Microtensile bond strength test bias caused by variations in bonded areas

While it has been shown that no method produces specimens with exactly the same cross-sectional bonded area (BA), BA variations within and between studies are a well-known covariate in microtensile test results. However, no method has yet been described to accurately account for its influence. A procedure is presented that allows controlling for variations of BA effects on results. Further, a proposal for reporting is presented which enables results of different studies to be compared.

Partially using the results of the report in which the microtensile test was originally described, 144 both general (caused by differences in BA) and specific (due to a material's performance differences and intrinsic biological variability of specimens) variabilities were separated through linear regression of microtensile (MPa) to BA (mm2) pooled results. Comparing the specific variability of specimens - the residuals to the regression line - of groups allowed assessing differences between groups.

A means comparison of residuals showed that specific differences were significant (t-test, p = 0.0004). The null hypothesis could be rejected: materials' performances were different. This could not be determined in the original report, since BA variability was very high. A proposal for reporting of results to facilitate their clinical interpretation and comparison between studies is presented.

Controlling for general variability caused by differences in BA size allows precise comparison of microtensile tests results.