Evaluation of alternative school feeding models on nutrition, education, agriculture and other social outcomes in Ghana: rationale, randomised design and baseline data

Table 3 Threats to internal validity (source: adapted from [12])

Indicator	Metric	Spill-over and contamination	Hawthorne and placebo effects
Schooling	Enrolment, attendance, drop-out and completion	Children may attend school from neighbouring communities to have access to meals	Expectation of coming programme in control communities
Cognitive ability	Raven’s matrices, digit span and/or other tests	Very unlikely	Teachers’ and children’s attempt to over-perform in both project and control communities
Attention	Digit span and/or other test	Very unlikely	Teachers’ and children’s attempt to over-perform in both project and control communities
Learning achievement	Scores on language and maths tests	Very unlikely	Teachers’ and children’s attempt to over-perform in both project and control communities
Physical growth	Anthropometric measures of height and weight	Children from other communities may access school meals	Very unlikely
Physical activity levels (PAL)	Parents’ perceptions	Very unlikely	Very unlikely
Diet diversity	Household consumption	Very unlikely	Very unlikely
Micronutrient intake	Iron status, anaemia	Children from other communities may access school meals	Very unlikely
Income	Farm profits	Unlikely, if food purchases are made in control communities	Very unlikely

ISSN: 1745-6215