Peeling Hologram Off Licence

Warming

As you know, and if not, I tell you, on December 7 out PISA 2009. PISA is an international assessment piloted by the OECD which measures students' knowledge of 15 years in three skills: reading, math and science. In this edition Star competition is reading, and involves over 60 countries under review no less than about half a million students.

What results is going to get Spain in PISA 2009? Well, basically, the same as in previous years. Why? Because little has changed in education English years, except in a more savvy management. Changes in the LOE (2006) on the Logse (1990) were minimal, and its effects on performance will be also. We will analyze the changes we can expect to find in the English results in PISA 2009 than in PISA 2006.

1. A technical improvement.

In PISA 2006 there was a problem with reading data from English. Specifically, the tests written in Castilian (and therefore does not affect Catalonia, Catalan pass the screening to all students, and to a lesser extent Galicia and Basque Country) was an issue that the results of the scale reading were about 20 points lower right. What is the problem, I do not know for sure, as I have heard two versions: a problem with the hardness of the editors and a problem with an item in a pilot test. The second seems more plausible, but as both have similar consequences and I know I have not yet measured enough to unravel the issue (although I lack less), if you can, because I can read here.

What I do know is that it is a technical problem that the Institute of Evaluation (IE) has been aware at least a couple of years, and on that basis have all worked. I hope that IE does not throw the bells ringing about the alleged improvement when it is perfectly aware of the technical problem, and put to say what good the LOE. [For example, in the English Report PISA 2006 (126 pp.) Table does not appear in any single previous data: see this].

Therefore, an improvement in the level of reading, say, 15 points in PISA 2009 than in PISA 2006, and that is accompanied by improvements in Math and Science scales is talk of a null effect the measures taken so far. And a minor improvement a complete failure because there will not be broken the downward trend in the Reading area since 2000. I understand that, being the main scale reading in 2009 and caring a fig for our political reality, the Ministry will not be able to resist trying to take advantage of this, but at least you can compare the results with PISA 2003 before take them seriously.

2. Have we reduced the yield somewhat lower PISA?

Almost the only measures introduced by the LOE that may affect the performance of pupils aged 15 are the introduction of large-scale educational reinforcements and little diversification posed by PCPI. The PCPI (Initial Vocational Training Program) is a diversification for students 15 and older who have repeated and it is judged that the normal way will not get a title. It's a way designed (badly, of course) to facilitate access to the title or at least, to give students a minimum qualification before they leave school. Probably the best we can expect from this measure is that students FPGM reinstatement through the entrance exam. But, of course, we can not expect the PCPI increase students' skills in reading, math or science, which is what PISA measures.

The second measures the extent of the reinforcements across the BOW program, it can have an effect on performance. It is a program that has been applied unevenly in each of the autonomous regions, and therefore the effects on performance will be uneven. Indeed, some have used the program to give titles, and others so that students know more, and that is going to be very clear in PISA 2009. As you recall, the plans to strengthen achieved a reduction of school failure in 2008 , but it is doubtful that in all regions reinforcements are used to increase the knowledge of pupils, leading them to obtain the title- but to give the title of ESO, ie, to make up statistics.

That is, if the reinforcements have worked well in PISA you notice a slight overall improvement, improvement that may be because students with low performance level would have a level of mediocre performance. It would also have an effect on the standard deviation, which some call ( bad ) equity. That is, communities which cut through the failure of reinforcements know who did well and who tried to deceive us.

3. Reinforcements Primary Reading

Some communities began years ago to realize that many students had reading problems in primary, and took steps to correct it by special reinforcements. And, within that group, some take longer than others. To my knowledge, only one, small scale, have been carried out to affect students in 2009 were 15 years old. That is, little to do with this issue on that side, but maybe next we see its fruits.

4. Other problems

Spain has other issues on which nothing is done. The main ones are repetition and low proportion of students at higher levels. There is no national plan to tackle these problems and you can see in PISA 2009, so the results will be exactly the same as in other years, except in communities where they had specific plans for improvement on these issues. Although some still or have been aware, only fools and madmen can expect to get a different result after doing the same thing.

5. The country of fairness to the country of "multi-speed"

Although there are many who celebrate the English equity, we actually have a country where being born in a community or another has a significant influence on the performance of students. If this has not been seen in PISA is because they were not going well not show up. In the General Assessment of Primary Diagnosis and find any indication of what I'm saying, but what of PISA 2009 will be obvious. Even though there are three communities, Castilla-La Mancha, C. Valencia and Extremadura, all featuring a high-school failure who had refused to appear in this issue.

