Stake tests are a critical part of evaluating durability of wood in ground-contact, but there is a lack of criteria for interpreting stake test results. This paper discusses criteria that might be used to determine if short term ratings indicate satisfactory longterm performance. Ratings of 19 by 19 mm stakes from multiple plots in the Harrison Experimental Forest, Mississippi, were analyzed to determine how well performance criteria at years three, four, or five predicted durability over the longer term (nine years). Results revealed that with only three years of data, false prediction of satisfactory performance was common even when strict criteria were applied. Predictions of future performance improved with four and five years of data, although some instances of false prediction still occurred. These analyses indicate that with three to five years of data, even very slight differences between ratings of a Test treatment and a Reference preservative may indicate unsatisfactory future performance. However, variability in the ratings of the standard reference preservative can make it difficult to detect slight differences and increasing the number of replicates may be warranted. More complex statistical approaches that incorporate all years of rating data into comparisons may help to account for year to year variation in ratings. This analysis was limited to one test site and a similar type of analysis for other test sites would be beneficial.
Lebow, Stan T.; Lebow, Patricia K.; Kirker, Grant T. 2017. Comparison of performance criteria for evaluating stake test data. In: McCown, C.; Gothard, T.; Staula, B., eds. Proceedings, One hundred thirteenth annual meeting of the American wood protection association. Birmingham, AL: American Wood Protection Association: 246-256.