Our approach to analysis

We have carried out a longitudinal analysis on a set of schools that remains the same through time. A common issue raised in the feedback was that external factors, such as student background, have a major influence on educational performance. It was argued that it is, therefore, not possible to draw conclusions on the influence of private finance without controlling for such external factors. The techniques we continued to use to ensure that conclusions can be drawn included the following.

•  The sample was the entire population of state secondary schools in England (apart from a small number of randomly distributed exclusions). This ensured that the findings remained representative of the population studied. The representativeness of the sample was also tested as outlined later in this section.

•  The focus was on assessing the difference in performance in the same dataset before and after an event. This eliminated the impact of random external variables within the group of schools studied. In other words, the same schools were studied before and after they were rebuilt.

•  The data was de-trended against the national average or, put another way, the performance difference was analysed relative to the average in England. This minimised the impact of random external variables in different years of the time-series data.