News & Updates

DISPROPORTIONATE IMPACTS OF THE PERSISTENT OFFENDER LAW: DISPARITY BY RACE, COUNTY, AND AGE FOR INDIVIDUALS ELIGIBLE FOR RE-SENTENCING UNDER SB 5164
September 10, 2021 | Alison Pagan

Robin Hardwick | Washington Defender Association
June 3rd, 2021

INTRODUCTION
This report is in response to Washington Senate Bill 5164, which calls for the re-sentencing of incarcerated individuals sentenced under the “Three Strikes” portion of the Persistent Offender Law with Robbery in the second degree as one of their strikes. Using data from the Washington State Department of Corrections, I take a closer look at individuals convicted of Robbery 2 by their race, ethnicity, and age while controlling for population.
In doing so, we find that we must prioritize Black and American Indian/Alaska Native community members in our re-sentencing efforts, especially in Clark, Okanogan, Pierce, Cowlitz, and Thurston Counties. Additionally, the proportion of younger individuals convicted are overwhelmingly people of color (especially Black), indicating that we must focus on re-sentencing Black individuals who were under the age of 30 when sentenced.

COUNTY LEVEL
The following section breaks down the total number of individuals who have been convicted of Robbery 2 in Washington State by the sentencing county. While these individuals may have more than one offense not limited to Robbery 2, the following data weighs every individual the same.

Unweighted by Population

The counts of individuals sentenced with a Robbery 2 conviction on the bar chart above roughly reflects the population size of each county. Population demographics indicate that King County is most populous by far, with over two million residents, followed by Pierce County with about 877,000 residents. Interestingly, Spokane outranks Snohomish in the above plot even while Snohomish is more populated, indicating that Spokane sentences more folks with Robbery 2 convictions than another, more populated county. We also observe that all of the other counties have sentenced less than five individuals. With such a small sample size per county, we are reluctant to assign significance to these counties even if the proportion of individuals convicted is high.

Counties who have not sentenced anyone under the Persistent Offender Law with a Robbery 2 strike are excluded from the above visualization.

Weighted by Population

The flaw in last section’s visualization is that it does not reflect each county’s different population size. For instance, King County looks like the biggest culprit of convicting folks with a Robbery 2 charge, however, this is simply because King County has more people and thus more individuals being sentenced for all charges. To control for population, I divide the total number of individuals the county has sentenced with a Robbery 2 conviction by the county’s population. With small populations of 7,578 and 22,421 respectively, Ferry and Asotin sentencing one individual with a Rob 2 conviction impacts a larger percent of their population than King County or Thurston County, both of which placed high on the last chart.

Perhaps the most significant disparity we see by county is how high Spokane scores, with the 17 individuals they have sentenced impacting a much larger portion of their residents than other counties. The same applies here to Cowlitz and Pierce Counties, who sentence a large number of individuals with a Robbery 2 conviction compared to proportion of individuals impacted in other counties.

RACE

The following section explores the total number of individuals sentenced with a Robbery 2 conviction by the individual’s recorded race.

Please note that because Hispanic is recorded as an ethnicity, not a race, it is not included as a category here with Asian/Pacific Islander, American Indian/Alaska Native, Black, and white. Refer further down the page to the ETHNICITY AND COUNTY section to explore the impacts of a Hispanic identity on Robbery 2 convictions.

Unweighted by Population

The above graph paints a stark picture of who is targeted by Robbery 2 convictions and the court system as a whole in Washington State. While reading the above graph, take care to note that the counts are not weighedWhite individuals make up 69.3% of the population while Black individuals make up 3.7% of the state’s population—yet we observe the exact same number of individuals in both of these racial groups sentenced with Robbery 2 convictions. This means that the Black community has been unfairly targeted with Robbery 2 charges in Washington State, representing 47.57% of all people in the state with a Robbery 2 conviction, despite only making up 3.7% of our state’s population. This pattern will become even more apparent in the following graph as we control for population.

Weighted by Population

This graph controls for the population of each racial category in Washington State, using data from the Office of Financial Management’s racial estimates for April 1st, 2021. Unsurprisingly, the graph shows us that the racial group most impacted by Robbery 2 charges are Black individuals, followed by American Indian/Alaska Native individuals, then white individuals.

With this in mind, the data provides clear support for the legislature’s passing SB 5164, as it is evident that Robbery 2 charges unfairly target minority races. Additionally, these results also indicate that in the effort to re-sentence folks with Robbery 2 convictions, courts must prioritize the re-sentencing of Black and American Indian/Alaska Native community members, as they have been impacted the most harshly and unfairly by the Persistent Offender Law’s three strikes rule.

