In July of 1976, the US Supreme Court set forth the framework that states must follow to comply with the ban on cruel and unusual punishment in the Eighth Amendment with the Gregg v. Georgia decision. Since then, 1532 people were executed in the United States. Death penalty and executions are subject to local laws and not every state carries them out to the same extent (map).
Texas carries out the majority of these executions, having executed more people than the next top 6 states combined. It’s also where the first execution by lethal injection in the world took place. In Texas, capital crimes generally involve murders.
The execution procedure varies from state to state, but in Texas, inmates are given the chance to make a last statement right before execution. These statements are made public in the Death Row Information page of the Texas Department of Criminal Justice (TDCJ).
An interesting dataset; I wanted to see what I could learn from it by looking at the very last words of people sentenced to death (who’ve spent many years in isolation for 20+ hours a day). Though the crimes are heinous, I wanted to acknowledge that these are still human beings, some possibly innocent.
Few statements have both verbal and/or written components, and I’ve included them all in this analysis.
The simplest statistic we can look into here are the individual words that comprise these final statements. What would you say, if it were you? I suppose this also depends on who’s in the room to witness your execution. Let’s see:
Interesting. It looks much more hopeful and positive than I would’ve imagined. love is the most commonly used word, surrounded by family, sorry, forgive and terms like death, pain, kill are uttered much less. Even religious terms are secondary in these statements. hate appears ~20 times less than it’s opposite.
What about the length of statements? Is there a limit to them? Can you try to filibuster your own execution? Would you bother, after spending years and years in isolation on death row?
Considering that the average rate of speech in the US is about 150 words per minute, most last statements take less than a minute. By the same calculation, the longest statement could’ve taken around 8.5 minutes.
One way to identify phrases (without going all-in linguist) is with n-grams. It’s a convenient way to isolate unique phrases in text and speech. In this case, I’ve picked n=3, also called a trigram. This is useful as it will provide a bit more context into the most common words we’ve seen above.
This trigrams visualization was inspired by “Web Trigrams” by Chris Harrison.
I love you, I am sorry, love you all, thank you for all make sense to me. Make sure to play around and see if you can spot other unique phrases. For more on this, check out “Death in Texas” by Jon Millward.
The first thing to note is that TDCJ redacts profanity from the text of the last statements. Still, it’s replaced with some kind of notice of the redaction, with phrases like [Profanity directed toward staff.] and [Portion of statement omitted due to profanity]. Only 3 statements among 454 contain any such notation. There are rarely swear words flying around. This is further evidenced by the overall sentiment of last statements, which tend to lean towards positive.
Google’s Cloud Natural Language provides sentiment analysis for text. What’s interesting here is that it can identify “entities” in text. Here are some select “people” entities, with their average sentiment and the average “salience”, which is a measure of the word’s importance in the text.
It’s easy to get caught up in the exact numbers, but here we see a portrait of which people are referenced in last statements and under what light. Lots of family members are mentioned with goodbyes. warden generally has a positive sentiment whereas people involved in the criminal justice system such as judge, police, investigators and jury are spoken to in a more negative light.
An interesting finding here is that the individual officers have a positive sentiment, whereas police as an institution of people have a far more negative tone.
I imagine the negative sentiment of victim is because it’s often spoken about in a sad, remorseful, indignant or apologetic tone. irene of course is one of the most positively mentioned names… Wait what? Who’s Irene?
This is one of the things I never saw in any of the other related projects I found researching this. Irene? I came across the name Irene when messing around with the data quite often. Let’s see the top 5 mentions by name:
Jesus: pretty obvious, considering the demographics of Texas this isn’t a shocker.
John: one of most common names in the US.
Jack: ok, I’d imagine that it was up there with John.
Father: is sometimes a religious minister, sometimes god or “Heavenly Father”.
But Irene? The 632nd most popular name in the US appears 18 times in the last statements of 454 executed people in Texas? I had to dig. Are these 18 different “Irene"s, or could it be 1 person?
Irene in some statements appears as “Irene Wilcox”. Some searching led me to TDCJ Chaplains' Network. Her husband’s name: Jack. It seems that Irene and Jack were TDCJ Chaplains that offered spiritual counselling.
Sadly, Irene passed away in 2013 but I find it interesting that someone - who has clearly touched so many of these people and left a lasting impression - showing such compassion can come up as an outlier in otherwise mundane data.
Stays of Execution
In the US, all death sentences are stayed automatically until reviewed by an Appeals Court. Executions are stayed (i.e. postponed until a later date for reexamination) quite frequently: one-third of people executed went through more than one death date. Sometimes executions are stayed mere hours before an execution, and in one case the inmate had already been strapped to the gurney with the needle in his arms when the stay of execution came through.
The combination of years of extended isolation and the multiple stays in addition to the stress of the court process is said to cause the Death row syndrome. Suicide rate on death row is 10x the rate for US in total, and 6x the rate of the US prison population.
I wanted to get a visual sense of the timelines involved, as well as the reasons for stays. A clerical error delaying an execution must be hard to handle.
Innocence and Exoneration
In the beginning, I suggested some people might’ve been innocent and still executed. This is more than pure speculation, there is strong evidence that up to 10 people might’ve been wrongfully executed in Texas. Courts rarely go back to review capital cases after execution, and defense lawyers must move onto advocate for other people.
On the other side of the coin, since 1972, 16 people convicted of capital crimes were later exonerated on appeals.
You can read more about the exonerations here.
One Last Ride
I invite you to take a moment to reflect on an inmate’s last journey, given the information above. Capital punishment is a polarizing issue with strong feelings and evidence for both sides. Still, it might be valuable to take 60 seconds and see what goes through your mind.
Thanks for reading, if you’ve made it this far! This dataset is analyzed by many others before me, and a google search will get you a lots of other gems. Here’s one I like, definitely check out “Bye, Warden” by Genevieve Milliken.