Skip to main content
. 2015 Apr 15;18(17):3224–3233. doi: 10.1017/S1368980015000919

Table 3.

Impact of supermarket purchase on child/adolescent nutrition, Central Province of Kenya, July and August 2012

Model 1 BAZ Model 2 HAZ Model 3 HAZ Model 4 Stunted (dummy) Model 5 Stunted (dummy)
Explanatory variables ME se ME se ME se ME se ME se
Buys in supermarket (dummy) 0·18 0·34 0·63** 0·27 −0·05 0·1
Supermarket purchase share (%) 0·03*** 0·01 −0·00 0·00
Age (months) −0·00** 0·00 −0·01*** 0·00 −0·01*** 0·00 0·00*** 0·00 0·00*** 0·00
Female (dummy) 0·11 0·13 0·08 0·15 0·13 0·15 −0·022 0·05 −0·03 0·05
Household expenditure (1000 Ksh/AE per month) 0·00 0·01 0·03* 0·02 0·02 0·02 −0·01** 0·01 −0·01** 0·01
Education of person responsible for food (years) 0·03 0·02 0·00 0·03 0·00 0·03 0·00 0·01 0·00 0·01
Married household head (dummy) −0·11 0·16 0·14 0·20 0·18 0·20 −0·07 0·05 −0·08 0·05
Malaria or respiratory infection (dummy) −0·44* 0·26 −0·43* 0·24 0·04 0·09 0·04 0·08
Height of female adult (cm) 0·06*** 0·02 0·06*** 0·02 −0·01*** 0·00 −0·01*** 0·00
Age of female adult when the child was born (years) 0·03** 0·01 0·02** 0·01 0·00 0·00 −0·00 0·00
Household treats drinking water (dummy) 0·36** 0·15 0·35** 0·15 −0·07 0·05 −0·06 0·05
Distance to nearest health-care centre (log of km) −0·04 0·07 0·02 0·07 0·05* 0·03 0·04 0·03
Age of female adult (years) 0·01* 0·01
Physical education at school (h/week) −0·02 0·03
Leisure-time physical activity (h/week) −0·00 0·01
Distance to nearest district hospital (log of km) 0·01 0·06
Constant −0·61 0·45 −10·76*** 2·57 −10·71*** 2·54
Number of observations 216 216 216 216 216
χ 2 test statistic 169·34*** 211·09*** 156·79*** 336·572***

BAZ, BMI-for-age Z-score; HAZ, height-for-age Z-score; ME, marginal effect; se, robust standard error; Ksh, Kenyan shillings; AE, adult equivalent.

Results of instrumental variable regression models are presented, with dependent variables shown in the column headings (first-stage results are presented in Table 4S in the online supplementary material). Models 2 and 3 differ only in terms of how supermarket purchase is defined, as shown in the first two rows; likewise for models 4 and 5.

*P<0·1, **P<0·05, ***P<0·01.