Scottish Government

Gaelic-medium versus English-medium: an attainment comparison

The SSA survey has provided a wealth of information that will be of interest and value to Gaelic-medium educators and others. But a critical question remains: what is the evidence, if any, that receiving science education wholly or partially through the medium of Gaelic has any effect, positive or negative, on children's science learning? In an attempt to answer this question the attainments of the Gaelic-medium pupils were compared with those of a group of English-medium pupils who had been assessed in the main survey.

The pupils who comprised the comparator group were randomly selected from among those English-medium pupils who had taken the same test booklets as the Gaelic-medium pupils, viz. booklets 17 and 28 at P5 and booklets 33 and 44 at P7, and who lived in one or other of the 11 authorities from which the Gaelic-medium pupils were drawn. The group was also matched in terms of gender and deprivation. In practice five such comparator samples were created, and compared with each other, to ensure that the English-medium group used in the study would not be aberrant in any discernible sense. No differences in fact emerged between the five English-medium subsamples, in terms of their compositions or the attainments of their pupils, and so a single subsample was randomly selected for use in this language-medium comparison.

Table 17 compares the distribution of the two pupil groups, English-medium and Gaelic-medium, over the 11 authorities. As the table shows, the authority distribution was not identical for the two groups. This was unavoidable, given the unbalanced distribution across authorities of the Gaelic-medium pupils: in Highland and Eilean Siar, in particular, there were few English-medium pupils in the main survey from which members of the English-medium group might be drawn. At each stage, therefore, the English-medium group was necessarily completed with disproportionate numbers of pupils from the other authorities.

Table 17
The geographical distribution of the Gaelic-medium and English-medium pupil groups
(192 P5 pupils and 166 P7 pupils in each group)

What, then, were the test performances of the two groups at each stage? Having attempted the same pair of booklets at each stage, test scores for the same single-level tests became available at P5 for Levels B, C and D, and at P7 for Levels C, D and E. The mean percentage test scores are shown in Table 18. There was a small difference in mean test scores at Levels B, C and D at P5, and at Levels C and D at P7, consistently in favour of the Gaelic-medium pupils, with no difference at Level E for P7. The overall test score differences can be attributed to statistically significant differences on some specific tasks within the tests; for most tasks there were no marked performance differences between the two groups of pupils at either stage. It is interesting at this point to note that a previous attempt to compare the science attainments of Gaelic-medium learners and English-medium learners at P7 reported an advantage in favour of the English-medium learners ⁹. However, in this earlier study the English-medium learners were pupils in the same schools as the Gaelic-medium learners, and no attempt was taken to control for gender and deprivation differences between the two groups.

Table 18
Average single-level test scores for the two language groups
(% mean scores and standard deviations for the test at the given level)

Differences in average test scores are statistically significant for P5 Levels B (p <= 0.01) and C (p <= 0.05), and for P7 Level D (p <= 0.001). Generalizability coefficients for relative group differences, which take task sampling into account, are 0.9 for Level B, 0.7 for Level C, 0.7 for Level D at P5.

Applying the 65% cut-off score to the test results produces the level attainment proportions shown in Table 19. Again, with the exception of Level E, we see that consistently higher proportions of Gaelic-medium pupils compared with English-medium pupils succeeded in reaching the criterion cut-off score for 'well-established' knowledge and understanding in science.

Table 19
Level attainment proportions* for the two language groups
(% pupils reaching the 65% cut-off score on the test at the given level)

The proportions shown here are of pupils reaching the criterion score of 65%, who can therefore be assumed to be well-established at the given level or higher

The distributions for the two groups over levels, with levels determined as explained earlier (in the section on 'Pupils' test-based level attainments'), are shown in Table 20 and illustrated in Figure 5.

Table 20
Science test outcomes for the two language groups
(% pupils put at the different levels by the SSA tests)

The Gaelic-medium pupils demonstrated slightly better performances than the English-medium pupils. For example, while 48% of the English-medium pupils demonstrated attainment at Level B or higher, the proportion of Gaelic-medium pupils doing the same was 52%. And at P7, while 37% of the English-medium pupils were classified at Level C or higher, the corresponding proportion of Gaelic-medium pupils was 43%.

Figure 5
Test-based level distributions for the two language groups
(% pupils put at the different levels by the SSA tests: 192 P5 pupils per group, 166 P7 pupils per group)