---
title: "Lewin's three-step model: what the 1947 paper says"
description: "Is unfreeze, change, refreeze really Lewin's model? The 1947 text set against Cummings, Bridgman and Brown's history, with a worked check for SME owners."
url: "https://staging.khoi.co.uk/en/blog/lewin-ba-buoc-co-that-khong"
lang: en-GB
author: "Lê Quốc Khôi"
date: 2026-11-03
updated: 2026-11-03
topics: ["Change management"]
tags: ["Change management", "Kurt Lewin", "Group dynamics", "Management history"]
---

# Did Lewin really create the three-step model? What his 1947 paper says

## Key takeaways

- Lewin (1947) wrote that a successful change has three aspects: "unfreezing (if necessary) the present level", "moving to the new level" and "freezing group life on the new level" (Lewin, 1947, p. 35). It is one sub-section of a 37-page article, with no diagram of its own and no label saying "model".
- Cummings, Bridgman and Brown (2016) argue that Lewin never developed a three-step model for change agents. It took form after his death (Cummings et al., 2016).
- The word "refreezing" is not in Lewin's paper. The authors found it first in a 1950 conference paper by Festinger and Coyle (Cummings et al., 2016).
- The textbook version drops the group, which Cummings and colleagues call Lewin's main unit of analysis (Cummings et al., 2016): the place where change happens, with each level of group life held by a force field (Lewin, 1947, p. 35).

Many change-management textbooks show three boxes joined by arrows: unfreeze, change, refreeze, with Kurt Lewin's name beside them. The owner of a 12-person motorbike spare-parts shop, about to change how stock is counted at closing time, thinks of turning the boxes into a three-week plan. Two questions come first. What did Lewin actually write, and who drew the boxes? The shop is a hypothetical example.

## What does the textbook three-step model look like?

A straight line of three steps. The library on this site records step one as challenging the status quo, step two as carrying out the change while involving people, and step three as building the new behaviour into routines and sustaining it with feedback and reward.

> **Figure:** [Lewin's three-step model](https://staging.khoi.co.uk/en/framework/lewin-three-step) The three steps as textbooks present them. According to Cummings, Bridgman and Brown (2016), Lewin did not draw this linear diagram.

Cummings and colleagues open by citing authors, textbooks among them, who call the model the "fundamental" or "classic" approach to change, plus a line from Schein saying that the fundamental assumptions about change in a human system derive from Lewin (1947) (Cummings et al., 2016).

## What did Lewin actually write in 1947?

A short sub-section titled "Changing as Three Steps: Unfreezing, Moving, and Freezing of Group Standards" (Lewin, 1947, p. 34). It exists. But it is about the standards of a group, not about a change agent.

Lewin starts from an observation: a change towards a higher level of group performance is frequently short-lived, and after a "shot in the arm" group life soon returns to the previous level (Lewin, 1947, p. 34). A successful change, he writes, has three aspects: "unfreezing (if necessary) the present level", "moving to the new level" and "freezing group life on the new level" (Lewin, 1947, p. 35).

"Freezing" here does not mean writing the change into procedure. He explains that "the new force field is made relatively secure against change" (Lewin, 1947, p. 35). Unfreezing, he adds, "may involve quite different problems in different cases", and sometimes it takes a deliberate "emotional stir-up" to break the shell of complacency (Lewin, 1947, p. 35). Summing up on page 36, he speaks of "unfreezing, change of level, and freezing on the new level" (Lewin, 1947, p. 36). I could not find the word "refreezing" anywhere in the paper.

Cummings and colleagues add that the passage sits in the 24th of the article's 25 sub-sections and carries no evidence or illustration of its own. They also searched 67 articles, chapters and books by Lewin, plus the archives at Iowa and the Tavistock Institute, and found no other origin for the three steps (Cummings et al., 2016).

What Lewin gave room to is the group-decision work that follows. On the maternity ward of the State Hospital of Iowa, a nutritionist spent about 25 minutes with a single mother, and the same time with a group of six; at four weeks every mother in the decision group was giving the baby the advised amount of cod liver oil (Lewin, 1947, pp. 35–36). He adds a warning: managers rushing into a factory to raise production by group decisions "are likely to encounter failure", because "there are no patent medicines and each case demands careful diagnosis" (Lewin, 1947, p. 36).

