Methods FAQ

The Eviction Lab is a research organization dedicated to studying the prevalence, causes, and consequences of eviction. The database it has built represents the largest accumulation of U.S. court records related to eviction ever compiled.

The data we collected is composed of formal eviction records from all 50 states and the District of Columbia. Eviction records include information related to an eviction court case, such as defendant and plaintiff names, the defendant’s address, monetary judgment information, and an outcome for the case. We combined these records with demographic information from the Census to paint a better picture of the areas in which these evictions are happening.

No. Evictions can happen outside of the courtroom, as when landlords pay renters to leave or execute illegal lockouts. There is some evidence that “informal evictions” are more common than “formal”, court-ordered evictions. Informal evictions are not captured in our dataset. Moreover, while we have tried to collect every recorded court-ordered eviction case, going back to 2000, some records were unavailable. Some courts seal eviction cases; others have not archived data; still, others make recording eviction cases time-consuming and difficult.

First, we requested a bulk report of cases directly from the courts. These reports included all recorded information related to eviction-related cases. Second, we conducted automated record collection from online portals, via web scraping and text parsing protocols. Third, we partnered with companies that carry out manual collection of records, going directly into the courts and extracting the relevant case information by hand.

We have accumulated over 99.9 million records related to eviction. The Eviction Lab directly collected court records from a number of states. But many states either did not centralize their eviction data or were unwilling to release this information. Accordingly, the Eviction Lab then purchased more comprehensive datasets of public eviction records from two companies: LexisNexis Risk Solutions and American Information Research Services Inc.

We also collected state-reported, county-level statistics on landlord-tenant cases filed from over half of the states. This information was collected either from online reports or by contacting state judiciaries directly.

A “filing rate” is the ratio of the number of evictions filed in an area over the number of renter-occupied homes in that area. An “eviction rate” is the subset of those homes that received an eviction judgment in which renters were ordered to leave. The filing rate also counts all eviction cases filed in an area, including multiple cases filed against the same address in the same year. But an eviction rate only counts a single address who received an eviction judgment.

For the denominator of our rate, we used the number of occupied renting households in each area. Information on the number of renter homes in an area comes from the U.S. Census and ESRI Business Analyst demographic estimates.

Despite aggregating more than 99 million eviction records and 30,000 county-level annual reports on eviction filings—representing an average of 2.6 million annual eviction filings—our records are still missing reliable data for counties that were home to 33% of renting households.

To overcome this challenge, we developed a statistical model that drew on different sources of data to estimate the annual prevalence of eviction filings and households threatened with eviction for the missing counties. This model is based on socioeconomic factors that are detailed in our Methodology Report (PDF).

All estimates contain uncertainty. To express this uncertainty, the estimates are accompanied by minimum and maximum credible intervals. We have confidence that the true eviction filing rate lies between the minimum and maximum credible intervals. The narrower the interval, the more confidence we have in our estimate.

An eviction filing does not necessarily mean that the threatened household leaves its home. A common practice is what is known as “serial filing,” where a landlord issues a series of eviction filings against a single household.

We created the households threatened rate so that we can clearly identify how many households annually receive at least one eviction filing. When calculating this rate, we looked at the number of households receiving an eviction filing per 100 renter households. The eviction filing rate (the number of filings per 100 renter households) is still available in our database, and this will measure the total of evictions filed, no matter how many happened to the same tenants.

We can use information in our case records to identify when a household receives multiple eviction filings in a given year. By counting the number of households who ever received a filing in a given year, we can calculate the households threatened rate. The eviction filing rate will always be higher than the households threatened rate.

To create the best estimates, all data we obtained underwent a rigorous cleaning protocol. This included formatting the data so that each observation represented a household; cleaning and standardizing the names and addresses; and dropping duplicate cases. The details of this process can be found in the Methodology Report (PDF).

Eviction records contain addresses. We had these addresses geocoded and spatially joined by the Environmental Systems Research Institute (ESRI). This process allowed us to match each address record to a latitude and a longitude. Then, by overlaying the geographic boundaries on top of these points, we can observe how many evictions took place in that area.

We used a common practice for missing data called “imputation.” For more information on this, please see the Methodology Report (PDF).

We linked individual records from overlapping data sources. Through this procedure, we determined that the content of the records (e.g., address, name information) had a high rate of validity. In addition, we obtained statistics from 27 states, New York City, and the District of Columbia that gave us numbers of cases filed per year at the county level. These figures allowed us to know what proportion of all cases filed in an area were present in our data.

Yes. You can download the full report (PDF) here. You can also download the original methodology report, which accompanied the release of our original 2000-2016 data set.

We’re here to help. Please contact us via email at