Total
100.00
80.00
60.00
40.00
Year
Number Discontinued
Percentage Discontinued
2007 21 29 72 2008 22 30 73 2009 55 47 118 Table 2: Students who discontinued study
Total number of all Students
20.00
. 00
Figure 4: Boxplots of the total marks for each year
Count
25
20
15
10
5
0
Figure 5: Student grades for each year Grade
Percent
HD DI CR P
40.0 %
30.0 %
20.0 %
10.0 %
0 %
Fail
Fail
Figure 6: Percentage of grades for each year
Year 2007 2008 2009 appears lower in 2009. This is confirmed by the ANOVA test that showed that the mean total mark for 2009 was significantly lower than 2007 and 2008.( F = 6.312, p = 0.002)
Having found a highly significant result, Bonferroni multiple comparisons were done. A difference was found between the 2007 and 2009 results( p = 0.043), and the 2008 and 2009 results( p = 0.003), but not between the 2007 and 2008 results( p = 1.000).
A similar comparison was done for all the assessment items revealing a similar trend towards lower marks in 2009, however not all of these were statistically significant. Only analysis of the total mark is reported here because it is a measure of all the assessment items.
The highly significant decrease in the mean total mark in 2009 occurred at the same time that contact hours were reduced. This resonates with the views of Nettelbeck et al.( 2007) who suggest that a decrease in contact hours leads to a decrease in proficiency. The proficiency of the students who passed the beginners’ Japanese unit in 2009, as recorded by their final mark, is on average lower than those students from 2007 and 2008. This may have a flow-on effect and may eventually contribute to the trend of fewer people graduating with advanced language skills in Australia( as noted by Pauwels, 2002).
Question 3 Were the overall grades lower in 2009?
Figure 5 is a clustered bar chart showing the number of students in each grade category for each year. We can clearly see that the distribution between grades in 2007 and 2008 are similar while 2009 is completely different. The number of students obtaining High Distinctions( 85 – 100 %) has noticeably reduced as a proportion of the grades in 2009 but the number of other grades, in particular the Pass grade( 50 – 64 %) has markedly increased.
In Figure 6, the distribution of grades as a percentage per year is shown. What is striking is that in 2009 the High Distinction( HD)( 85 – 100 %) marks have decreased to approximately 50 % of the 2007 and 2008 proportion of HDs. At the other end of the scale, Pass marks( 50 – 64 %) make up a much larger proportion of the marks, in fact, more than four times as much as the average number of the 2007 and 2008 Pass marks.
A Chi-Squared test confirms that the differences in the clustered bar charts in Figures 5 and 6 above are statistically significant. It shows that the grade and year are dependent.( x 2 = 16.768, df = 8, p = 0.033)
To further see the difference between the class hours the 2007 and 2008 class results were amalgamated in each of the grade categories and compared to the 2009 ones.
A significant difference between 2007 and 2008 combined and 2009 was also found thus showing that class contact hours was linked to grade.( x 2 = 14.693, df = 4, p = 0.005)
The significant decrease in grades in 2009 shows that the decrease in mean total mark was large enough to move a large proportion of students from one grade category to the next lower grade category. Overall, there were approximately 20 % fewer High Distinction grades and 15 % more Pass grades.
28 BABEL