Three-tier and Two-tier Specification - Results Analysis of Module 3 Intermediate, Foundation, and Higher Papers

Oct 20, 2026

Report presents results analysis based on item-level data across Module 3 papers under three-tier (Intermediate) and two-tier (Foundation and Higher) specifications. It compares candidates’ performance by raw marks and mean marks as a percentage of maximum mark, and evaluates facility indices for questions targeted at the final grade achieved. Findings show linear trends overall, but substantial within-grade variability in facility indices, with targeted questions working well for some grades while others display wide variability between items and grades.

LanguageEnglish

CategoryExam

FormatPDF

Pages17

Reading Time43 min

AccessFree to download

Results

Three-tier specification: Intermediate paper (33003/I)

Two-tier specification: Foundation paper (43003/F)

Two-tier specification: Higher paper (43003/H)

Tables 1-3 and targeted-item performance

award in both December 2006 and March 2007. In both series, recommended grade changes were made with a view to making the module outcome less severe. Only the Intermediate paper on the three-tier specification has been considered, as this tier has historically had the largest entry. In March 2007, for example, 27,343 candidates entered for the Module 3 Intermediate tier, compared to 3,822 on the Foundation tier, and 3164 on the Higher tier. The analysis conducted on the item level data for each paper was threefold: 1. Candidatesí performance (in terms of raw marks) across all questions targeted at a particular grade by final module grade was displayed graphically (see Appendix 2) . 2. Candidatesí performance (in terms of raw marks) on each item by final module grade was also graphed, in order to ensure that compensatory influences had not been masked by only examining candidatesí scores across a number of items targeted at a certain grade1 . 3. Candidatesí performance (in terms of mean mark and mean mark as a percentage of the maximum mark/facility index) on the questions targeted at their final module grade was assessed. So, for example, A* candidatesí mean mark on each of the questions targeted at A* was calculated (Tables 1-3) Results Three-tier specification Intermediate paper (33003/I) Figures 1-4 (see appendix 2) show the total mean mark that candidatesí achieved across questions targeted at each grade. For each group of targeted grade items a linear relationship was evident between final module grade and total mean mark. Analysis of each individual item revealed a similar pattern. Since there were no instances where C grade candidates, for example, scored more marks on an item targeted at grade B (or across a number of items targeted at grade B) than B grade candidates did, one may argue that targeted questions worked well in this instance. However, when candidatesí performance on items targeted at the final grade they achieved on Module 3 in March 2007 was considered, a different pattern emerged. Table 1 shows the mean mark and facility index (mean mark as a percentage of the maximum mark) that candidates awarded each grade at module level obtained on the questions targeted at that grade. So, for example, B grade candidatesí performance on each of the B grade questions is given. One may expect that candidates that achieved a particular grade at module level may have displayed a similar level of ability on each of the questions targeted at that grade. However, as Table 1 demonstrates, this was clearly not the case. Facility indices for the B grade candidates, for example, varied from 31.7 per cent to 90.0 per cent. Grade E candidates showed the greatest degree of variability across questions targeted at their final grade (1 .5 ñ 83.0%) and grade C the least (24 .7 ñ 74.7%) . Facility indexes across all questions targeted at each grade ranged from 40.0 per cent for grade E candidates to 54.7 per cent for grade D candidates. This suggests that on average candidates obtained around half of the marks targeted at the final grade they received. 1 For reasons of prudent practice, individual graphs for each item have not been included in an appendix. 2 Two-tier specification Foundation paper (43003/F) As with the three-tier Module 3 Intermediate paper, candidatesí scores on items on the two-tier Module 3 Foundation paper increased with module grade awarded (see Appendix 2, Figures 5-9) . However, candidatesí scores on the items targeted at their final grade varied substantially (Table 2) . In terms of mean mark as a percentage of the maximum mark, grade F candidates showed the greatest degree of variability across targeted questions (2 .0 ñ 82.0%) and grade D candidates the least (32 .3 ñ 79.0 %) . Facility indices across all questions targeted at each grade indicated that grade E candidates only obtained on average a quarter (25 .5 %) of the marks targeted at the grade. The most successful candidates

FAQs

How was item-level performance analyzed in the report?

The analysis used three approaches: graphical display of raw marks by targeted grade, item-level graphs to ensure no masking effects, and assessment of mean marks and mean mark as a share of the maximum/facility index for questions targeted at candidates’ final grades.

What pattern is observed for the three-tier Intermediate paper (33003/I)?

Total mean marks show a linear relationship between final module grade and mean mark. Item-level analysis reveals a generally consistent pattern, and targeted questions appear to work well since C-grade candidates do not outperform B-grade candidates on B-targeted items.

How did candidates’ performance vary on the two-tier Foundation and Higher papers?

For Foundation, scores on targeted questions vary substantially by grade, with facility indices showing that some grades obtain only a small fraction of targeted marks on average. For Higher, performance shows a positive linear relationship with awarded module grade, with variability across targeted items and differing ranges by grade.

Oct 20, 2026

Uploaded by Jordan Avery