Table 2.
Results of binomial logistic regression analysis of factors affecting proportion of total fruits allocated to SRE
| Term | Parameter estimate (±SE) | Wald chi-square statistic | df | p | Estimated odds ratio |
|---|---|---|---|---|---|
| Constant |
−4.88 (0.20) |
589.81 |
1 |
<0.001* |
- |
| Year = 2008 |
−0.40 (0.32) |
1.51 |
1 |
0.22 |
0.67 |
| Year = 2009 |
0.18 (0.23) |
0.61 |
1 |
0.44 |
1.19 |
| Year = 2010 |
0 |
- |
- |
- |
1.00 |
| Bolting month = June |
3.35 (0.17) |
389.09 |
1 |
<0.001* |
28.50 |
| Bolting month = July |
2.96 (0.17) |
300.49 |
1 |
<0.001* |
19.30 |
| Bolting month = August |
1.71 (0.18) |
87.45 |
1 |
<0.001* |
5.52 |
| Bolting month = September |
0 |
- |
- |
- |
1.00 |
| Size | 0.02 (0.00) | 0.37 | 1 | 0.54 | 1.01 |
The four bolting groups were experimentally produced using a photoperiod manipulation, and this manipulation was repeated for three years from 2008-10. Prebolting plant size was measured as the length (in mm) of the longest living leaf of the unbolted rosette. The regression model is a significant predictor of the response (χ2 = 2200.57, df = 32, p < 0.005). Odds ratios express relative likelihood of fruit being allocated to the SRE.
*p < 0.05.