## Who drew the boxes in the textbooks, and when?

According to Cummings and colleagues, a chain of writers over more than forty years, mostly after Lewin died:

| Year | Who | What the paper records |
|---|---|---|
| 1947 | Marrow, Likert, Lippitt | Tributes to Lewin do not mention three steps |
| 1950 | Festinger and Coyle | First "refreezing" the authors could find |
| 1958 | Lippitt and co-authors | Call it Lewin's three-phase model and build a seven-phase model for "change agents" on it |
| 1961 | Schein | First "refreezing" in a management setting; the steps become phases of influence, turning a way of observing group change into a tool for the person leading it |
| 1979 | Robbins, textbook first edition | One sentence on three stages, footnoted "Lewin, 1951" with no page |
| 1986 to 1993 | Robbins, third to sixth editions | A linear diagram appears with no mention of Lewin, then carries a title calling it Lewin's classic model |

Why a "foundation" only in the 1980s? The paper names five conditions. Best-sellers such as *Theory Z* (1981) and *In Search of Excellence* (1982) created demand for memorable frameworks, and textbooks added step diagrams that made big classes easier to teach. Change management also overtook organisation development in journal articles: by the paper's account of a study by Palmer and colleagues, organisation development dominated in 1980 and change management had completely surpassed it by 1993. A new field wanted a respectable ancestor, and Lewin, a social psychologist with an experimental record, fitted (Cummings et al., 2016). In the paper's view, the later complaint that the approach is obsolete in an unpredictable world has little to do with what Lewin wrote.

## What does the textbook version leave out?

The group. The 1947 paper treats the group, not the manager or the individual employee, as the place where change happens, and each level of group life as set by a force field (Lewin, 1947, p. 35).

| Textbook reading (as Cummings and colleagues describe it) | Lewin, 1947 |
|---|---|
| A leader overcomes each employee's resistance | It is usually easier to change individuals who form a group than to change any one of them separately (Lewin, 1947, p. 34) |
| Resistance is an obstacle to break | One cause of resistance lies in the relation between the individual and the value of group standards; change the standard and that part of the resistance disappears (Lewin, 1947, pp. 34, 39) |
| The organisation is a block of ice | A group's production level is a process like a river, which keeps changing its elements while speed and direction stay the same (Lewin, 1947, p. 15) |
| Persuade to overcome resistance | Two routes: add forces in the desired direction, or diminish opposing forces (Lewin, 1947, p. 26) |

The last row is the easiest to miss. In the production experiments, Lewin remarks that the methods seemed to rest, at least in part, on reducing the forces that keep production down rather than adding forces upwards (Lewin, 1947, p. 26). If you think only of adding force, nobody asks which opposing force could be removed.

Cummings and colleagues also mention the Harwood Manufacturing Corporation, where Lewin's team found the most successful approach was to have all stakeholders work through specific processes before agreeing the way forward (Cummings et al., 2016). Lewin's view of resistance points the same way as the argument that resistance belongs to the system, set out in [the post on four theories of resistance](https://staging.khoi.co.uk/en/blog/bon-ly-thuyet-khang-cu-thay-doi).

## How can I use Lewin's three aspects without repeating the textbook boxes?

Treat them as questions about a group, not as three weeks on a calendar. The table compares two plans for one hypothetical job: the 12-person motorbike spare-parts shop wants two people to count stock together each evening and enter it on one shared sheet, instead of each keeping a separate book.

| Question | Textbook plan | Plan from the 1947 paper |
|---|---|---|
| Which group standard does the change touch? | Week 1: a meeting on why change is needed | Ask what standard the team holds (hypothetical: "whoever counts is liable") and who upholds it |
| What pushes, what holds back? | Weeks 2 to 3: training, personal benefits | Write two columns of forces; find which opposing force can go before adding pressure (Lewin, 1947, p. 26) |
| Who decides the method? | The owner issues the new procedure | Let the team choose a trial version; Lewin warns this takes skill and general conditions (Lewin, 1947, p. 36) |
| What keeps the new level in place? | Week 4: add it to the procedure, tie it to bonuses | Look for an arrangement that holds itself, then check at two and four weeks as his experiments did (Lewin, 1947, p. 35) |

