In anticipation of public testimony on the subject of homeless encampments, the City of Redmond emailed data to concerned residents. This email included data on crime reports, normalized as the average number of calls per day, comparing times when an encampment was present and not present. While I appreciate the opportunity to study data, I question the accuracy of the conclusions made in the email and whether the analysis was done by someone with a doctorate in statistics or statistical criminology.

The letter contains contradictory statements. No “correlation” was found between encampments and crime report increases. But it also states that no review was conducted concerning the number of incidents, of any, that were actually related to an encampment.

Nevertheless the e-mail contained a “conclusion” stating, “The data shows that over the last two years there has been an increase in police calls for service and reported crime when we have had an encampment in Redmond, compared to when we haven’t had an encampment.”

No mathematical basis was provided for this hypothesis. There is no mathematical analysis proving that the variations in the data were any more than random fluctuations, or “noise,” Furthermore, the study shows no attempt to consider “confounding,” the presence of other variables that influence both an independent and dependent variable.

For example, one would expect seasonal variations in crime data. But in the Calls for Service section, data from January through late August 2014 is compared with data for June through October 2015. Clearly the latter includes a much higher percentage of summer days affecting the average. A conclusion is being made based on comparing apples and oranges.

Furthermore, there is data that specifically contradicts the alleged conclusion.

The largest number of incidents per day actually occurred when there was no encampment, May-June 2015.

The average for auto theft was, by far, the lowest when Camp Unity was present. Does this mean that the presence of homeless persons walking in the neighborhood actually deters auto theft? If so, perhaps some of the residents of Camp Unity should be hired as part-time patrol officers. Then they could also have a place to live while protecting the community.

The text below the data set specifically states that burglaries and car prowls did not come from Camp Unity. In other words, data is presented, but the accompanying text states that it is not relevant.

The lowest average level of quality of life calls occurred during Tent City. Noise complaints are included in the category of Quality of Life. It is well known that concert noise is more of a source of noise complaints than the presence of a homeless encampment that, to my knowledge, does not own band instruments or a sound amplification system.

The danger is that voters are often misled by simple one-sentence sound bites. A 30-second sound bite alleging that crime rates increase during encampments could take on a life of its own on social media. Indeed, a letter to the Redmond Reporter was published last week repeating the “conclusion” as if it were a fact.

Care must be taken not to draw or present “conclusions” unless a mathematically rigorous analysis is provided.

Linda Seltzer

Redmond