The Built Review rates a factor on the studies in its collection that measure it

A figure holds for the rooms it came from

The authors of a study chose the building, the people and the way to measure them, and their figure describes those people in that building. A team elsewhere that chooses differently gets another figure from its own conditions. I take a figure as evidence when I can name its study, its sample size and its method. The sample size is the number of people or rooms a study measured. Two figures from two studies can be compared when each comes with its sample size and its method.

For each figure in The Built Review, I name its study and give a dated source. For a point estimate, a single value a study calculates from its sample, I add the sample size and the method.

The Built Review has one author

The Built Review is an independent analysis of how the built environment affects the people who use it. I research, write and publish it alone, from Germany, and it has no editorial board. Its readers are real-estate investors and developers, corporate real-estate and workplace leaders, the architects who work with them, and HR. It covers six sectors: workplace, healthcare, education, hospitality, housing and urban.

The home page counts 15,314 scientific studies in the collection. For each report, I search these studies and analyse the ones on its question. I write The Built Review apart from neuecon partners. No project or client of neuecon partners changes a finding of The Built Review. I publish a report without a fixed schedule.

A factorsheet rates all the studies on one cause and its effect

A factorsheet covers a feature of a building and its effect on people, for example daylight and sleep. neuecon.com calls it an evidence sheet. I decide for each factor whether the studies in the collection are enough for a sheet. The sheet counts the studies that my review finds support the effect and those it finds contradict it, and for each study it names the author, the year and the design, with a link to the paper. A studysheet covers a single study and is written ready to cite, for people who are not scientists.

Reviews and randomised trials weigh most

Each study takes its grade from its design. A systematic review is Strong, and so is a randomised trial, in which chance decides who gets the treatment. A cohort study, a case-control study and an experiment without random assignment are Moderate. A cross-sectional study, a before-and-after comparison, a case study, a survey, an expert opinion and a figure quoted from another paper are Limited, and a study with no design on record is Unrated.

A factorsheet is Strong once two reviews or randomised trials support its effect, or once five studies support it and two of them have at least a Moderate design. Either way, at least 30% of the supporting studies must come from the last 15 years. A factorsheet is Limited with two supporting studies or fewer, or with no study of at least Moderate design. Every factorsheet between these is Moderate.

My rule follows the method of the Community Guide, a US guide to public health measures, which the public health researcher Peter Briss and colleagues set out in 2000. I can lower the grade of a factor and never raise it, as reviewers do in GRADE, the rating scale of the GRADE Working Group. I lower it when the studies disagree or match the factor only in part.

A disagreement between studies is part of the evidence

When two credible sources disagree, I report both. On a factorsheet, the contradicting studies stay in the list next to the supporting ones, and the sheet gives the count on each side.

Sources

• The Built Review, "About". https://thebuiltreview.com/about
• The Built Review, "How it works". https://thebuiltreview.com/how-it-works
• The Built Review, home page, with the count of 15,314 studies. https://thebuiltreview.com/
• The Built Review, "Factorsheets". https://thebuiltreview.com/factors
• Briss, Zaza, Pappaioanou, Fielding, Wright-De Agüero, Truman and colleagues 2000, "Developing an evidence-based guide to community preventive services: methods", American Journal of Preventive Medicine 18(1):35-43. https://doi.org/10.1016/S0749-3797(99)00119-1
• GRADE Working Group. https://www.gradeworkinggroup.org/