Evaluation Framework
Roger Kaufman looked at Kirkpatrick's four levels and made two changes: he split the first level in two, and he added a level above organizational results for impact on society. One of those changes is quietly useful for everyday evaluation. The other is mostly discussed and rarely measured.
Kaufman published the model with John Keller in 1994, under the title "Levels of evaluation: beyond Kirkpatrick." It sits inside his larger Organizational Elements Model, which is why the levels run from resources all the way out to society.
| Level | Name | What it evaluates |
|---|---|---|
| 1a | Input | The resources themselves: materials, content, facilities. Were they available and fit for purpose? |
| 1b | Process | The delivery: how the program ran, whether participants found the experience acceptable and efficient. |
| 2 | Acquisition (micro) | Did individuals and small groups learn it? Kirkpatrick's Level 2. |
| 3 | Application (micro) | Are they using it? Kirkpatrick's Level 3. |
| 4 | Organizational output (macro) | Payoffs for the organization. Kirkpatrick's Level 4. |
| 5 | Societal outcomes (mega) | Contributions to clients and society beyond the organization. |
The middle of the model is Kirkpatrick wearing different labels, and Kaufman never pretended otherwise. The differences live at the edges.
A standard reaction survey blends two different complaints into one score. "The handouts were unreadable" and "the facilitator lost the room after lunch" are different problems with different owners, and a single satisfaction question averages them into mush. Kaufman's input/process split fixes that at the question level.
Input (1a)
"The program materials were clear and usable." A question about the artifacts, which the designer owns.
Process (1b)
"The sessions ran at the right pace and the time was well used." A question about delivery, which the facilitator owns.
For a consultancy running programs with multiple facilitators, the split matters even more: pair the process questions with per-facilitator survey links and you can see whether a weak score belongs to the materials or to one delivery. You don't need to adopt Kaufman's whole model to take this. Most good evaluation surveys already do, without crediting him.
Level 5 asks what the program contributed to clients and society. Kaufman took it seriously: his "mega planning" work argues organizations should define their purpose in terms of societal value added, and evaluation should reach that far.
In commercial L&D, almost nobody measures it. The chain of attribution from a leadership workshop to a societal outcome is too long to survive contact with data, and most organizations stop at Level 4 for the same reason most stop at Level 1: each step up costs more and proves less cleanly.
Where the mega level earns its keep is in sectors whose mission already is the societal outcome: public health, education, development programs. If your client is a government agency or an NGO, Kaufman gives you a framework-level reason to evaluate against mission rather than internal metrics, and that may be exactly the language their funders speak.
Input and process questions in one short survey, with per-facilitator links so you know which is which.
Start your free trial30-day free trial. No credit card required.