Page 1536 - SUBSEC October 2017_Neat

Table 10:
Total hardship cases reported for May-June 2017 sitting

Source Frequency Percentage

Weather 533 89%

Medical 45 8%

Compassionate 11 2%

Technical 8 1%

Total 597 100%

224. FAC further heard that in estimating the grades for these weather-affected
candidates, robust statistical procedures were used to model patterns of performance based
on (i) candidates from the current sitting with complete data; (ii) traditional performance of
the affected centres, and (iii) order of merit predictions from current candidates’ teachers.
Using data from these three sources, the statistical model derived was used to estimate
missing profile level scores for affected candidates. With respect to validation, for a sample
of students with complete information was randomly selected to test the accuracy of the
model equations developed for each subject. The candidates were isolated by territory,
profile and/or paper performance.

225. FAC noted that in Table 11 below, the model was found to be robust within the
established accepted statistical standards and methodology. Further discussion of the
procedures may be found in Appendix II

Table 11:

The distribution of resulting estimated grades is as follows:

Grade Distribution

Subject CSEC Frequency I-III IV-VI % I-III

Agricultural Science SA 113 93 20 82%

Theatre Arts 1 1 0 100%

Biology 2 2 0 100%

Industrial Tech (Mechanical Engineering
34 22 12 65%
Technology)

Textiles, Clothing and Fashion 40 27 13 68%

47