An exploratory analysis of Nebraska Correctional System's inmate database
In this analysis, we will be answering questions specifically relating to race and life without parole (LWOP), where the minimum sentence is life (as opposed to a set number of months or years). While LWOP has more or less always existed under varying levels of legality, we have separated LWOP as a minimum life sentence that begin after 2002 (due to a change in legislation).
Note that we are not currently looking at inmates whose minimum sentence is a definite number of years, even though that number may be large (for example, 80 years).
Cleaning and filtering the data
The database exists as two separate sheets, which can be joined by an ID number (we are not sure as of now if the IDs are an accurate way to join the datasets, but our current analysis does not rely on the merged dataset).
Here we just drop the columns we don't need and adjust datatypes for the remaining columns.
Note that we have deleted duplicates by ID from the second sheet, so that we only have unique
IDs in the crime description. We're also filtering the dataset for only minimum sentences that
Next, we have merged both datasets by ID in
merged. The second cleaned dataset we will be
merged_lwop, which is minimum life sentences that begin after 2002.
Here's a look at the first five rows of the
Analyzing the data
The first analysis we did was by county: We sorted the values by county and counted for each race. The results of this can be seen below.
It's evident that Douglas County, which has a population of approx. 50/50 white and Black, has glaring disparaties by race (60 Black inmates vs. only 14 white inmates). Another interesting data point is in Lancaster county, which has a small population of Black people (about 4%), has 6 Black inmates and 6 white inmates. These are two counties that would be interesting to focus on in future analyses.
The race description only allowed for one value, and Native American and Hispanic inmates do not appear to have been counted reliably throughout the years. We are not focusing on their values during this analysis.
Analysis of race
Here is an analysis of minimum sentence life for all years (in
merged) broken down by race.
The first graph displays these values—given that there are only about 4% Black people in
Nebraska currently, it is clear that there is a significant difference in the proportion of white
and Black people who are given minimum life sentences.
Below, we have compared the number of inmates by race for both pre- and post-2002.
There is an interesting difference in the two graphs above. There are significantly more Black people sentenced to LWOP after 2002 than before 2002. This difference may warrant more investigation, perhaps due to a change in prosecutorial leadership or due to the new legislation in 2002 surrounding LWOP.
Finally, we looked at race sorted by gender for LWOP after 2002. These results are shown above.
Sorting by offense type
We took a look at how the dataset post-2002 looked for each type of offense given LWOP. There are many more 1st degree murders given LWOP than 2nd degree.
When these values are further broken down by race (as shown above), we find that many more Black people are sentenced to LWOP for 1st degree murder than white people are for the same crime.
While this dataset does not have enough information to analyze as to why this difference occurs, it raises some significant questions in relation to examples of past cases—could Black people have been convicted of a worse crime for less significant involvement than white people, which would appear as racism in criminal investigation? This would turn into an analysis in criminal heinousness. Were more white people convicted of 1st degree murder but not sentenced to LWOP and instead given a lesser sentence (such as life with a chance at parole)? This may show up in an analysis of the full dataset, which we investigate in
Analysis by age
The graph below depicts LWOP broken down by race as well as age. There is an incredibly evident difference in the age a Black person is sentenced to LWOP (the blue bars) versus a white person (the red bars). Black people tend to be sentenced to LWOP much younger than white people.
More analysis into this topic over the full dataset may reveal more results—how long do inmates actually stay in prison on average (or total prison years), broken down by race?
Further Analysis by Offense Type
Here, we will analyze more of the original dataset (including all minimum term years, instead of just LFE). This analysis is only for data post-2002. We've just imported the data once again and instead filtered by maximum LFE term.
Here are the total counts of the new dataset (with all types of minimum sentences) by offense arrest description.
Below is the same data, but split by race. There's evidently some suspicious discrepancy when it comes to 1st degree and 2nd degree murder and the split between Black and white races. (63 vs. 43 for 1st degree, and 18 vs. 34 for second degree, respectively).
The above table describes the split between race for 1st degree murder convictions that ended in a minimum life term. There is a clear difference in treatment for white and Black people convicted of 1st degree murder—white people are a) far less likely to be convicted of 1st degree murder in the first place, and b) if they are, they are more likely to be given a lesser sentence (there are white inmates given as low as a 76-year minimum sentence for 1st degree murder, whereas inmates of all other races are given a minimum LFE sentence).
Offense description analysis by county
This is a particularly interesting table. The counties that white sentences come from are very varied, which makes sense given how small the black population is in a lot of these counties. The main counties are Douglas (14), Lancaster (6), and Sarpy (3) - the 3 most populous ones. Also, the only non-LFE sentences come from Sarpy and other small counties. The counties that black sentences come from are very few, but mainly Douglas (56) and Lancaster (3). Douglas county is made up of 11.4% African Americans, and Lancaster is just 4.2%. Notably, in Douglas, the ratio of black:white is 1:10 (approx), but the ratio of 1st degree sentences is 4:1.
The data is more surprising for 2nd degree murder (table shown below). First, with whites, the cases are more varied as usual, with Douglas (7), Lancaster (7), and Lincoln (5) leading the pack. With Black people, it is mainly just 2 counties - Douglas (10) and Lancaster (6) again. When it comes to 2nd degree murder, Douglas county still has pretty skewed numbers given the population, and while 2 black people were given LFE there, no whites were. However, the years given are actually pretty similar. We see both really harsh punishments for African Americans in Douglas county, and a huge preference for 1st degree over 2nd degree murder. However, while the numbers are definitely much lower for whites in Douglas county, even for them the 1st degree is more, which could just mean Douglas county is really harsh?
On the other hand, Lancaster county is a little more interesting. Whites were given some of the lowest years here, even the bare 20. However, for African Americans, 3 LFE sentences were given, and the minimum otherwise is 60 - this is the biggest gap seen so far. The black:white population of Lancaster country is just 1:25, yet 1st degree sentences is 1:2, 2nd degree sentences is literally 1:1, and even within 2nd degree murder, the actual number of years given out have a huge difference. Also, while LWOP is more in Lancaster for Black people, 2nd degree is more common for whites. Lancaster county is a model county that displays all the trends we have discovered so far.
It would appear that any biases that led to many of the biases shown in the data are very prominent in these two counties, which results in very skewed numbers. An important fact to note is that after Douglas County, which has the highest Black population, there are several counties including Dawson and Johnson which don't even show up in any of our findings. IN fact, Lancaster is only 7th in terms of Black population.
These are ideas and points of further investigation.
- Look into empirical data (a few individual cases) where white people were charged with a lesser crime/sentence than Black people for a similar crime
- Why are pre-2002 and post-2002 numbers so different by race? could this just be because of the smaller dataset for each, or maybe the fact that pre-2002 it is more recent?
- Timeseries line graph of LWOP by race
- Actual total (or average) years in prison by race (as opposed to sentences)
- Differences in "crime heinousness" (would require more information)
- Offense descriptions of the full database broken down by race
- Did spikes in LWOP sentences correlate to election years or years of major racial unrest?
- Determine population of African Americans over time