Chi Square Test of Significance

Race Observed Number of Individuals Sentenced with Robbery 2 Conviction Expected Number of Individuals Sentenced with Robbery 2 Conviction
American Indian/Alaska Native 4 1.991389
Asian/Pacific Islander 1 11.310402
Black 49 4.687573
White 49 85.010636

To test the results of the above bar charts, we must run a chi-square test of independence. While the relationships are stark, there is a smaller number of cases, and thus it is important to test the statistical significance of race on for individuals eligible for re-sentencing. To do this, I took the observed number of individuals convicted of Robbery 2 by their race and compared it to the expected number of folks of each race that would be convicted out of the 103 individuals statewide charged with Robbery 2, if the numbers accurately reflected the race’s measured proportion of the population (calculated from the OFM estimates). The above table shows these values.

Secondly, we run a chi-square test of independence. It is important to note here that our data meets the six assumptions of a chi-square test:

  1. The data is counts, not transformed into percentages or another form.
  2. A subject only is assigned to one category of race.
  3. Each subject is counted in just one cell.
  4. Data are not paired, and thus is independent.
  5. The variable being measured is categorical.
  6. While two of the four expected cells are greater than five, a third cell is also almost greater than five, with a value of 4.68. Additionally, none of the cells are under one. While it would be more ideal to have more data for this chi-square test, the data is a comprehensive recording of all individuals sentenced with a Robbery 2 conviction. Thus, we cannot obtain more of this data.

Our results are highly statistically significant: the chances of race having no impact on who is convicted of a Robbery 2 charge are next to zero (, to be precise). Additionally, the reported Cramer’s V value indicates that we can attribute 51.93% of the variation in convictions to race, a very high value for a Cramer’s V. I would like to emphasize that statisticians can very rarely create models that explain half of the variation with just one variable. For instance, I have created models with over twenty variables only to explain just 39.6% of variation. The fact that we can so easily build models that predicts something as impactful and long-lasting as whether or not someone is sentenced to a crime in Washington State with just race is sickening. These findings point to the need for swift and radical change in the Washington State court system.

RACE AND COUNTY

Unweighted by Population

The above graph shows the unweighted number of individuals charged with Robbery 2 in each county, colored by race. Immediately, even without being weighed each race’s population size, it is clear that Black people are targeted more than white people, especially in King and Pierce County. Both of these counties have convicted over twice the amount of Black folks as they have white folks, despite the fact that Black folks make up just 6.76% and 7.22% of their populations, respectively.

We should also find concern with Spokane County, which convicts seven Black individuals to ten white individuals. Additionally, while many counties have convicted under three individuals of Robbery 2, it is crucial to consider what it means for counties such as Okanogan and Ferry to not convict any white members, but have instead convicted a member of one of their most vulnerable populations (which in Okanogan only makes up 0.44% of their population, or 189 individuals to the county’s total population of 43,310).

Weighted by Population

The most outstanding part of this graph is the small quantity of pink (convictions of white people) across all counties. Also, while controlling for population we can see concerning relationships that the last graph masked. For instance, we are concerned with Cowlitz County, as they have convicted the largest percent of their Black community than all other counties. We see the same pattern with Spokane County, as they convict a larger portion of their Black community than their white residents. This graph also makes evident how targeted American Indian/Alaska Native individuals are, because they make up such a small portion of the population, yet face Robbery 2 convictions in counties where the proportion of white individuals impacted is next to none. American Indian/Alaska Natives are heavily targeted, especially in Clark, Okanogan, Pierce, and Thurston Counties.

It is also significant to note that Ferry County is omitted from the above graph, as they have convicted over 2% of their Black population. Including their county in the graph skews the x-axis, rendering all of the other bars too small to take any conclusions from.

ETHNICITY AND COUNTY

This section explores the relationships between Hispanic and non-Hispanic individuals on the county level for Robbery 2 convictions.

Unweighted by Population

The above graph proves that while not many Hispanic individuals are targeted through Robbery 2 convictions, certain counties are much more likely to target their residents than others. This is especially true in Whatcom County—which has only convicted the Hispanic community of Robbery 2. However, Whatcom has only sentenced one person with a Robbery 2 conviction, meaning there is not enough data to make many conclusions from the results. But it is worth asking how Whatcom has only ended up convicting someone from a marginalized community, a community that in total makes up less than 0.1% of the county’s population.

Unweighted by Population

This graph reaffirms the previous one, demonstrating that Spokane and Whatcom both have convicted a large percent of their Hispanic population. Significantly, we notice that when controlling for population size of the ethnicity groups, King County does not seem to target Hispanic individuals more, nor do the other counties (excluding Spokane and Whatcom). In short, we do not see the same concerning patterns that race shows us for conviction records of Robbery 2 charges, though a couple of counties demonstrate problematic behavior toward their Hispanic population.

AGE

The following section takes a closer look at the age of individuals when they were convicted of Robbery 2 in Washington State.

