaviour, and does influence them to adopt more constructive attitudes and less risky behaviours. 

The measured effects of the programme are not large, but this may in part be due to practical difficulties in obtaining a fully equivalent comparison sample (see Appendix 1). The more detailed analysis of individual questionnaire items that follows below should be read in the context of this main finding.

It is often argued that safe sex knowledge and attitudes do not necessarily translate into appropriate behaviour, and looking at the correlations between knowledge, attitude and behaviour in this sample one can see that this is indeed the case. 

Knowledge and attitudes are fairly weakly but statistically highly significantly correlated, i.e. there is a tendency for learners who know about safe sex also to have appropriate attitudes towards safe sex. Similarly condom use and risk behaviour is highly negatively correlated, i.e. learners who use condoms sensibly also engage in fewer risk behaviours. (The magnitude of this latter correlation is artificially inflated because the risk scale contains some of the same items as the condom use scale.) 

However, knowledge and attitudes are not significantly correlated with behaviour, i.e. learners who know about safe sex and have positive attitudes towards it are not necessarily the same ones who use condoms sensibly or the ones who show other low-risk behaviour patterns.

2 These correlation matrices can be interpreted as follows: The top figure in each cell is the correlation coefficient. 0 indicates no relationship between the two variables, 0.3 is a moderate strength relationship, 0.6 is a strong relationship and so on up to 1 which is a perfect relationship. 

Correlations with minus signs are negative correlations, indicating that learners who obtains high scores on the variable tend to obtain low scores on the other. The second figure in each cell (after "N=") is the number of learners for whom scores were available for both variables. 

The last figure in each cell (after "p<") is the probability level, indicating if the correlation is likely to be a chance event or statistically significant. A p of smaller than 0.05 is considered significant, smaller than 0.01 highly significant.

However, when learners from Sapler schools only are considered, the pattern is somewhat different. Although the most significant correlations are still between knowledge and attitudes on the one hand and between the two behavioural measures (condoms and risk taking) on the other, in addition attitudes also correlate significantly (albeit fairly weakly) with condom use and risk behaviour. 

This could tentatively be interpreted as an indication that the work of the Sapler pioneers may be beginning to forge a link for learners between attitudes and behaviour.

Such links are not evident from the correlation matrix for learners from non-Sapler schools. 

However, as will become clear from the detailed analysis of responses to the questionnaire, there is still a pervasive lack of linkage for all learners between knowledge and attitudes on the one hand and behaviour on the other. Along with other organisations in the field of safe sexuality education, this is likely to remain a major challenge for Sapler.

[ Home ] [ Introduction ] [ Method ] [ Att.& Behaviours ] [ Conclusion ] [ App.1: Sampling ] [ App. 2: Scale Dev. ] [ Appendix 3 ]