TIMSS 2006/07: Trends in Year 5 mathematics achievement 1994 to 2006

Publication Details

This report describes the mathematics achievement of Year 5 students in TIMSS 2006/07. Trends in New Zealand’s achievement over the 12 years from 1994 to 2006 are examined, along with comparisons with other countries. Analyses of achievement by sub-groupings (such as gender and ethnicity) and background information are also presented. It was originally published in December 2008 and revised in September 2009 due to the mislabelling of the content domains knowing and applying. The current version rectifies this error.

Author(s): Robyn Caygill & Sarah Kirkham [Ministry of Education]

Date Published: December 2008

Mathematics achievement by ethnicity, language, and country of birth

This section will examine the mathematics achievement of students in TIMSS across different ethnic groups, by use of English at home, and by country of birth. These three characteristics of students are interrelated so in the final part of this section they are examined together. This section will examine relationships with mathematics achievement, but it should be noted that the existence of a relationship does not infer a causal link.

Mathematics achievement by ethnicity

Five broad ethnic classifications are used to describe ethnicity in New Zealand. They are: Pākehā/European, Māori, Pasifika, Asian, and ‘Other’ ethnic groupings. The majority of Year 5 students in New Zealand were identified by their schools1 as Pākehā/European (61%) or Māori (19%). Pasifika (10%) and Asian (7%) students made up most of the rest of the ethnic groupings, with four percent of students categorised in the Other ethnic grouping.

Previous cycles of TIMSS have shown that average mathematics achievement varies across ethnic groups. Although the variation in achievement is not caused by ethnicity per se, education policies have been introduced so that all students may realise their potential. Specific areas of focus for the Ministry of Education include the achievement of Māori and Pasifika students (Ministry of Education 2007). The results at the Year 5 level in TIMSS 2002/03 (Caygill, Sturrock, & Chamberlain, 2007) showed an increase in mathematics performance, on average, for Māori and Pasifika students since the first cycle in 1994/1995.

In TIMSS 2006/07, Asian (546) and Pākehā/European (510) students had significantly higher mean mathematics achievement than did their Māori (453), Pasifika (427) and Other (491) counterparts, as shown in Figure 9. On average, Asian students performed significantly higher in mathematics than Pākehā/European students and Māori students performed significantly higher than Pasifika students.

The distribution of achievement of Asian students and those in the Other ethnic grouping was the widest, while the distribution for the Pākehā/European students was the narrowest. Note that the 5th and 95th percentiles of achievement for the students in the Pasifika, Asian, and Other ethnic grouping should be treated with caution as there are few students at these ends of the distribution due to the smaller number of students in these ethnic groupings.

Figure 9: Distribution of New Zealand Year 5 mathematics achievement for each ethnic groupingImage of Figure 9: Distribution of New Zealand Year 5 mathematics achievement for each ethnic grouping.
Notes:
  1. The distribution and ranges for the students in the Pasifika, Asian, and Other ethnic groupings should be read with caution as there is a lot of uncertainty at the extremes of the distribution.
  2. Standard errors are presented in parentheses

Benchmarks for ethnic groupings

Within all ethnic groupings, there were students who reached the advanced benchmark; in terms of the benchmark definitions, they demonstrated the ability to complete tasks requiring applying their "understanding and knowledge in a variety of relatively complex situations and explain their reasoning". Similarly, within all ethnic groupings there were students who did not reach the low benchmark; that is, they did not demonstrate the ability to complete a reasonable number of the simplest mathematics tasks which TIMSS seeks to measure.

Higher proportions of Asian and Pākehā/European students reached the advanced benchmark compared with each of the other ethnic groupings (as shown in Table 14). Around one-third of Pasifika students did not reach the low benchmark, while a quarter of Māori students did not reach this low benchmark.


Table 14: Proportion of Year 5 students reaching each international benchmark by ethnic grouping
Ethnic grouping
Percentage of Year 5 students reaching each benchmark
Advanced
High
Intermediate
Low
Pākehā/European
5 (0.8)
32 (1.3)
70 (1.1)
91 (1.0)
Māori
1 (0.6)
12  (1.5)
41 (2.4)
75  (2.5)
Pasifika
1 (0.6)
7 (1.1)
28 (2.6)
62 (3.1)
Asian
19 (1.9)
50 (3.1)
78 (2.8)
95 (1.5)
Other
5 (1.8)
27 (3.8)
60 (3.4)
85 (3.4)
Note:
  1. Standard errors are presented in parentheses.

Another way of looking at this information is to examine the composition of the group who did not reach the low benchmark. Fifteen percent of New Zealand students did not reach this benchmark as shown in Figure 10. The majority of these students were Pākehā/European (5.3%) or Māori (4.7%). However, Māori and Pasifika students were over-represented in this lower-achieving group compared to their proportions in the population.

Figure 10: Ethnic composition of the students who did not reach the low benchmarkImage of Figure 10: Ethnic composition of the students who did not reach the low benchmark.
Note:

  1. The values presented in the pie chart are proportions of the whole population and therefore add to 15%, the proportion of students in the ‘below low’ group.

