Modelling Determinants of Tree Planting and Retention on Farm for Improvement of Forest Cover in Central Kenya

Table 6

Likelihood ratio tests and model classification of tree retention determinants using binary and multinomial logistic regression.

Determinants	−2log likelihood		d.f		𝑃 value		% Model classification
	*Logt	*Mult	Logistic	Multinomial	Logistic	Multinomial	Logistic	Multinomial
Site	281	45.1	3	6	0.05	0.04	59	50
Gender HH*	273	24.8	1	2	0.12	0.03	55	51
Occupation	275	20.6	1	2	0.06	0.16	55	50
Age	262	238	1	126	0.93	0.06	52	67
Education	267	40.0	3	6	0.04	0.14	58	52
Marital status	279	18.7	1	2	0.38	0.32	53	51
NMH*	275	85	1	2	0.47	0.47	53	52
Income	220	168	1	80	0.00	0.00	60	53
Land tenure	276	43.9	2	10	0.76	0.15	52	51
Land size	231	130	1	6	0.00	0.00	68	68
Tree use	240	27.3	2	4	0.35	0.53	55	48
Motivation	225	27.5	1	2	0.20	0.01	56	52
Technical skills	262	21.1	1	2	0.05	0.13	56	49
Skill effect	71.8	16.6	1	2	0.05	0.14	66	66
Labour and cost	232	25.8	1	2	0.05	0.01	57	50
Extension services	263	18.8	1	2	0.07	0.06	53	50
Harvest permission	229	23.9	1	2	0.04	0.05	58	50
Harvesting regulation	208	25	1	2	0.22	0.02	55	50
Forest associations	234	22.1	1	2	0.83	0.11	51	50
Membership	253	18.1	1	2	0.70	0.79	51	49
Ready market	195	19.8	1	2	0.24	0.22	55	51
Marketing problems	203	21.7	1	2	0.16	0.07	55	50

∗Logt: Logistic regression values; *Mult: multinomial logistic regression values; HH*: household head; NMH*: number of members in the household.