Appendix B

The learning effect of the diagram type was measured in the independent variable ROUND, whose values 1, 2 and 3 corresponded to the first, second and third period respectively.

For the codification of the carryover effect we found two strategies [13, pp. 545-546]:

1.       Milliken and Johnson [18, pp. 440-450] included carryover effects in the model as covariables: C1 and C2.

2.       Ratkowsky, Evans and Alldredge [24, pp. 4-7] introduced the carryover effects in the model as another factor: CARRY.

The analysis of ANOVA full model was carried out using both alternatives. The conclusions obtained were the same both for TSEC and for NRESP. With the purpose of abbreviating in Tables B.1 and B.2 the results for TSEC are shown according to the first strategy and for NRESP with respect to the second method. In both tables the learning effect is not statistically significant for TSEC (Sig. = 0.333 > a) nor for NRESP (Sig. = 0.696 > a). The same happens with carryover effects for the two response variables, given the statistic non-relevance of the covariables C1 (Sig. = 0.605 > a) and C2 (Sig. = 0.656 > a) and of the CARRY factor (Sig. = 0.152 > a).

Another important detail is that the term CARRY, instead of having three degrees of freedom, one less than its four levels 0, 1, 2 and 3, only has two, because the effects of the first period and carryover effects in the first period are confounded (that is, they cannot be estimated independently). In the case of the covariables C1 and C2, their degrees of freedom sum two.

Tables

Table B.1. SPSS results of ANOVA full model for TSEC according to [18]

Table B.2. SPSS results of ANOVA full model  for NRESP according to [24]

References

[13] Kuehl, R. O., Design of Experiments: Statistical Principles of Research Design and Analysis, Duxbury Thomson Learning, California (USA), 2000.

[18] Milliken, G. A. and Johnson, D. E., Analysis of Messy Data. Volume I: Designed Experiments, Chapman & Hall, New York, 1992.

[24] Ratkowsky, D. A., Evans, M. A. and Alldredge, J. R., Cross-over Experiments: Design, Analysis and Application, Marcel Dekker, New York, 1993.