Trends in mean mathematics achievement for ethnic groupings

Pākehā/European, Māori, Pasifika, and Other students all demonstrated significant gains in mathematics achievement, on average, between 1994 and 2002. Asian students showed a positive change over the eight years between 1994 and 2002 but because of the large standard errors involved, this is not statistically significant. While between 2002 and 2006, the average performance of Māori students decreased significantly, the average mathematics achievement of Māori students in 2006 was significantly higher than in 1994. Between 2002 and 2006, the average performance of Pasifika students returned to the lower level of achievement observed in 1994. Asian students continued to show significant gains between 2002 and 2006, making a total gain of 62 scale points for Asian students between 1994 and 2006, a greater increase than for the remaining ethnic groups (see Table 15).

Table 15: Trends in mathematics achievement over four cycles of TIMSS by ethnic grouping
Ethnic grouping
Mean mathematics achievement
Change
 1994 to 2006
1994
1998
2002
2006
Pākehā/European
493 (3.9)
502 (5.0)
506 (2.7)
510 (2.1)
17 (4.4)
Māori
427 (8.2)
445 (7.3)
479 (4.8)
453 (4.4)
25 (9.3)
Pasifika
412 (11.0)
416 (15.1)
464 (6.3)
427 (5.1)
15 (12.2)
Asian
483 (16.9)
516 (9.9)
500 (6.0)
546 (4.9)
62 (17.7)
Other
475 (15.1)
481 (14.8)
504 (9.8)
491 (6.0)
16 (16.3)
Note:
  1. Standard errors are presented in parentheses.
  2. Due to rounding some results may appear inconsistent.

Mathematics achievement of boys and girls within ethnic groups

As mentioned earlier, there was no significant difference in mean mathematics achievement between boys and girls in TIMSS 2006/07. This result was also observed when gender differences were examined within each of the ethnic groups.

Mathematics achievement by regularity of English speaking at home

Most students reported that they always or almost always spoke the language of the test (in this case English) at home (87% - 74% always and 13% almost always).2 Few students (1%) reported that they never spoke English at home. Students who always or almost always spoke English at home had higher mathematics achievement, on average, than those who sometimes or never spoke English at home (see Figure 11). This pattern of higher average achievement for those who spoke English at home was also evident across the previous three cycles of TIMSS (see Caygill, Sturrock, & Chamberlain, 2007). However, it is interesting to note that the difference between these two groups of students has reduced over time from 77 scale score points in 1994 to 41 scale score points in 2006.


Figure 11: Year 5 mean mathematics scores by regularity of English speaking at homeImage of Figure 11: Year 5 mean mathematics scores by regularity of English speaking at home.
Note:

  1. The bars on the graph represent the proportions of Year 5 students while the points represent mean scores. Lines extending from the points represent the 95% confidence interval, i.e. the range within which we are 95 percent confident that the true population value lies.
  2. Standard errors are presented in parentheses.

Mathematics achievement by country of birth

Another factor that interacts with language and ethnicity is the immigrant status of the student and their parents. This information was collected in TIMSS by asking the student if they and their parents were born in New Zealand. Around one-fifth of students had neither parent born in New Zealand, one-fifth had only one parent born in New Zealand and the rest had both parents born in New Zealand. One-quarter of students were born outside of New Zealand. Of these students born outside of New Zealand, nearly half of them (44%) came to New Zealand as school-age children.

Mathematics achievement was lower for those students born outside of New Zealand, on average, compared with the New Zealand born students (38 scale score points difference). The majority of the students born outside of New Zealand were Pākehā/European in ethnic origin, as shown in Figure 12.


Figure 12: Proportions of students born out of New Zealand by ethnic groupingImage of Figure 12: Proportions of students born out of New Zealand by ethnic grouping.
Note:
  1. The values presented in the bar are proportions of the whole population and therefore add to 26%, the proportion of students in the ‘born out of NZ’ group.

Interaction of use of English at home, ethnicity and country of birth

In order to confirm the relationships between use of English at home, ethnicity, country of birth and mathematics achievement and also to see how they interacted together, multiple-regression techniques were used. Only these background characteristics were included in the investigation. The resulting statistical model showed that speaking English infrequently at home, belonging to the Māori or Pasifika ethnic grouping, and being born outside of New Zealand, were all associated with lower mathematics achievement when the other factors were taken into account.3 The model also demonstrates that all of these background characteristics were significant when explaining differences in mathematics achievement. Note that differences in achievement were smaller when the other factors were taken into account. For example, the difference between those who regularly spoke English at home and those who did not was reduced from 41 scale score points when analysed in isolation to 27 scale score points in the model. However, there are a limited number of factors included in this model. Taking into account socio-economic or educational resource factors may change this result (see the section later in this report entitled "Discussion of interactions").

Footnotes

  1. Based on enrolment information supplied by parents.
  2. In TIMSS 2006, as in 1994 and 1998, students who had the majority of their teaching in te reo Māori were excluded from the assessment. See technical notes and definitions for further details of exclusions.
  3. The model showed that when the other factors were taken into account, speaking English infrequently (-27 scale score points or ssp), Māori (-52 ssp), Pasifika (-62 ssp), born outside New Zealand (-35 ssp) were all associated with lower achievement.