Prospect-Refuge Theory: What It Claims, What Holds Up

6 cited studies · updated 2026-11-11

Prospect-refuge theory, proposed by geographer Jay Appleton in 1975, holds that people prefer places where they can see out widely while feeling sheltered.#

Last updated: 2026-11-11. Studies are summarized from the papers in DwellScore's research index and from publisher records; each section links the original.#

The Short Answer#

Prospect-refuge theory says we like places that offer prospect, an open view of our surroundings, and refuge, a protected spot to watch from. Appleton argued the preference is inherited from ancestors who survived by seeing danger and opportunity without being exposed. Tests support the prospect half well, especially open views in streets and rooms, but the evidence for refuge is mixed, and a 2016 meta-analysis concluded that support for the theory as a whole "remains inconsistent."#

Where the Theory Came From#

Appleton was a British geographer interested in why some landscapes and landscape paintings please us. In The Experience of Landscape (1975) he built on what he called habitat theory: that aesthetic pleasure in a landscape comes from perceiving it as a good place to meet our biological needs. Prospect-refuge theory is a specific version. A prospect is a place with an unobstructed view; a refuge is a place to hide or shelter. The ideal is a vantage point that has both. A related term, hazard, covers features that signal danger.#

The theory was built for landscapes. The architectural historian Grant Hildebrand carried it indoors. In The Wright Space (1991) he analyzed Frank Lloyd Wright's houses and argued that their most consistent pattern is a low, sheltered core, typically around the hearth, opening onto long views through bands of windows and terraces. Since then, prospect and refuge have become standard vocabulary in architecture and biophilic design, often cited as if proven.#

Prospect-refuge theory belongs to the same evolutionary family as the biophilia hypothesis, and it shares that hypothesis's main weakness: an evolutionary origin is easy to propose and hard to test.#

Study 1: The Original Theory#

Citation: Appleton, J. (1975). The Experience of Landscape. Wiley.#

Scope:#

  • A theoretical book drawing on landscape painting, literature and geography; it reports no experiments#

Key Claims:#

  • Humans take pleasure in landscapes that signal opportunities to see (prospect) and to hide (refuge)#
  • The preference is inherited from our evolutionary past and operates without conscious thought#
  • Landscapes that balance the two are judged most satisfying#

Limitation: The book proposes the theory; the testing came later and from other researchers.#

Study 2: The Theory Applied to Architecture#

Citation: Hildebrand, G. (1991). The Wright Space: Pattern and Meaning in Frank Lloyd Wright's Houses. University of Washington Press.#

Scope:#

  • Spatial analyses of Frank Lloyd Wright houses, from the Prairie houses to Fallingwater and the Usonians#

Key Claims:#

  • Wright's houses repeatedly pair sheltered, low-ceilinged interiors with expansive outward views#
  • Hildebrand adds complexity and order as further qualities derived the same way, and argues the combination is what makes these houses distinctive#

Limitation: This is architectural interpretation, not a test. It shows that the pattern can be found in admired buildings, not that it causes people to prefer them.#

Study 3: Fear of Crime on a Campus#

Citation: Fisher, B.S., & Nasar, J.L. (1992). Fear of crime in relation to three exterior site features: Prospect, refuge, and escape. Environment and Behavior, 24(1), 35-65. doi:10.1177/0013916592241002#

Design:#

  • Three studies on the Ohio State University campus: observation of behavior, responses to site plans, and on-site ratings of safety in areas that varied in prospect, refuge and escape#

Key Findings:#

  • Fear of crime was highest where potential offenders had refuge (places to hide) and potential victims had little prospect or escape#
  • The authors suggest designers can reduce both fear and opportunities for crime by adjusting these features#

Why it matters here: Refuge is not always good. The same hiding place that shelters you can shelter someone else, and what counts is who controls it.#

Study 4: Enclosure and Judged Safety in Street Scenes#

Citation: Stamps, A.E. (2005). Enclosure and safety in urbanscapes. Environment and Behavior, 37(1), 102-133. doi:10.1177/0013916504266806#

Design:#

  • Ratings of 21 artificial street scenes that varied in walls, ground, lightness, depth of view and open sides#

Key Findings (from the abstract):#

  • Replicated earlier work: impressions of enclosure were driven by the share of the scene covered by walls and ground, how light or dark it was, the depth of view, and the number of open sides#
  • Judged safety was most strongly correlated with lightness of the scene (r = .82)#

A companion paper by the same author reported four experiments in which impressions of enclosure depended more on whether one could see through gaps, but impressions of safety depended more on whether one could walk through them (Stamps, 2005, Environment and Behavior, 37(5), 587-619, doi:10.1177/0013916505276741).#

Limitation: Artificial scenes rated on screen; the results describe perception, not health.#

Study 5: When Nature Is Not Restorative#

Citation: Gatersleben, B., & Andrews, M. (2013). When walking in nature is not restorative: The role of prospect and refuge. Health & Place, 20, 91-101. doi:10.1016/j.healthplace.2013.01.001#

