Multi-criteria Assessment Tool

Multicriteria Analysis Tool for Digital Agriculture

The Multicriteria Analysis Tool (MAT) brings together evidence on how digital technologies affect farming in CODECS Living Labs across Europe. Economic, environmental and social impacts are measured in very different units – euros, kilograms of CO₂, statements from interviews – so they cannot simply be added up. MAT places every assessment on a common scale from 1 (very negative) to 5 (very positive), where 3 means no change.

You decide how much each dimension of sustainability matters. A policy-maker may give more weight to the environment, a farmer to the farm economy. The tool does not hand out a single right answer. It shows how the result shifts with priorities, and makes visible why two people can disagree.

  • Economic10 criteria
  • Environmental8 criteria
  • Social9 criteria

The evidence is not uniform

The 9 Living Labs collected their assessments in different ways, for example individual scoring, workshop discussion, group consensus or conversion from an earlier workbook. They also assessed different numbers of criteria. That is why every result is shown together with its scenario, its coverage and the way the evidence was collected.

How it works

  1. Choose the evidence

    Select a Living Lab and a scenario. When a Living Lab has no assessments for that scenario, the tool uses one it does have and says so.

  2. Set the priorities

    Pick a weight profile. It decides how much the economic, environmental and social dimensions count in the overall score.

  3. Read the result

    Impact and certainty are combined into a robust score. Every score is shown with its scenario and the number of criteria behind it.

  4. Compare and rank

    Put two Living Labs side by side, rank all of them under the same weights, or score a case of your own.

What you can do with MAT

MAT is made for farmers, advisors, researchers and policy-makers. It draws on 290 assessments from 9 CODECS Living Labs and offers four ways to work with them.

  • Analysis

    Look at one Living Lab, a technology group or all the evidence under one scenario and one weight profile. See impact, certainty and the robust score for each dimension and overall, and which criteria drive the result.

  • Compare

    Put two Living Labs side by side under the same scenario and the same weights, and see where they differ in impact, certainty and the robust score. Every difference reads as the first Living Lab minus the second.

  • Rankings

    Order all Living Labs under common weights, a common scenario rule and a minimum amount of evidence. Every score comes with the number of criteria behind it, and Living Labs below the minimum are not ranked.

  • Custom case

    Enter your own impact and certainty for all 27 criteria and see how your case compares with the most specific CODECS evidence available.

Every result depends on the evidence available and on the scenario, filters and weights you choose. A higher score does not mean that one technology or Living Lab is better in general.

Go to the calculator

Methodology

MAT reproduces the calculation model of the CODECS multicriteria workbook, built in Task 7.3 at the University of Pisa. The evidence was collected by the Living Labs through questionnaires and validation workshops in Task 4.6.

A common 1–5 scale

  1. 1very negative
  2. 2
  3. 3neutral, no change
  4. 4
  5. 5very positive

Impact says whether a technology makes things better or worse. Certainty uses the same range, but says how sure the assessment is, not whether the effect is good or bad.

Weight profiles

ProfileEconomicEnvironmentalSocial
Expert weights39 %29 %32 %
Policy-maker30 %40 %30 %
Farmer / business55 %20 %25 %
Advisor / knowledge35 %20 %45 %
Equal weights33 %33 %33 %
Custom40 %30 %30 %

Within a dimension, the weight is shared equally among the criteria that were actually assessed. A Living Lab that assessed two economic criteria gives each of them a larger share than one that assessed ten. A dimension without any assessed criteria is left out, and the other weights are scaled up to fill its place.

The robust score

Robust = Impact × Certainty ÷ 5

Certainty scales an assessment down: at certainty 5 the robust score equals the impact, and lower certainty discounts it proportionally. Each record is converted on its own before anything is averaged, so a certainty can only ever discount the impact it was given for.

Two consequences worth knowing. The robust score runs from 0.2 to 5, not on the 1–5 impact scale, so it has no neutral point. And because certainty only scales downward, an uncertain warning scores lower than a confident one — the score measures how strongly a piece of evidence contributes, not whether the news is good. That is why impact and certainty are always shown next to it.

A dimension score is the average robust score of its assessed criteria. The overall score is the weighted average across the dimensions that have evidence.

Reading the results

Overall and dimension scores

  • Very strong robust contribution4.0 and above
  • Strong robust contribution3.3 to under 4.0
  • Moderate robust contribution2.7 to under 3.3
  • Weak robust contributionBelow 2.7

These bands describe how strong a contribution is, never whether the impact behind it was positive or adverse. A confident assessment of a small effect and an uncertain assessment of a large one land in the same band.

Signals for single criteria

  • Robust benefitImpact ≥ 4 and certainty ≥ 4
  • High potential / uncertainImpact ≥ 4 and certainty < 3.5
  • Robust concernImpact < 3 and certainty ≥ 3.5
  • Mixed / monitorAll other combinations

These thresholds are rules of the original workbook, not scientific cut-offs. They leave a deliberate gap: a high impact held with middling certainty falls to the catch-all rather than to its own label.

Explore the Living Lab evidence

Analyse, compare and rank what CODECS Living Labs report about digital technologies in farming, under your own priorities, or score a case of your own against their evidence.

Go to the calculator