Given Hapunda

208 design in which the data were collected within the same period of time (Morroni & Myer, 2007), whereas the implementation and impact study pursued a case-control study design (Morroni & Myer, 2007). Three phases of the study are illustrated in Figure 6. Figure 6: Overview of experimental study design. Study population Sample size was determined using Stata programme version 12.0MP in order to collect and produce statistically representative data. The parameter used was a 95 confidence level ( p =0.05) with a power of 90%; the alternative mean was 12 with a standard deviation of 8.4. The calculated sample size was 84 for each group. As a result, 100 children were selected for each group. Extra 16 respondents in each group was to ensure that the sample size would still be sufficient in case of drop-outs. A random sampling technique was used to select the participating pre-school children from the list of consenting caregivers and from the purposively selected pre-schools (n=300). This technique was considered to be suitable for this study because it was unique in representing a completely experimental Baseline survey: Pre- intervention  Socio-demographic, Health  Anthropometric (Weight (kg), Height (cm), MUAC)  Biochemical measurements Intervention: Feeding programme  Feeding programme: 10-month period Post-Intervention: Outcome analysis  Anthropometric (Weight (kg), Height (cm), MUAC)  Biochemical measurements Phase 1 1 Phase 2 Intervention Phase 3