The Atlas of the Middle

The studies we found on middle managers, nearly all first published from 2007 to 2026: 173 studies across 25 journals, 105 of them about change, each summarized in plain language and placed on the five parts of the Unfrozen Loop. Counts are as of 30 October 2026 (Version 2).

Each dot is one study. A study that informs several parts appears in each one. Hover or tap a dot to read it.

Map of the 173 studies

  1. 1Surface both ways27
  2. 2Make sense together, top included37
  3. 3Start it close to the work67
  4. 4Review on a rhythm, then hold the pace6
  5. 5The middle's own weather42
  6. ·About the middle, not one part49
Supports the partMixedChallenges itAbout the middle, not one partBigger dot = stronger evidence

All studies

Browse all 173 study pages, each with its grade and summary.

How the Atlas was built

Why the middle. In 21 core management journals, 2007 to 2018, English-language research abstracts mentioning executives outnumber those mentioning middle managers 693 to 65 by keyword: about 10 to 1 after correcting for missed papers (a provisional estimate until an independent human coder checks it). The Atlas gathers the research there is on the middle. Read the study behind the count.

The studies come from the Unfrozen Research Index, a database of about 35,000 articles from 29 management and organization journals. Version 1 (October 2026): a search for middle-manager research on change first published from 2007 to 2018 found 196 candidates; 117 were judged to be about middle managers. One of them, Noda and Bower, turned out on checking to be a 1996 classic that the index had dated 2007 (the index has carried its print year since 1 November 2026); it stays in. Version 2 (30 October 2026) added studies first published from 2019 to 2026, from the same index journals (core and expansion), screened from their abstracts with the same coding guide. Of 71 candidates, 59 were judged to be about middle managers, and 56 were added after two independent fact-check passes and my approval. Three were held back: one appears to reuse another study's interviews, one reports no results in its abstract, and one is an essay. Each study was coded for which parts of the Unfrozen Loop it informs, whether it supports, complicates or challenges that part, and how strong its evidence is.

Every summary was written in plain language by Unfrozen, not copied from the publisher, then checked twice against the study's abstract or full text. Coding, including the 56 studies added in Version 2, is exploratory and provisional until an independent human coder checks it.

Data and reuse. Every study's record is also in one file, studies.json, with its page link, grade, Loop parts and summary. The Atlas's own summaries, grades and Loop coding, on these pages and in that file, are shared under CC BY-NC 4.0. The license covers those only, not the original papers, which stay with their authors and publishers. Cite as: Lopez, G. (2026). Atlas of the Middle. unfreeze.us/atlas/

Found a mistake? Suggest a correction or email info@unfreeze.us. Corrections are made and noted.

Also from the index: The Backbone, a case study of the middle with a mini atlas of 18 graded studies on veterans and NCOs, the military's middle managers.

How studies are graded

The grade measures evidence strength only, meaning how much weight a study's own design, execution and finding can bear. I ask three questions of each study.

  • Design. Did the researchers test the idea against a comparison they controlled, such as a randomized trial? If not, did they study real organizations systematically, through a survey, a comparison of organizations or long fieldwork in one? Or is it an argument, a practitioner article or an account of one manager or one program?
  • Execution. Is the sample big enough for that design? And does the study avoid more than one of five common limits: data from a single organization, accounts gathered years later, causes and results reported by the same people at the same time, researchers who ran the change they studied, and students standing in for managers?
  • Clarity. Does the study show a specific link between a condition and an outcome, rather than a list of themes, an untested proposal or a result nobody measured? And is its main result clear rather than marginal?

A grade of High needs a controlled test that passes both of the other questions. A controlled test that fails either one is Medium. A systematic study is Medium if it passes at least one of the other two questions and Low if it passes neither, and an argument, a practitioner article or a single account is always Low.

  • HighA controlled test, big enough, with a clear main result.
  • MediumA systematic study of real organizations, or a controlled test with a clear limit.
  • LowAn argument, a practitioner account, a single account, or a study that fails on both execution and clarity.

By "big enough" I mean at least 100 people or 20 randomly assigned groups for an experiment, 200 respondents or 50 organizations for a survey, and 20 informants or a year of repeated contact for fieldwork. These cutoffs are my conventions, not a published standard.

Three things stay out of the grade. One is how closely a study bears on the point it is used for. Another is fit to the Unfrozen Loop, because fit to my own model isn't evidence that a study is strong. The third is how far it was read: a grade made from an abstract or from coverage is provisional and could change with a full reading.

The Atlas's first grades were drafted mostly from the studies' abstracts, and they also weighed how well a study informed the Loop. The studies used in the book were checked again against this rubric, and five grades changed. The Nokia study (Vuori & Huy, 2016), the trucking trial (Logan & Ganster, 2007) and the support-to-mutiny study (Huy, Corley & Kraatz, 2014) moved from High to Medium; the Philips study (Karsten et al., 2009) moved from Low to Medium; and the Lego study (Lüscher & Lewis, 2008) moved from Medium to Low. The tech-firm strategy study (Huy, 2011), which is not in the book, was checked the same way and moved from High to Medium. The Atlas now shows those grades, so no study in it is currently rated High. The other studies keep their first grades until they are checked the same way.

Suggest a correction

If you think a grade is wrong, tell me. If a challenge holds up, I'll correct the Atlas page and the next edition of the book.

Your email is used only to reply about this challenge. It doesn't sign you up for anything.