Table 3.

ANOVA results for the quadratic models describing arabitol concentration and glycerol conversion yield as function of glycerol concentration (A), temperature (B), pH (C), and DOT (D) together with model performance parameters

Response	Source	Sum of Squares	Degrees of freedom	Mean Square	F value	p-value Prob > F
Arabitol	Model	1272.84	6	212.14	31.51	< 0.0001
	A	185.93	1	185.93	27.61	< 0.0001
	B	416.67	1	416.67	61.88	< 0.0001
	C	106.68	1	106.68	15.84	0.0006
	AB	186.32	1	186.32	27.67	< 0.0001
	A2	331.54	1	331.54	49.24	< 0.0001
	B2	76.42	1	76.42	11.35	0.0027
	Residual	154.86	23	6.73
	Lack of fit	145.90	18	8.11	4.53	0.0512
	Pure error	8.95	5	1.79
	R-squared					0.89
	Adj R-squared					0.86
	Pred R-squared					0.78
	Adeq precision					21.6
Conversion yield	Model	0.090	6	0.015	10.83	< 0.0001
	A	6.534·10− 3	1	6.534·10− 3	4.72	0.0404
	B	0.031	1	0.031	22.25	< 0.0001
	C	0.024	1	0.024	17.10	0.0004
	AB	6.806·10− 3	1	6.806·10− 3	4.91	0.0368
	A2	0.017	1	0.017	12.23	0.0019
	B2	7.425·10− 3	1	7.425·10− 3	5.36	0.0299
	Residual	0.032	23	1.385·10− 3
	Lack of fit	0.031	18	1.739·10− 3	15.62	0.0032
	Pure error	5.568·10− 4	5	1.114·10− 4
	R-squared					0.74
	Adj R-squared					0.67
	Pred R-squared					0.52
	Adeq precision					11.1

Only the terms with significant effect (P < 0.05) and those that have significant interactions were retained in the model and reported herein