Comparing scores to ielts



Yüklə 232,96 Kb.
Pdf görüntüsü
səhifə4/7
tarix04.06.2022
ölçüsü232,96 Kb.
#88747
1   2   3   4   5   6   7
461626-cambridge-english-qualifications-comparing-scores-to-ielts (1)

6
5
4
3
2
1
0
C
B2 First
B
B2 First
A
B2 First
C
C1 Advanced
B
C1 Advanced
A
C1 Advanced
4
IELTS
5
IELTS
6
IELTS
7
IELTS
8
IELTS
9
IELTS
Mean self-r
ating (logits)
Can Do self-ratings and grades
7


Cambridge Assessment English
Comparing scores to IELTS
5
In 2009, we undertook to benchmark Level C1 as represented by C1 Advanced against IELTS scores. For this 
exercise an empirical validation study was undertaken, where registered IELTS candidates were invited to also
take C1 Advanced, and registered Advanced candidates were invited to take IELTS, and their scores compared. 
This counterbalanced design accounted for preparation or motivation-related effects on one exam or the
other. As C1 Advanced targets the higher end of the IELTS candidature population, participants’ performance
was on average higher than that of the global IELTS candidature, as expected. Correlations between scores on
the two exams were calculated to see how related the two tests are. The correlations between the different
parts of the two exams are generally moderate, whereas the correlation for the overall scores is, as might be 
expected, stronger. 
To compare results on the two exams, the equipercentile linking method was used, and pre-smoothing using
the polynomial log-linear method (Holland and Thayer 2000) was employed to increase the precision of the 
linking. This method was adopted because indices are available for evaluating goodness of fit and appropriateness 
of the linking (Kolen and Brennan 2004). Because smoothing resulted in C1 Advanced scores that were not 
integers, linear interpolation was used to determine IELTS raw marks that corresponded to CEFR Levels B2, C1 and 
C2 on each of the four skills, and standard conversion tables were used to express the outcomes in terms of the 
nine-band IELTS scale. Classification consistency between the two exams on the three levels and across the
four skills averaged 80%. 
In the meantime, the IELTS partners had approached Chad Buckendahl of Alpine Testing Solutions to lead a 
standard-setting study aligning IELTS bands to the CEFR levels. The standard-setting study involved 19 panellists 
using two different standard-setting methods for the four papers that comprise IELTS. For Speaking and Writing,
a modification of the Analytical Judgment method (Plake and Hambleton 2000) was used. Panellists were asked 
to read samples of writing and view samples of speaking, and to classify each into appropriate CEFR levels
which was subsequently refined to identify performances at the border of each level. These judgements were 
then replaced by the original score that those performances received to arrive at the cut score. For Listening and 
Reading, the Yes/No variation of the Angoff (1971) method (Impara and Plake 1997) was adopted. This standard-
setting project is discussed further in Lim, Geranpayeh, Khalifa and Buckendahl (2013). 
Our advice as to the alignment of C1 Advanced scores and IELTS bands is therefore based on the results of the 
external validity study comparing IELTS and C1 Advanced performances, with supporting evidence drawn from
the 2009 IELTS standard-setting project and earlier studies. 


6

Yüklə 232,96 Kb.

Dostları ilə paylaş:
1   2   3   4   5   6   7




Verilənlər bazası müəlliflik hüququ ilə müdafiə olunur ©genderi.org 2024
rəhbərliyinə müraciət

    Ana səhifə