Table 3 Statistical results of supercooling temperature as a function of magnetic field strength, as described in the “Methods” section, for the groups shown in Fig. 3

Treatment		Total # analyzed	# Exactly observed	# Right censored	Mann–Whitney–Wilcoxon test (2-tailed Wilcoxon rank-sum test)
Treatment		Total # analyzed	# Exactly observed	# Right censored	Z value	n score	p-value	Comment
A. 1.0 mT	Field	91	21	70	2.061	+4.245	0.039	Field-treated freezes before sham
A. 1.0 mT	Sham	79	12	67	2.061	+4.245	0.039	Field-treated freezes before sham
B. 1.5 mT	Field	80	22	58	−0.434	−1.065	0.664	Indeterminate
B. 1.5 mT	Sham	66	25	41	−0.434	−1.065	0.664	Indeterminate
C. ≧2.0 mT	Field	175	90	85	−2.020	-9.345	0.043	Field-treated freezes after sham
C. ≧2.0 mT	Sham	180	116	64	−2.020	-9.345	0.043	Field-treated freezes after sham

Field and sham experiments at a given current level were done using garlic bulbs from the same allocation to reduce the variability that might arise from uncontrolled factors such as age and magnetite content. Out of the 721 garlic cloves examined, there were 543 total cloves measured with thermocouples (Supplementary Data 1), of which 286 froze (Exactly Observed) according to the thermocouple spikes. The rest were censored as described in the “Methods” section. Use of the Wilcoxon test rather than a Log-Rank test is appropriate here because the probability of ice nucleation events is not constant with temperature⁵⁷ (see ref. ⁵³ for a discussion on the statistics).

Quick links

Search