211 In summary Dr Burman only postulated a partial explanation of the difference between the results and the specified requirements under the contract.
212 One of the problems about the tests is, of course, that they were spread over all the different levees. For instance, for the Tarrion Creek levee and Charlton Road there were two tests while at North Brewarrina there were three tests. In the Southern levee there were only two tests whereas in the Northern levee there were eight tests. Of the seventeen tests, six, being at least one in each levee, achieved the standard of compaction required under the contract and nine were between 90% and 95%. Two tests failed. In other words they were below 90%. This is a very small spread of results over each levee and there is no evidence that has dealt with the significance of those results in a statistical sense. Hence one has the substantial number of CETS tests during construction which achieved appropriate levels. In these circumstances I do not think that the Golder tests really assist in determining whether the levees in general were compacted as required under the contract.
213 It is also relevant in a general sense, although they are also subject to the same problem as the Golder tests, that the Douglas Partners tests all provided satisfactory results immediately after the conclusion of the works.
214 I am satisfied having regard to the CETS tests which I regard as accurate, and having regard to the undoubted evidence of compaction while the works progressed under the daily supervision of Mr Komp, that the levees themselves were compacted in accordance with the relevant specification.
215 I note of course the opinions of Professor Fell and Dr Truscott but their opinions are predicated on the acceptance of the Barnson testing which I have rejected. They also suffer from what was submitted by the plaintiff to be a problem in that their opinions do not seek to describe or elucidate the process by which the logical step between the test data and their ultimate conclusions in respect of the condition of the levees was made.
216 In respect of Professor Fell and Dr Truscott, when their reports and affidavits were tendered or read, objection was taken to the relevant conclusions on the basis that such conclusions offended the principles which were recently referred to by the Court of Appeal in Makita (Australia) Pty Ltd v Sprowles. At paragraph 85 Heydon JA set out the principles for the admissibility of expert opinion evidence. There he said:
"In short, if evidence tendered as expert opinion evidence is to be admissible, it must be agreed or demonstrated that there is a field of 'specialised knowledge'; there must be an identified aspect of that field in which the witness demonstrates that by reason of specified training, study or experience, the witness has become an expert; the opinion proffered must be 'wholly or substantially based on the witness's expert knowledge'; so far as the opinion is based on facts 'observed' by the expert, they must be identified and admissibly proved by the expert, and so far as the opinion is based on 'assumed' or 'accepted' facts, they must be identified and proved in some other way; it must be established that the facts on which the opinion is based form a proper foundation for it; and the opinion of an expert requires demonstration or examination of the scientific or other intellectual basis of the conclusions reached: that is, the expert's evidence must explain how the field of 'specialised knowledge' in which the witness is expert by reason of 'training, study or experience', and on which the opinion is 'wholly or substantially based', applies to the facts assumed or observed so as to produce the opinion propounded. If all these matters are not made explicit, it is not possible to be sure whether the opinion is based wholly or substantially on the expert's specialised knowledge. If the court cannot be sure of that, the evidence is strictly speaking not admissible, and, so far as it is admissible, of diminished weight. And an attempt to make the basis of the opinion explicit may reveal that it is not based on specialised expert knowledge, but, to use Gleeson CJ's characterisation of the evidence in HG v R (1999) 197 CLR 414, on 'a combination of speculation, inference, personal and second-hand views as to the credibility of the complainant, and a process of reasoning which went well beyond the field of expertise' (at [41])".
217 There was debate as to the admissibility of the reports and at transcript 120 I ruled that at that stage the reports would be admitted but that the question of whether such evidence was properly admitted or, alternatively their weight, could be argued at a later stage.
218 As was made plain in the debate as to admissibility there are two principal problems with both reports in this area. The first of these is their statistical relevance of extrapolating a small number of tests over the large area of the levee banks. The second area is the logical step from the existence of non-compaction or compliance with requirements for linear shrinkage, to the proposition that the levee banks are likely to fail. I turn now to the first of these matters.
219 In their submissions the plaintiff made the following points:
"In essence the plaintiff's objection to the conclusions expressed by Dr Truscott and Professor Fell in respect of which Makita objections have been taken is that the opinions do not seek to describe or elucidate the process by which the logical step between the test data and their ultimate conclusions in respect of the condition of the levees was made.
Even accepting the Barnsons/GHD testing as reliable (which the plaintiff submits elsewhere it clearly is not) the data upon which the defendant's experts have sought to draw conclusions consists of the results of tests taken at a very small number of discrete points spread over five levees. Those tests can themselves be divided between tests taken at culverts, tests taken in pavements and tests taken in embankments (ignoring for the moment the contest over whether all these tests were in fact taken in fill required to be compacted in the course of the contract).
These discrete points are a very small proportion of the total population of areas of fill. Even at the chainages where tests were taken the test results only relate to the specific point at which the test was taken, they do not necessarily reflect the condition of the entire embankment at that point.
It is not permissible to draw the conclusion that any test or group of tests is representative of any area within the embankments without careful justification. The appropriate method of so doing (if possible) would have been by properly documented statistical analysis. Such statistical analysis would have required identification of those areas which were considered to constitute a statistically valid population and the reasons why they are so considered.
