2.5. Statistics
Statistical analysis was carried out in SAS 9.2 using the MIXED model procedure.A mixed model was fitted to each of the fruit locations (i.e. green v.yellow tissue) and colour measure combinations separately. Pressure,time,pressure x time interaction,day and sample order were all treated as fixed effects. However, grower was treated as a sample of the population of growers and if the experiment were repeated, a different sample of growers may have been selected.Therefore, grower was included in the model as a random effect.The respiration rate data sets were analysed for the same effects except that the sample order was not included in the model. To meet the model assumptions of normally distributed residuals with constant variance for the respirationrate data,all statistical tests were carried out on log transformed respiration data, with fitted means back transformed onto the original scale. The ethylene measures at 1 h contained a number of zeros. So half the minimum non-zero value was added to these before the log transformation. As the ethylene values were very low at 17 h and contained a very large numbers of zeros; only summary means for these. Without further analysis, are presented.Pair-wise comparisons involved the means for each pressure/time combination being tested against the control. All tests were conducted at the 5% level of significance, with dunnett’s adjustment used to account for multiple testing in the post hoc tests.