217
18.2 Methods
This study focused on a population in the Upper East and Upper West regions of
Northern Ghana, in the Lambussie Karni, Kassena Nankana East and Kassena
Nankana West districts. Interviews with 223 households were conducted in March
2017. Informants were selected randomly from 26 villages within the project area,
where informants were either project beneficiaries (101 households) or members of
a control group (122 households) of non-beneficiaries. The villages were selected
on the basis of their already established relationships with partner organisations.
Beneficiaries in the project were self-selecting, and so it can be assumed they had
more interest in shea compared to the general population. The control group, comprising households identified as future project beneficiaries, were chosen because
they are directly comparable to the beneficiary households.
The RHoMIS tool uses a modular, rapid (40–60 min) digital survey to derive
standardised indicators on agricultural practices, livelihoods, food security and
dietary diversity, as well as gender roles (Hammond et al. 2017). A survey module
was developed to collect information on use of non-timber forest products (NTFPs)
and woody environmental resources. The indicator used was food availability
(Frelat et al. 2016), which converts all household income and agricultural produce
into a calorie per person score. Food availability was chosen in preference to cash
incomes as it also takes account of self-produced and consumed items and thus
provides a more comprehensive perspective on the livelihoods of the very poor
(Ritzema et al. 2017). Other rapid and well-tested indicators were also gathered:
experience of hunger, quantified using the Household Food Insecurity of Access
Scale (HFIAS) (Coates et al. 2007); dietary diversity, assessed using the Household
Dietary Diversity Score method (Swindale and Bilinsky 2006); and food groups,
gathered using the Minimum Dietary Diversity for Women guide (MDD-W) (FAO
and FHI 360 2016). The Progress out of Poverty Indicator (PPI) was used to crosscheck the household income figures gathered from direct questioning (IPA 2015).
The use of these standard indicators allows evaluation of the project impacts in a
wider frame of reference, comparison to other locations, and evaluation of changes
over time should a further RHoMIS study be done at a later date.
Households were classified into three poverty classes based on their food availability scores. Households with access to less than 2500 kcal per male adult equivalent (MAE) person per day were classed as “below the calorie line.” Households
above the calorie line but with a total value of activities (i.e., actual cash income
plus the value of consumed agricultural produce) below US$1.90 were classed as
“below the poverty line.” Households with total value of activities above US$1.90
were classed as “above the poverty line.” Welfare indicators have been presented as
medians per household group, and incomes have been presented as trimmed means,
where 5% of the observations at either extreme of the scale were dropped to reduce
the effect of outliers. The Kruskal-Wallis test for significance was used when comparing between beneficiary and non-beneficiary households within paired poverty
classes, and unless otherwise stated all significance was attributed at the 0.95 level.
18 Shea Butter: A Pro-Poor, Pro-Female Route to Increased Income
18.2 Methods
This study focused on a population in the Upper East and Upper West regions of
Northern Ghana, in the Lambussie Karni, Kassena Nankana East and Kassena
Nankana West districts. Interviews with 223 households were conducted in March
2017. Informants were selected randomly from 26 villages within the project area,
where informants were either project beneficiaries (101 households) or members of
a control group (122 households) of non-beneficiaries. The villages were selected
on the basis of their already established relationships with partner organisations.
Beneficiaries in the project were self-selecting, and so it can be assumed they had
more interest in shea compared to the general population. The control group, comprising households identified as future project beneficiaries, were chosen because
they are directly comparable to the beneficiary households.
The RHoMIS tool uses a modular, rapid (40–60 min) digital survey to derive
standardised indicators on agricultural practices, livelihoods, food security and
dietary diversity, as well as gender roles (Hammond et al. 2017). A survey module
was developed to collect information on use of non-timber forest products (NTFPs)
and woody environmental resources. The indicator used was food availability
(Frelat et al. 2016), which converts all household income and agricultural produce
into a calorie per person score. Food availability was chosen in preference to cash
incomes as it also takes account of self-produced and consumed items and thus
provides a more comprehensive perspective on the livelihoods of the very poor
(Ritzema et al. 2017). Other rapid and well-tested indicators were also gathered:
experience of hunger, quantified using the Household Food Insecurity of Access
Scale (HFIAS) (Coates et al. 2007); dietary diversity, assessed using the Household
Dietary Diversity Score method (Swindale and Bilinsky 2006); and food groups,
gathered using the Minimum Dietary Diversity for Women guide (MDD-W) (FAO
and FHI 360 2016). The Progress out of Poverty Indicator (PPI) was used to crosscheck the household income figures gathered from direct questioning (IPA 2015).
The use of these standard indicators allows evaluation of the project impacts in a
wider frame of reference, comparison to other locations, and evaluation of changes
over time should a further RHoMIS study be done at a later date.
Households were classified into three poverty classes based on their food availability scores. Households with access to less than 2500 kcal per male adult equivalent (MAE) person per day were classed as “below the calorie line.” Households
above the calorie line but with a total value of activities (i.e., actual cash income
plus the value of consumed agricultural produce) below US$1.90 were classed as
“below the poverty line.” Households with total value of activities above US$1.90
were classed as “above the poverty line.” Welfare indicators have been presented as
medians per household group, and incomes have been presented as trimmed means,
where 5% of the observations at either extreme of the scale were dropped to reduce
the effect of outliers. The Kruskal-Wallis test for significance was used when comparing between beneficiary and non-beneficiary households within paired poverty
classes, and unless otherwise stated all significance was attributed at the 0.95 level.
18 Shea Butter: A Pro-Poor, Pro-Female Route to Increased Income