using the idb analyzer iv - european commission · 1 using the idb analyzer iv andrés...
TRANSCRIPT
1
Using the IDB Analyzer IV
Andrés Sandoval-Hernández – IEA DPC
Workshop on using PISA, PIAAC, TIMSS & PIRLS, TALIS datasets
Ispra, Italy- June 24-27, 2014
Note: These slides were prepared as part of the IEA training portfolio with the collaboration of IEA staff and resource persons.
2
Table of contents
• Correlations with TIMSS and PIRLS combined
• Regression with PIAAC
3
Correlations
How strong is the correlation between
achievement in reading and achievement in
mathematics?
ASMMAT1-5 ASRREA1-5
Variables of interest
4
What Do we Need to Do?
• Merge the data set for TIMSS and PIRLS 2011 by choosing the student background questionnaire and the selected countries
• Countries in the merged data file:
– Austria
– Italy
– Poland
– Slovak Republic
– Spain
• Open Analysis Module
5
Analysis: Correlations with TIMSS and PIRLS combined
• Analysis File: C:\EU_Workshop\Work\TIMSS_PIRLS_Merged.sav
• Analysis Type:
TIMSS/PIRLS (Using Student Weights)
• Statistic Type: Correlations
• Plausible Value Option: Use PVs
• Grouping Variables: IDCNTRY
• Plausible Values: ASMMAT1-5, ASRREA1-5
• Weighting variable: TOTWGT
6
Correlations Analysis
C:\EU_Workshop\Work\TIMSS_PIRLS_Merged.sav
C:\EU_Workshop\Work\ASMMAT_ASRREA.*
Analysis Type:
TIMSS/PIRLS (Using Student Weights)
Statistic Type:
Correlation
Plausible Value Option:
Use PVs
IDCNTRY
ASMMAT01-05
ASRREA01-05
7
Correlations Analysis
Start SPSS and run syntax
8
Country = Austria
.78 (.01)
There is a significant positive correlation between the achievement in math and achievement in reading for the students that participated in TIMSS and PIRLS 2011, grade 4.
Correlations Outcome
9
Correlations Outcome: Correlations
Correlation between ASMMAT1-5 and ASRREA1-5 per country
Standard error of correlations
10
Correlation Outcome: Descriptive
Mean and standard deviation for the variables
Standard errors of the mean and standard deviation
11
Table of contents
• Correlations with TIMSS and PIRLS combined
• Regression with PIAAC
12
Regression with continuous variable
To what extent does the level of education
explain differences in problem solving skills ?
ISCED_HF_C PVPSL1-10
Variables of interest
13
What Do we Need to Do?
• Merge the data set for PIAAC for the selected countries
• Countries in the merged data file:
– Austria
– Denmark
– Netherlands
– Slovak Republic
– Spain
• Open Analysis Module
14
Analysis: Regression with continuous variable
• Analysis File: C:\EU_Workshop\Work\PIAAC_Merged.sav
• Analysis Type:
PIACC (Using Full Final Sample Weight)
• Statistic Type: Regression
• Plausible Value Option: Use PVs
• Grouping Variables: CNTRYID
• Independent Variable: ISCED_HF_C
• Dependent Variable: PVPSL1-10
• Weighting variable: SPFWT0
15
SPSS: Run the analysis
16
Differences btw students belonging to schools of ‘Very high’ (ACDGEASH=1, the reference group) and ‘Medium’ (ACDGEASM=1) school emphasis on academic success in math achievement
Analysis : Results
17
Regression: Results
In AUSTRIA each additional education level attained is associated to 4.09 score points in the scale of problem solving. This association is statistically significant (α=0.05) as t-value > 1.96
* Check descriptive statistics!
18
Any Questions?
Thank you for your attention!