Before-and-after evaluation

Begin with the improved-looking bin

Imagine a kitchen changing its preparation and dispatch arrangements. Its own discarded-food mass falls, so the local result looks encouraging. Yet the receiving site reports more discarded food. An evaluation limited to the kitchen bin would miss that later stage.

Before-and-after evaluation asks what changed relative to a defined baseline. It requires the relevant food outcome and boundary, not only the easiest favourable measurement. The loss-measurement topics provide categories and methods; this article brings those records into an intervention review.

Keep the observations comparable

Use the same food-state basis, included categories and measurement method for each comparison. Retain period length, food quantities, meal requests and the places covered. A kilogram discarded in one short quiet period cannot be read as equivalent to a kilogram during a longer busy period without context.

Check whether an apparent reduction reflects a new exclusion or recording gap. Changing from all recorded discards to only one category changes the measurement. Missing observations should remain missing; do not treat an unweighed bin as zero waste or silently convert its volume to mass.

Follow the shifted material

In an original comparison, two nonoverlapping stages record discarded food on the same unchanged food-mass basis and comparable reporting boundary. Before, the kitchen records 20 kg and the receiver 10 kg: 30 kg total. Afterwards, they record 10 kg and 25 kg: 35 kg total.

The kitchen figure halves, but the included total rises by 5 kg. The diagram compares those accounts without proving why they changed. Redistribution, repackaging or a new destination may move a discard outside the original bin; moving it is different from preventing it. Follow the relevant next stage and retain other omissions.

Interpret the rate with its denominator

A separate example compares kitchen-only discard per completed meal service. Before: 20 kg over 100 meal services, or 0.20 kg/service. After: 10 kg over 40 meal services, or 0.25 kg/service. Absolute discard falls while the stated per-service ratio rises.

These are count-normalised observations, not nutritional portions or proof that every meal was equivalent. Food mix and service need can differ. Choose a functional comparison that matches the question, and report both totals and ratios where relevant rather than selecting whichever number looks better.

Test the causal explanation

A change between periods can reflect demand, ingredients, weather, recording or several operating changes as well as the proposed intervention. A descriptive difference alone cannot isolate the intervention’s effect. Appropriate comparison design, replication and uncertainty analysis can strengthen a relevant causal investigation.

Explain the proposed mechanism and observe whether the expected stage actually changes. Retain safety, meal sufficiency, quality, work and cost outcomes. Less recorded discard achieved by serving too little food or tolerating unsafe conditions would fail the intended food goal.

Use the review to choose the next step

A useful review states the actual result, its boundary and the remaining questions. It can support another bounded test, a modification or keeping an arrangement that demonstrably serves the goal. One favourable day is not proof of a dependable annual benefit.

For the kitchen, investigate the downstream increase and the changed service denominator before declaring prevention successful. Composting or another recovery route may change discard destination without preventing generation. Keep those milestones distinct so the next decision responds to evidence rather than a smaller-looking bin.

Worked examples

Check your understanding

1. Does a smaller kitchen discard prove less discard across the relevant chain?

   Show answer and reasoning

   No.

   A downstream stage or omitted destination can change differently.

2. Should an unmeasured discard quantity be entered as zero?

   Show answer and reasoning

   No.

   Missing evidence is not a measured absence.

3. Does sending existing scraps to compost establish that fewer scraps were generated?

   Show answer and reasoning

   No.

   Diversion changes a destination; prevention concerns the generation of excess or discard.

Key ideas

• Use comparable categories, periods and denominators.
• Follow downstream outcomes rather than only a local bin.
• Keep descriptive change, causal evidence and recovery distinct.