What evidence we have of it? Then seeing the results of the Autonomous Communities in PISA 2006. Both in 2006 and 2003 was possible to identify the centers of each CCAA and make an approximation of the results. Naturally, the sample is smaller and the selection of centers may not be representative of the region, but was remarkably stable between 2003 (there were four communities) and 2006 (where there were ten), and therefore I hope that you in 2009. On Tuesday we will check, but for now, I offer the results of all English regions in 2006 and also the correction of ISEC (sometimes, the selection of schools can make the selected sample has a higher or lower socioeconomic status than the population of the Community, and this fix avoids). Let

first with the table of students tested in each CCAA (there are a few sites not allocated, which are labeled as "Unknown"):

Error Error Typical % 0.06 2.0 Baleares Canary 0.24 4.2 17,296 Castilla y León Ceuta and Melilla Navarra 147 Unknown 0.92 Basque

Students examined in PISA 2006, by C. Autonomous
	Sample Population			Weight
	students	student		each CCAA	typical
		No No	No	%
Andalucía		1,463 2,506 81,437			21.3 0.58
Aragon		1,526 9,467 318			2.5 0.09
Asturias		1,579 7,594 197
		123 847 5868			1.5 0.22
		15,868 206 899
Catalonia		1,527 1,787 56,987			14.9 0.46
Castilla-La Mancha	226			907 0.23 4.5
Cantabria		1,496 4,534		98 0.03 1.2
		1,512	19,697 419 5.2		0.12
		108 1441 74		0.4	0.02
Extremadura		141 660 9343		2.4	0.17
Galicia	1573		22,578 365 5.9		0.12
Madrid	525		43,847 1,018 11.5		0.26
Murcia	174		11,843 368 3.1		0.11
	1,590 4,678			67 0.02 1.2
			12,931 3,613 3.4
		14,707 3,929 203		0.08 3.9
	La Rioja 1,333 2,494			13 0.7 0.01
C. 426 Valencia			39,077 1,136 10.2
0.30 SOURCE: Compiled from data from PISA 2006.

The first column indicates those students who are actually evaluated, the second points to the population they represent, and the fourth the percentage of students in each community on the whole of Spain. Typical errors are those with all statistical PISA, and we're talking about samples.

The following tables show the results in each of the scales in PISA 2006: Science, Reading and Mathematics. As usual, the data from small samples have a higher standard error.

4 Navarra 508 1579 9 Andalucía 474 14 11.8

scale results in PISA 2006 Science, by AC
		Valid	Media	ET
1	Castilla y León		1,512 520 3.9
2	La Rioja		1,333 520 2.5
3	Aragon		513 1526 3.9
			511 1,590 2.9
5	Cantabria		1,496 509 3.6
6	Asturias			4.9
7	Madrid		506 525 11.7
8	Galicia		1,573 505 3.4
	Unknown		147 23.0 500
	Murcia		174 500 6.3
10	Basque		3,929 495 3.5
11	Catalonia	1,527 491		5.1
12	Castilla-La Mancha		226 489 11.9
	SPAIN		19,604 488 2.6
	13 1463			4.0
	Canary		206 472 19.2
15	C. Valencia		468 426 6.4
16	Extremadura		141 452 4.5
17 Ceuta and Melilla		437 108
18	Baleares		123 433 20.7
Source: from data from PISA 2006.
Notes: Valid is the number of students tested with valid values \u200b\u200bfor this variable. Media is the average score of students in each community on the scale of Sciences. ET is the standard error associated with that media.

ET 1 13 14

Results on the PISA reading scale in 2006, by AC
		Valid	Media
	Rioja	1333	492	2.6
2	País Vasco	3929	487	4.2
3	Aragon		483 1526	5.2
4	Navarra	1590	481	2.7
5	Galicia	1573	479	3.4
6	Castilla y León	1512	478	3.4
	Unknown		478 147 20.9
7	Asturias		1,579 477 4.7
8		Catalonia	1,527 477 5.1
9	Cantabria		1,496 475 4.0
10	Madrid		525 473 8.9
11	Castilla-La Mancha	226 463		15.3
	SPAIN		19,604 461 2.2
12	Murcia		458 174 8.0
	Andalucía		445 1463 4.1
	Canary		443 206 11.9
15 Ceuta and Melilla			434 108 9.3
16	C. Valencia		434 426 5.4
17	Extremadura		419 141 7.1
18	Baleares	413 123		13 , 9
Source: Compiled from data from PISA 2006.
Notes: Valid is the number of students tested with valid values \u200b\u200bfor this variable. Media is the average score of students in each community in the scale reading. ET is the standard error associated with that media.