For example, the areas surrounding culverts cannot be considered to be part of the same population as the embankments since they were clearly constructed and compacted by different methods. Equally, there are significant differences between levees constructed as overlays and those constructed from scratch. In the plaintiff's submission, a proper analysis of these issues would have required justification for treating even discrete parts of the same levees as part of the same population.
Both Dr Truscott's and Professor Fell reports are signally devoid of any such analysis.
The next step in seeking to generalise from a limited number of tests to a whole population must be the statistical analysis itself. It is simply not appropriate to draw the conclusion that a limited number of tests is representative of the whole population from which they were drawn without proper statistical analysis. No attempt at such analysis is contained in any of Dr Truscott's or Professor Fell reports.
One factor that ought to loom large in any attempt to assess whether the test results indicate that the levees were not properly compacted at the time of construction is an assessment of the degree of accuracy of the tests themselves. Again there is no consideration of this issue by either Professor Fell or Dr Truscott.
As canvassed elsewhere in these submissions the evidence shows that the degree of precision of Hilf compaction tests is no more than 4%. When allied with the 2% limit of accuracy of the Nuclear densometer gauges, it is apparent that relative density testing may be out by as much as 6%. As Dr Truscott noted in cross examination (T376-377) in considering whether some results below 95% could be considered as indicative of an improperly compacted levee bank it 'depends how below and it depends how many'. In other words, it requires a statistical analysis, and neither of the defendant's experts has sought to perform such an analysis."
220 If one looks at Professor Fell's reports, one sees his somewhat incomplete and an inaccurate reference to the testing in Table 1 and the dynamic cone penetration tests in Table 2. After referring to these tests in paragraph 17 he then gives his results in paragraph 19. These are repeated in a slightly different form in the answer provided at paragraph 46. It is plain that he gives no logical reason or explanation as to how he comes to the conclusions as to the level of compaction in the different levees from the particular results. When one looks at Dr Truscott's evidence and particularly his conclusions at section 12 of his report one does not see in the preceding material any explanation as to why test results are indicative of the whole of the levees.
221 There was no evidence from any of the experts that one should engage in some statistical analysis. However, this is really self-evident and it is worth noting that the parties specified a regime under their contract which led to there being upwards of 300 compaction tests during the course of the job. In contrast, the Golder tests numbered only 18 tests. Presumably the engineers who drew the specification requirements for testing gave consideration to what would be an appropriate level of testing to ensure the desired result.
222 In considering these objections it is probably useful to bear in mind the purpose for which the matter is being considered, namely, there is a decision to be made as to whether or not the plaintiff complied with the contract in respect of the compaction requirement. If it did not, the next question that arises is whether there has been a substantial performance of the contract. This question of substantial performance clearly must take account of, if one were to rely on Professor Fell's and Dr Truscott's evidence, the second question in respect of their reports to which I have referred above. Namely, whether there is a proper elucidation of their reasons for moving from the proposition that the levees contain weaknesses and the proposition that the levees are likely to fail. In this respect the plaintiff's submissions were as follows:
"The second step in respect of which Professor Fell and Dr Truscott fail to provide proper elucidation of their reasons consistently with the requirements of Makita v Sprowles is in moving between the proposition that the levees contain weaknesses and the proposition that the levees are likely to fail. Although there is discussion of 'piping' and cracking, there are a number of intermediate steps which are left unexamined. It is first necessary to draw the conclusion that areas of weakness extend right through the levees (that is through a whole cross-section). If that conclusion cannot be drawn it is necessary to identify the extent of the areas of weakness and then identify the manner in which such limited areas of weakness are said to expose the levees to the risk of failure.
A proper analysis of the foregoing issues would require some quantitative assessment of the impact of such deficiencies in the construction of the levees as could properly be established on the extent of piping and cracking (in particular by reference to that which could have been expected in the levees had such deficiencies not been present). Neither Dr Truscott nor Professor Fell sought to or was apparently able to carry out such quantitative analysis. In respect of linear shrinkage the best Dr Truscott could do was 'if it is lower it is better if it is higher it is worse' (T396.4) and 'it is not just as good whether it is significantly worse I don't know' (396.42). At the same time all the tests concerning the dispersivity of the fill used in the construction of the levee suggested it was not dispersive. It is dispersive soils that are more susceptible to piping. (T473.42)
Finally the defendant's experts seek to dismiss the CETS results on the basis that they cannot be correct. Yet again there is no attempt to engage in any form of statistical analysis to assess whether the subsequent test results actually demonstrate that that is the case. If the subsequent test results were to be used to impugn the CETS results it was incumbent upon the defendant's experts to provide a coherent explanation for that conclusion. Again, in contravention of the requirements of Makita v Sprowles , no such explanation is provided.
For the foregoing reasons the plaintiff submits that those parts of the various expert reports to which Makita objections were taken should not be admitted into evidence. However, even if this submission is not accepted, consistently with the penultimate sentence of paragraph 85 of Makita v Sprowles , the evidence should be accorded little weight."