The effect of book stocks per pupil on Key Stage 2 test results
Are low levels of book spending in primary schools jeopardizing the National Literacy Strategy? is an article in the March 2006 issue of the Curriculum Journal, by Steve Hurd, Malcolm Dixon and Joanna Oldham. (Some readers will be able to access the full text of this article through their Athens account.)
The article analyses data to show that there is a positive correlation between spending on books (and a smaller, but positive, correlation between spending on ICT) and the acquisition of literacy by primary school pupils. Citing evidence from secondary schools that there is competition between spending on books and spending on ICT, the article suggests that it is this which explains why book-spending has been falling whilst learning-resource spending has been rising overall. Here is an extract:
Book expenditure per pupil is the most important resource variable accounting for differences in test performances between schools. It has a positive effect which rises directly in line with the level of spending. Every £1 spent on each child on books raises average test results by 0.004. This implies that if we raised average spending on books by £100 per child then average test results would rise by 0.4 points (or by 1.5% per child). The effect of variations in the school book stock is shown in Figure 3. (At the top of this post.) The overall effects on standards are small. However, they suggest that standards can be raised by increasing the number of books to about 63 per child, more than double the current level of 25 books per pupil. ICT expenditure per pupil also has a significantly positive effect on pupil performance. For every £1 spent on ICT per child the average test score rises by 0.002, so increasing spending by £100 per child raises average test scores by 0.2 of a point (a rise of 0.72% per child). It is evident, therefore, that, at current levels of spending, putting more money into books is twice as cost effective in raising Key Stage 2 test scores as a similar amount spent on ICT. Increasing the stock of computer hardware beyond current levels, indicated by pupils per PC, has no significant effect on test scores. In line with previous studies, variations in the teacher-pupil ratios (i.e. class sizes) around the current average levels do not significantly affect pupil test scores. Thus the findings do not support a case for reducing class sizes any further. They argue that, instead, money should be put into improving the provision of learning resources, and of books in particular.
The article mentions in passing that in Norway book spending in schools is seven times higher than in England, and it also highlights that since 2004 data about book spending by schools is no longer collected by DfES, with only the ICT component of learning resource spending separately identified.
Comments