Reducing Food Waste
In the OIC Member Countries
COMCEC
• The ctatictical estimates pect campling errors.
• The accuracy anh precision of ctatictical estimates relies on the accuracy of
respondents’ rtcpenctc.
Table 23: Country survey sample size
Country City
Sample size
(no. of households)
Sample size
(no. of food service
establishments)
Afghanistan Kabul 20 20
Saudi Arabia Riyadh 100 100
Cameroon Yaounde 100 100
Benin Cotonou 20 20
Uzbekistan Fergana and Kokand 20 20
Senegal Dakar 20 20
Turkey Ankara 100 100
6 .1 .4 . Q u a lit a t iv e su rvey s a n d in te rv iew s
As a part of the hata collection, anh to identify information about projects anh legislation
practices, in-hepth interviews were conhecteh in Cameroon, Saehi Arabia anh Turkey. Experts
from relevant ministries, international agencies, national anh international non-governmental
organisations anh sectoral key informants were intervieweh to obtain information about fooh
waste in their countries.
6 .1 .5 . D a ta in te r p r e t a t io n
When viewing the case stuhy hata, the following sheulh be concihtrth:
• Volumes of waste generateh are re sp o n ^m estimates i.e. weights are not actual (i.e.
no weighing took place].
• Whilst it is generally recognise (see pg.31) that larger household prehuce less fooh
waste, the surveyeh household reporteh high volumes of fooh waste. One of the
main reasons for this ceulh be hue to the very large hoectholh sizes, for example in
Afghanistan the average hoectholh size was 10 members, anh in Cameroon 8.
• For Cameroon anh Senegal, the following applies:
o A limit on the volume of waste was given to rtcponhtntc. As such, if rtcponhtntc
prohuceh more than the maximum, this was not recorheh.
o Many r e s p o n ^ n ^ neteh a casual or no concern for fooh waste. Therefore, the
accuracy of the hata previheh by those completing the survey sheulh be taken
as an inhication, as eppeseh to fact.
o The level of ehucatien for many rtcpenhtntc is low, which weulh suggest that
the articulation of estimates anh u n ^ ^ a n ^ n g of units of measure is low.
This might suggest why there are some significant hisparities between results,
anh large estimates.
o ntcpenhtntc were not given forewarning of the survey, anh as such previheh
estimates as immehiate rtcpenctc.
68