The above chart is a histogram by an individual’s age when their Robbery 2 conviction was filed. How the data is right-skewed is noteworthy, meaning the majority of folks convicted of a Robbery 2 charge and currently incarcerated are young—under about 30 years old. Credibility is lent to this by the average age, represented by the yellow line and calculated as 28.64 years old. It is also notable that 20-year-olds are the age with the most Robbery 2 convictions filed, with over 17 individuals having been convicted at this age.

AGE AND RACE

This following section explores how race interacts with an individual’s age when they are convicted of a Robbery 2 charge in Washington State.

 Unweighted by Population

While the above chart is unweighted, we can start to glean some problematic trends between age and racial category. For instance, people of color are convicted when 25 years old and younger at a much higher rate than their white counterparts (including Black, Asian/Pacific Islander, and American Indian/Alaska Native individuals). This indicates that the youths of marginalized racial communities are targeted unfairly by the court system for Robbery 2 charges. The following graph controls for population and further supports this conclusion.

Weighted by Population

Upon controlling for population, the previous graph’s trend becomes even more stark. The youngest members of marginalized racial communities are targeted through Robbery 2 convictions, as nearly 100% of individuals under the age of 28 convicted of Robbery 2 are members of a marginalized racial category (after controlling for population size). The Black community especially is hurt by Robbery 2 charges, with their community making up just about all of these convictions for 28-year-olds and under (as well as a majority of the older individuals convicted), with some higher proportions attributed to the American Indian/Alaska Native community.

When re-sentencing individuals using SB 5164 in Washington state, I argue that it is crucial for prosecutors, the defense community, and judges alike to prioritize re-sentencing young Black incarcerated individuals, as they have been grossly misrepresented in Robbery 2 convictions. This means that they have been denied an equal chance at life, one that is clearly granted to most white individuals without an afterthought. To live up to the spirit of SB 5164 and undoing harm, we must advance re-sentencing for all individuals impacted by Robbery 2 charges, with a focus on marginalized communities and their younger members as they have been sentenced due to unfair and cruel discrimination in the court system.

ANOVA Test of Significance

The boxplot above plots the distribution of ages for individuals convicted of a Robbery 2 charge by race. As we can see, marginalized racial communities are convicted more when they are younger, while white individuals tend to be convicted at an older age (above 30). To test the significance of this relationship, I ran an ANOVA test—which compares the variation both within and between the racial categories to determine if there is a significant relationship between race and age of conviction. With an F-value of 1.983 and significant at the p < 0.15 level, I find that there is a 12.9% chance that the relationship between age and race is unrelated. While these results are not as significant as those in the chi square test for race and number of convictions, we can conclude there to be a relationship between age and race, though it is not as strong as previous results in this report.

Additionally, we meet the majority of the assumptions for an ANOVA test:

  1. The sample distribution is approximately normal (symmetrical and unimodal).
  2. Observations consist of the entire population (a census) of Robbery 2 convicted individuals, and thus meets the requirement for random sampling.
  3. Group sample sizes must be equal to or greater than ten, which is true for the Black and white racial categories. However, there are not enough American Indian/Alaska Native or Asian/Pacific Islander individuals convicted of Robbery 2 to meet this assumption, meaning we should treat our results with a bit of caution.

CONCLUSION

In short, this report indicates which communities and counties must be prioritized in the wave of re-sentencing for SB 5164 in the coming months. For instance, after controlling for population, we determined that Spokane, Cowlitz, and Pierce Counties have sentenced the largest proportion of their population with Robbery 2 as a strike. Further worrying is the clear targeting of the Black community with these convictions, while American Indian/Alaska Native individuals are also disproportionality impacted. Combining these two variables indicates that Clark, Okanogan, Pierce, and Thurston Counties have been targeting Black and American Indian/Alaska Native people, with higher proportions of these communities impacted than white or Asian folks. Ethnicity does not prove as significant as race, though Whatcom and Spokane Counties individually are proven to target Hispanic individuals disproportionally. Finally, we found that communities of color (especially Black individuals) disproportionally find themselves sentenced at a younger age. Thus, we must also prioritize the re-sentencing of people who were incarcerated while under the age of thirty, as these individuals were likely treated harsher by the court system due to their race and youthfulness.

There is much work to be done in order for SB 5164 to truly undo the harm that Robbery 2 convictions in the Persistent Offender Law has caused. This report offers several suggestions about communities to prioritize and places to start for counties and prosecutors who are unsure where to start when beginning re-sentencing.

BIBLIOGRAPHY

The author is thankful to both the Washington State Department of Corrections (specifically Courtney Bagdon-Cox in RDA Data Analytics for compiling the report and Trisha Newport for requesting it) and the Washington State Office of Financial Management for publishing their 2020 estimates of population by race and Hispanic origin and county.

Please reach out to the author at robhardwi@reed.edu with any questions you may have.

This study was supported in part by the Paul K. Richter & Evalyn Elizabeth Cook Richter Memorial Fund.