290
A. Song and R. Chuenpagdee
group could reduce the reliability of the results of this study. However, as other studies
of similar methodologies and intent have shown (c.f. Chuenpagdee et al. 2003 ; Quah
et al. 2006 ; Bose and Crees-Morris 2009 ) , in-group consistency was quickly reached
with the number of respondents obtained in the study. Thus, increasing the number of
respondents would be of no pragmatic value from the view of both cost-effectiveness
and time-ef fi ciency (Bose and Crees-Morris 2009 ) . Whenever possible, a one-on-one
setting was preferred when conducting the survey in order to minimize any strategic
bias that may arise from social pressure or fear of reprisal. Hence, most surveys
were conducted in a quiet, sheltered environment in the absence of other community
members. Each survey took an average of about 25 minutes to complete.
Fig. 15.2 Map of the Southeast Arm of Lake Malawi (Survey sites are shown as dark circles ;
double circles indicate town centres) ( Source : Song 2009 )
A. Song and R. Chuenpagdee
group could reduce the reliability of the results of this study. However, as other studies
of similar methodologies and intent have shown (c.f. Chuenpagdee et al. 2003 ; Quah
et al. 2006 ; Bose and Crees-Morris 2009 ) , in-group consistency was quickly reached
with the number of respondents obtained in the study. Thus, increasing the number of
respondents would be of no pragmatic value from the view of both cost-effectiveness
and time-ef fi ciency (Bose and Crees-Morris 2009 ) . Whenever possible, a one-on-one
setting was preferred when conducting the survey in order to minimize any strategic
bias that may arise from social pressure or fear of reprisal. Hence, most surveys
were conducted in a quiet, sheltered environment in the absence of other community
members. Each survey took an average of about 25 minutes to complete.
Fig. 15.2 Map of the Southeast Arm of Lake Malawi (Survey sites are shown as dark circles ;
double circles indicate town centres) ( Source : Song 2009 )