11 We carried out statistical analysis in order to explore trends in the number and value of changes made in PFI projects and in the staffing levels of project teams.
Results for the extent of change in operational PFI projects
12 We carried out stepwise regression analysis to explore reasons for the extent of change carried out, defined here as the amount spent on changes in 2006.24 We entered all the variables that might conceivably have affected the level of change carried out by projects, including project age, sector and size; however, all were eliminated in the process of selection.
Results for general staffing levels
13 We carried out linear regression analysis to explore the extent to which variations in staffing levels could be explained by the size of projects (defined as the unitary charge payment for 2006-07). We found that the size of projects explained about 40 per cent of the variation in staff resources. There was a correlation between project size and staff resources for all sectors; however, the correlation was weaker for school projects than other sectors. This may reflect the fact that the size of PFI schools deals may be a weaker indicator of complexity than in other sectors: some deals in the schools sector cover the maintenance of dozens of existing buildings, whilst other deals of a similar size cover only one or two newly constructed schools, with far simpler operational requirements.
Results for staff time spent on managing changes
14 We asked project teams to estimate the proportion of time spent by staff specifically on managing changes. This was multiplied by staffing levels to give a figure for the total staff time spent on managing changes. We then explored the relationship between this figure and the extent of change (defined as the amount spent on change in 2006), again through linear regression. We found that approximately 25 per cent of the variation in staff resources devoted to managing changes can be explained by the level of change. However, repeating the analysis by sector shows that the relationship exists only for changes made in PFI schools projects. For hospitals and other sectors, there is no statistically significant relationship between the amount spent on changes in 2006 and the associated staff resource.