Table 3.

Discriminant function analysis. Comparison of the models constructed based on three sets of predictor variables.

	Model 1		Model 2		Model 3
	Model Parameters
Wilks’ Lambda	0.4849		0.5777		0.3871
F	6.9045		3.6546		6.3324
p-value	<0.0006		<0.0128		<0.0004
Sensitivity	94.44%		77.78%		94.44%
Specificity	69.23%		76.92%		84.62%
Total Group Membership Classification	83.87%		77.42%		90.32%
	Variables Included to the Model
	Variable	Partial Wilks’ Lambda	Variable	Partial Wilks’ Lambda	Variable	Partial Wilks’ Lambda
	Asn	0.9926	1488.59	0.8619	1488.59	0.9607
	EtN	0.9310	2913.15	0.9290	2913.15	0.9966
	Cit	0.9890	3430.44	0.9966	4111.89	0.9846
	Val	0.9975	4111.89	0.9884	Asn	0.9082
			4437.76	0.9999	EtN	0.8845
					Cit	0.9064