2 Ipsos MORI conducted an online survey of PFI project teams across England, between 1 November 2008 and 31 March 2009. The survey was used to collect information on the key study questions:
• Have PFI construction projects been completed to the contracted timetable?
• Have PFI construction projects been completed to the contracted price?
• Have PFI construction projects received good quality ratings?
3 The survey was sent to all projects identified as being within the defined population. Our PFI population was defined as all English PFI construction projects which were due to be complete between 1 January 2003 and 31 August 2008, with a capital value of £20 million or above. PFI projects were identified from databases maintained by the Treasury16 and PUK17 and were included if either one of the databases identified a project as being valued at £20 million or more. The resulting population totalled 153 projects.
4 The survey was completed by 114 respondents. The majority of these were completed online, though a number of projects recorded key information on a shorter Word version of the survey. This represents a 75 per cent response rate from the defined population. The number of respondents varies between questions (identified below each chart) as the number of respondents stating 'don't know' to questions varies.
5 A range of values has been presented for key statistics on time and price, as those who responded to our survey were only a sample of the total population. We cannot therefore be certain the figures obtained are exactly those we would have obtained if every project responded. In 2003 this was not necessary as a census was conducted; the whole population (except one project) responded to the census, rather than just a sample of the population as in 2008.
6 A 95 per cent confidence level has been utilised for the confidence intervals stated in this report. That is, for each percentage stated, there is a 95 per cent chance that the 'true' percentage falls within the specified range. Confidence intervals stated in this report are wider than those used for finite samples, to take account of the fact that 153 may not be the exact number in the population. It must also be noted that confidence intervals require random sampling of a population which may not be the case in this instance due to the self-selection of survey respondents. For this reason, confidence intervals should be used as a guide only.
7 A number of projects were subject to a validation exercise which was performed after survey responses were received. While not all questions were validated, the key questions referring to time and price performance, and the reasons for these, were followed up for a majority of projects reporting poor time or price performance. A random sample of respondents reporting delivery to timetable and price were also validated. This was performed by phoning projects and talking to project teams about their experiences in more detail. The NAO did not take this analysis any further. More work would need to be undertaken to get a comprehensive picture of PFI construction performance.
_________________________________________________________________________________
16 PUK Projects Database http://www.partnershipsuk.org.uk/PUK-Projects-Database.aspx
17 The Treasury's database of PFI projects http://www.hm-treasury.gov.uk/ppp_pf_stats.htm