ET 3 Navarra Andalusia 17

Results on the mathematics scale in PISA 2006, by AC
		Valid	Media
1	La Rioja		1,333 526 2.2
2	Castilla y León		1,512 515 3.3
			515 1590	3.5
4	Aragon	1526	513	4.5
5	Cantabria	1496	502	2.6
6	País Vasco	3929	501	3.4
7	Asturias	1579	497	4.9
8	Madrid	525	496	10.5
9	Galicia		1,573 4.1 494
	Unknown		493 147 17.9
10	Catalonia		1,527 488 5.2
11	Castilla-La Mancha		482 226 8.4
	SPAIN		19,604 480 2.3
12	Murcia		475 174 6.6
	13		463 1463 4.2
14	C. Valencia		456 426 4.9
15	Baleares		123 449 10.8
16	Extremadura	448 141		8 , 6
	Canary		206 446 11.0
18 Ceuta and Melilla		442 108
8.9 Source: Compiled from data from PISA 2006.
Notes: Valid is the number of students tested with valid values \u200b\u200bfor this variable. Media is the average score of students in each community on the scale of Mathematics. ET is the standard error associated with that media.

These tables show the direct results, but, as we said, in communities without a broader sample, we run the risk that the selection does not represent the population. We do not have much with which to compare, but it seems clear that the socioeconomic and cultural level Extremadura is not above the English average. We can therefore say that the students selected in the sample Extremadura have a higher level than the population real Extremadura, and thus conclude that the results of the above tables are overestimated. The ISEC through each of the Communities (is an index with zero mean and standard deviation one for all students in the OECD) is as follows:

Basque 0.158 -0.094 5 0.052 -0.136 0.035 -0.141 15 0.111 -0.562

ISEC average in PISA 2006 CCAA
		Valid	Media	ET
	Unknown		147 0.024 0.228
1			3892 0.039 -0.039
2	Madrid	524
3 Balearic			121 0.281 -0.116
4	La Rioja		1331 0.026 -0.124
	Aragon 1520
6	Navarra 1586
7	Catalonia		1,514 0.080 -0.146
8	Asturias		1,573 0.068 -0.150
9	Cantabria		1,483 0.047 -0.162
10	Castilla y León		1,508 -0.212 0.046
11	Extremadura		140 0.135 -0.244
	SPAIN	19,499	-0.311 0.031
12	C. 423 Valencia		-0.320 0.087
13	Galicia	1566	-0.341 0.066
14	Ceuta and Melilla	105	-0.344	0.120
	Canary		205 0.275 -0.381
16	Murcia		174 0.216 -0.542
17	Castilla-La Mancha	226
18	Andalucía		1461 0.056 -0.635
Source: Compiled from data from PISA 2006.
Notes: Valid is the number of students tested with valid values for this variable. Media is the average score of Socioeconomic and Cultural Index (ISEC) of students in each community. ET is the standard error associated with that media.

But that does not help much. To know which are the net scores of each community, ie, the score would be if the students had a half ISEC equal to the average of the OECD, is the following table. Data are linear regression coefficients and their significance is in the notes to the table. As it is not something that has an average reader to understand, the most basic is this: the first column (B0) is the net score, which average should be 500, and the last (R2) is the percentage change in the results of the ISEC students are able to explain. In short, this is the table:

ET B1 12.0% 2.2 13.2% 2.2 9.5 21.4% 9.2 14.9% 2.8 Extremadura 7.9% 7.8

Net gain on the scale of Sciences PISA 2006, by AC
		B0			ET	R2
1	Castilla y León	525		3.6 3.0 23		8.4%
2		La Rioja 523 2.4			25 8.4% 2.8
3	519 Aragon		3.3		32 14.4% 2.3
4	Navarra		516 3.0 30
5	Cantabria		516 3.0 32
6	Asturias	514		3.5 32 2.5		15.5%
7	Galicia	514		3.0 25 2.7		8.9%
8	Murcia	512		5.8 21 3.9		10.8%
9	Madrid	510		7.0 36 5.7		16.3%
10	Castilla-La Mancha	505			26 8.3% 7.8
	Unknown		501 17.3 40
		499 SPAIN 1.9			31 13.7% 1.3
11	Basque		497 2.8 28			10.9% 2.0
12	Catalonia		497 4.0 31			13.5% 3.1
13	Andalucía		494 3.1 32
	Canary 14 484		13.2		29 14.8% 4.6
15	C. 480 Valencia		6.5		30 11.3% 3.7
	16 464			6.2 9.3 37		19.0%
17 Ceuta and Melilla			445 13.0 22
18	Baleares		441 18.3		34 11.9% 21.2
Source: Compiled from data from PISA 2006.
Notes: B0 is the net value of the performance of students in each CCAA on the scale of Sciences. That is, the average score of students in each community if everyone had an ISEC zero or, in other words, if the conditions commo cultural and family were equal in all regions. B1 is increased score associated with an increase of one unit (in this case, one standard deviation) in the ISEC. R2 is the percentage of variation in student achievement explained by ISEC of each student. ET is the standard error associated with each coefficient.

So far, the predictions. On Tuesday morning, the moment of truth.