Sample Size / Design:#

  • An online survey (n = 269) rating natural scenes for restorative potential#
  • An experiment with walks through real outdoor settings (n = 17) and videos of the same walks in a laboratory (n = 17)#

Key Findings:#

  • Natural settings with high prospect and low refuge (open, with clear sight lines) were restorative#
  • Settings with low prospect and high refuge (dense woods, places for others to hide) were not, and may have increased stress and attention fatigue#

Limitation: The experimental groups were very small. Here refuge meant places to hide in dense vegetation, which a walker may read as places where someone else could hide. That is close to the reverse of the sheltered vantage point Appleton had in mind, and it shows how loosely the terms are used across studies.#

Study 6: The Meta-Analysis#

Citation: Dosen, A.S., & Ostwald, M.J. (2016). Evidence for prospect-refuge theory: A meta-analysis of the findings of environmental preference research. City, Territory and Architecture, 3, 4. doi:10.1186/s40410-016-0033-1#

Design:#

  • A review and classification of quantitative studies testing prospect-refuge theory and related preference variables (complexity, mystery), grouped by natural, urban and interior settings. The authors say their aim was to classify findings for and against the theory, not to judge each study's rigor.#

Key Findings (from the abstract):#

  • Urban and interior studies supported prospect and were more neutral about refuge#
  • Studies of natural environments gave evidence for both prospect and refuge, but also included evidence against and neutral results#
  • Some complexity in interiors seems to be preferred, but how much or where is unclear; results for mystery were mostly neutral or contrary#
  • Conclusion: "The quantitative evidence for prospect-refuge theory remains inconsistent." The results most often cited in architecture come from studies of natural landscapes, not of interiors or cities#

How Strong Is the Evidence?#

  • Prospect (open views): moderate and fairly consistent. People prefer scenes and rooms with open views, especially in built settings.#
  • Refuge: mixed. Neutral in urban and interior studies, and sometimes negative when refuge means hiding places for others (Studies 3 and 5).#
  • The balance of both: weakly tested. The specific claim that the combination is best has less direct support than either half.#
  • An evolutionary origin: untested. Studies measure present-day preference; none can show where it came from.#
  • Effects on health: thin. Most studies measure preference or perceived safety, not health outcomes.#

DwellScore's evidence matrix grades its Prospect-Refuge row C with a Minor magnitude, the lowest grade among the factors in its Attention Restoration score.#

What It Means for Choosing a Home#

Everything in this section is application, not proven effect: the theory suggests what to look for, but no study has shown that these features improve residents' health.#

  • Window seats and bay windows. A bay or a window seat is a classic prospect-refuge spot: sheltered on three sides, with a wide view. Bay windows also tend to bring in light from more than one direction.#
  • Parlor floors. In a brownstone, the parlor floor sits half a story above the street, with tall windows. You see the street; passers-by see less of you.#
  • Stoops. A stoop gives a raised, semi-private perch over the sidewalk, a place to watch street life with the house at your back.#
  • A set-back room with a street view. A sitting area a few feet back from a window, looking out over a street or a tree canopy, offers a view without exposure.#
  • Top floors vs. garden floors. A high floor maximizes prospect (long views, more sky and light) but can feel exposed if the rooms are all glass. A garden floor gives refuge and a private outdoor space, but less view and light; it can feel enclosed if the yard is ringed by tall walls. The evidence favors prospect more consistently than refuge.#
  • Light over enclosure. In Stamps's study, how light a scene looked predicted judged safety more strongly than any enclosure variable. A bright room may matter more than a cozy one; see natural light and circadian health and the apartment orientation guide.#

How DwellScore Uses It#

Prospect and refuge enter DwellScore's Attention Restoration score as one component, measured from street enclosure: the ratio of average building height to street width around the listing, from building-footprint and street data, with lot size from the city's PLUTO data as a fallback. Streets that are neither wide open nor canyon-like score highest. Because the underlying evidence is graded C, read this component as the least certain part of the score: it measures a street shape the theory favors, not a proven health effect. For how this fits a purchase, see buying an apartment in NYC.#

References#

  • Appleton, J. (1975). The Experience of Landscape. Wiley.#
  • Dosen, A.S., & Ostwald, M.J. (2016). City, Territory and Architecture, 3, 4. doi:10.1186/s40410-016-0033-1#
  • Fisher, B.S., & Nasar, J.L. (1992). Environment and Behavior, 24(1), 35-65. doi:10.1177/0013916592241002#
  • Gatersleben, B., & Andrews, M. (2013). Health & Place, 20, 91-101. doi:10.1016/j.healthplace.2013.01.001#
  • Hildebrand, G. (1991). The Wright Space: Pattern and Meaning in Frank Lloyd Wright's Houses. University of Washington Press.#
  • Stamps, A.E. (2005). Enclosure and safety in urbanscapes. Environment and Behavior, 37(1), 102-133. doi:10.1177/0013916504266806#
  • Stamps, A.E. (2005). Visual permeability, locomotive permeability, safety, and enclosure. Environment and Behavior, 37(5), 587-619. doi:10.1177/0013916505276741#

Where these studies are used in DwellScore