The last row is my application, not Lewin's words. He says only that it is sometimes possible to set up "an organizational set up which is equivalent to a stable circular causal process" (Lewin, 1947, p. 35). In the hypothetical shop, that might be counting pairs who swap partners weekly, so that everyone is cross-checked. Nobody has tested the idea.

The sourced reading fits if the change touches work a team does together and you have room to meet, trial and adjust. It does not fit if a rule forces the change within a day, or if it touches one person only; there the three-step list still works as a quick checklist.

## Going deeper: what does the Cummings paper rest on, and where does it stop?

The paper takes a counter-historical approach after Michel Foucault: instead of telling history as a climb from honourable foundations, it asks how the foundation was built (Cummings et al., 2016). The textbook analysis uses Robbins as its main example, because the book has been in print since 1979 and a preface to one edition, quoted in the paper, calls it the best-selling organisational behaviour textbook.

The authors suggest teaching the change chapter around the group (Cummings et al., 2016). They describe later frameworks built on the three steps, such as [Kotter's eight-step change model](https://staging.khoi.co.uk/en/framework/kotter-8-steps), as innovating while staying true to Lewin's spirit. The library also has an entry on [the critique of refreezing](https://staging.khoi.co.uk/en/glossary#refreezing-critique).

Three limits. "No other origin" is a negative result and holds only within their search. The textbook analysis rests mainly on one series of books. And the paper does not measure whether using three steps helps organisations change better.

What did I read? The authors' manuscript of Cummings and colleagues, from the Victoria University of Wellington open repository, not the typeset journal version, so I give no page numbers for it. Lewin (1947) I read in full from a scan on the Internet Archive; page numbers follow the printed numbers at the foot of each page and match the page 36 that Cummings and colleagues cite for "may be thought of". I did not read Field Theory in Social Science, Lewin's chapter in the Newcomb and Hartley volume (1947), the texts by Lippitt, Schein and Festinger, or the Robbins editions. Everything said here about them is as Cummings and colleagues report it.

## Frequently asked questions

### Who was Kurt Lewin, and where did he work before the 1947 paper?

According to Cummings, Bridgman and Brown (2016), Lewin studied philosophy and psychology and worked at the Psychological Institute of the University of Berlin until 1933, when he left Germany. In the United States he studied children's behaviour at Cornell, then from 1935 to 1945 worked at the Iowa Child Welfare Research Station, where he gave his title as Professor of Child Psychology. "Frontiers in group dynamics" appeared weeks after he died, and it is not a management book.

### How should I reference the three-step model in a report?

Cite Lewin (1947) with page numbers for what he wrote, and cite whoever built the model for the rest. Cummings and colleagues found that among nearly 10,000 Google Scholar citations of "Lewin, 1951" (Field Theory in Social Science, edited by Cartwright four years after his death), none of the 100 most cited gives a page number. The passage on three steps in that book runs to 131 words in 338 pages, so without a page number your reader cannot tell which passage you mean.

### Does the Cummings paper say the three-step model is useless?

No. Its question is where the model came from and how it became a "foundation", not whether it helps managers. The paper does not test the three steps inside real organisations, so no conclusion about effectiveness follows from it.

## References

- Cummings, S., Bridgman, T. and Brown, K.G. (2016) ‘Unfreezing change as three steps: rethinking Kurt Lewin's legacy for change management’, *Human Relations*, 69(1), pp. 33–60. https://doi.org/10.1177/0018726715577707
- Lewin, K. (1947) ‘Frontiers in group dynamics’, *Human Relations*, 1(1), pp. 5–41. https://doi.org/10.1177/001872674700100103

---

Canonical HTML version: https://staging.khoi.co.uk/en/blog/lewin-ba-buoc-co-that-khong