Explaining Reading Achievement in PIRLS by age and SES

1
Explaining Reading Achievement in PIRLS by age
and SES

• Jan Van Damme
• Lobke Vanhee
• Heidi Pustjens
• Katholieke Universiteit Leuven (Belgium)

22nd International Congress for School
Effectiveness and School Improvement Vancouver,
Canada, January 4-7, 2009
2
PIRLS 2006

• Progress in International Reading Literacy Study
• Large-scale international assessment
• - 40 countries
• - 4th graders, 10-year-olds
• - Primary schools

3
1 Problem

• 1) Quality Average achievement
• Between countries differences in age
• 2) Equity
• There is no general SES-index

4
2 Research Questions

• 1) What is the impact of age differences on the
  achievement of different countries and/or on the
  change in achievement between different
  measurement occasions?
• 2) Is it possible to create an acceptable
  SES-index across countries?

5
3 Hypotheses

• 1) Age effect
• a) We expect that the increase or decrease in
  average age is an important explanation
  of the progress or decline of a country
  in reading achievement.
• b) We expect that a correction for age yields a
  different ranking of countries according to
  their average achievement.
• 2) SES

6
4 Method

• Age effect
• a) The change between PIRLS 2001 and
• 2006 at country level correlations
  between change in achievement and change in
  age
• b) Ranking
• Rindermann, 2007
• gain of students in one year 42 points
• Cliffordson Gustafsson, 2007
• 1/3 or 14 points age effect
• 2/3 or 28 points grade effect

7
4 Method

• 2) SES-index
• - indicators
• books at home, childrens books at home,
  educational aids, home belongings, financially
  well- off, occupational status, highest
  educational level, employment situation
• - linear regression analysis proportion
  explained variance of a single SES-variable
  within each country
• - PCA across countries books at home,
  educational aids, occupational status, highest
  educational level

8
5 Results

• 1) Age effect
• a) Average age and average achievement in PIRLS
  2001 and 2006 correlations at the country
  level

9
5 Results

• 1) Age effect examples

10
5 Results

• 1) Age effect
• b) Ranking uncorrected

11
5 Results

• 1) Age effect
• b) Ranking corrected

12
5 Results

• 2) SES-indicators

13
5 Results

• 2) SES PCA

Total variance explained by each component of the
principal component analysis
14
5 Results

• 2) SES PCA

Component Matrix
15
5 Results

• 2) SES-index country average

16
5 Results

• 2) Regression of PVs on SES R²

17
5 Results

• 2) Regression of PVs on SES

18
5 Results
19
5 Results
20
6. Conclusion and discussion

• We demonstrated the importance of age and
  schooling on the ranking of countries concerning
  their achievement level.
• But the correction made is only justified if we
  can assume that aging by one year has more or
  less similar effects in every country.
• Can we confirm this proposition?

21
6. Conclusion and discussion
22
6. Conclusion and discussion
23
6. Conclusion and discussion
24
6. Conclusion and discussion
Grade and age effects for Sweden, Iceland and
Norway