In other words, denying the zero assumption that there is no difference between the two sets of test results would simply allow us to say that the tests do not match; non-proof of this hypothesis would not prove that the tests are consistent. Chen CC, Barnhart HX. Comparison of CCI and CCC to assess compliance for data without replication and with. Comput Stat Data Anal 2008;53:554-64. The first step is to study the data. A simple diagram of the results of one method against another (Fig. 1) is certainly a useful start without a regression line, but normally the data points are grouped close to the line and it will be difficult to assess the differences between the methods. A representation of the difference between methods relative to their average may be more instructive.