84 Joseph Omotoso Ajayi
Data Collection
Rapid Rural Appraisal (RRA) (transect walks, identification and
inspection of farmlands) has been used widely in the region (Ajayi, 2015;
Ajayi, 2014b; and Nzeadibe et al., 2011) for data collection. The RRA has the
advantage of encouraging the respondents to describe their relationship with
their natural resources (Ajayi 2015; Ajayi 2014b; and Nzeadibe et al., 2011).
The RRA also has the advantage of identifying variables of importance to the
cassava farmers and in the formulation of questions that may be included in
the more formal semi- structured interview schedule in locally meaningful
terms (Ajayi 2015; Ajayi 2014b; and Nzeadibe et al., 2011). Two single
gender-based Focus Group Discussions (FGDs), one for male and another for
female were held with cassava farmers across the States with the number of
participants ranging from 10-20. The six FGDs used were very helpful in
eliciting clearer information from the farmers.
Analytical Techniques and Model Specification
Likert Scale
To examine gender-based climate change awareness of cassava farmers in
the study area, Vagias’s 5-point Likert scale of level of awareness (Vagias,
2005) was used. Means, grand means and standard deviations comparatively
for both male and female cassava farmers were used. The 5-point rating scale
was graded “Not at all aware” =1, “slightly aware”=2, “somewhat aware”=3,
“moderately aware”=4 and “extremely aware” =5 (Vagias, 2005).
Multinomial Logit Regression Model
To examine the factors influencing gender-based cassava farmers’ choice
of adaptation strategies in the study area, multinomial logit model (MNL) was
be used. This involved running two (2) separate comparative multinomial logit
models. One for male cassava farmers, the other for female cassava farmers.
MNL model has been used extensively in studying climate change adaptation
strategies. For examples (Ajayi, 2016; Fatuase and Ajibefun, 2012; Sowumi
and Akintola 2010; Deressa et al., 2008; Koch, 2007; and Green, 2000).
Others have also studied climate change adaptation strategies or adoption
using other models. For examples (Amusa et al., 2015; Komba and
Muchapondwa 2015; and Tihahum and Bedemo 2014) used Heckman’s
double stage model. Meanwhile, the MNL according to (Sofoluwe et al., 2011)