7 of legislative texts
7.7 Test on functional comprehensibility .1 Method
7.7.3.1 Effects on comprehensibility
The main aim of the test was to measure the effect of the change of per-spective. We will first consider the results attained by the three groups tested for the fourteen questions. Correct response frequencies for the questions are shown in Figures 7.2 and 7.3. Figure 7.2 shows the fre-quencies for those who had the original law-text (-O): EMP-O, TU-O and LAW-O, Figure 7.3 for those who had the alternative text (-A): EMP-A, TU-A and LAW-A. The questions are listed in the order they were pre-sented in the test. The subjects were instructed to answer the questions
in this order, and the lower frequencies for the latter questions therefore partly reflect the lower response frequency for those questions.
As we can see from Figure 7.2, the LAW group had the highest aver-age percentaver-age of correct responses, 71.7 per cent, among those who were given the original law-text. This group had the highest percent-age of correct responses for all questions except one (no. 9). If we com-pare the two non-law groups, we find that TU-O had a higher correct response frequency than EMP-O for most of the questions. The average for TU-O was 37.2 per cent and for EMP-O 29.2 per cent. The average difference between these O-groups was thus 8 per cent.
A comparison of the average results listed below the percentages in Figures 7.2 and 7.3 shows that in all three test groups those who
100 Correct response percentage
%
90 80 70 60
40 50
30 20 10
1 Average percentage Question no.
EMP-O 29.2%
EMP-O = Employed persons given the original version.
TU-O = Trade union officials given the original version.
LAW-O = Law students given the original version.
EMP-O TU-O LAW-O
TU-O 37.2%
LAW-O 71.7%
2 3 4 5 6 7 8 9 10 11 12 13 14
Figure 7.2 Correct response frequencies for the 14 test questions achieved by EMP-O, TU-O and LAW-O (Figure 2 in Gunnarsson, 1984: 91)
had the alternative text had higher correct response frequencies than those with the original text. LAW-A had an average of 79.4 per cent, which is 7.7 per cent higher than LAW-O. The TU-A group had an aver-age of 43 per cent, i.e. a gain of 5.8 per cent. EMP-A had an averaver-age of 40.2 per cent, an improvement of 11.0 per cent.
As we can see from Figure 7.3, the LAW-A group had the highest percentage on all questions. Even with the alternative text, their com-prehension was greater than for the two non-law groups. The correct response frequencies of EMP-A and TU-A can be said to have been on the same level. EMP-A had an average of 40.2 per cent and TU-A 43.0 per cent. The average difference between these two groups was therefore only 2.8 per cent, less than that between EMP-O and TU-O.
If we consider the individual questions, we find that TU-A were better on seven questions while EMP-A were better on six.
100
EMP-A = Employed persons given the alternative version.
TU-A = Trade union officials given the alternative version.
LAW-A = Law students given the alternative version.
EMP-A
Figure 7.3 Correct response frequencies for the 14 test questions achieved by EMP-A, TU-A and LAW-A (Figure 3 in Gunnarsson, 1984: 9)
As for differences between the individual questions, the groups showed similar tendencies. On the whole, all three groups found the same questions easy and difficult, in spite of their different back-grounds. There were, however, a few exceptions. The most noteworthy was question no. 9, where the TU group had a remarkably high fre-quency (TU-O higher even than LAW-O). For this question, which was ‘Has the trade union any way of postponing the shut-down?’, the pre-knowledge specific to the trade union officials was thus import-ant. The answer to this question is to be found in § 14 (cf. 6 above) and it can also be noted that the changes made in the alternative text led to a substantial improvement for the law students, LAW-A was 28.9 per cent higher than LAW-O.
Another way to analyse the results is to compare the average number of correct answers, incorrect answers and non-responses, for all sub-jects and for the groups separately. This comparison points to a higher average number of correct responses among those who were given the alternative text, than among those who had the original law-text, and a corresponding decrease in the average number of incorrect responses.
The improvement was greatest in the EMP group, second highest in the LAW group and smallest in the TU group. For all subjects, the differ-ences between the alternative text and the original text were significant at the 1 per cent level, as far as the average numbers of both correct and incorrect answers were concerned. The improvement was, however, not very great. The changes made in the alternative text can best be seen as a step towards a comprehensible law-text. From a further scrutiny of the results, I have come to the conclusion that the alternative text seems to have compensated the EMP group and the LAW group for their lack of trade union experience. It did not, however, compensate the two non-law groups (EMP and TU) for their lack of legal knowledge, or for their lower education. Although the alternative text led to fewer non-responses, the average number of incorrect answers hardly changed.
The alternative text was comprehended better, but not more quickly.
These results indicate that the alternative text led to a higher response rate and to more correct answers. The function-centering of the alternative text as a whole had a positive effect for the groups tested. I have, however, also analysed the results in relation to the hypotheses made regarding the importance of the different text levels underlying this function-centring.
As we already know, the alternative text was elaborated system-atically on the basis of an analysis of pragmatic levels, the content level and the text-structural level. Syntax, vocabulary and typography were used as auxiliary means of realizing the function perspective.
The original law-text, however, was not written consistently, as was
the alternative text. The differences between the two texts, therefore, vary from rule to rule. I examined this variation between the texts by means of a rule-by-rule analysis of text differences on the following levels: content, text-structural, syntactic, anaphoric, lexical and typo-graphical. This comparison was then used for a detailed analysis of the effect of different text levels on comprehensibility.
I will not elaborate any further on this analysis here, only mention that the results of these detailed analyses support the assumption that function-centering primarily based on changes at deeper text levels leads to improved comprehensibility as far as action-oriented law-text reading is concerned.