Thursday, July 25, 2013

Stand Your Ground, and Statistics!

Former stats students, what can we say about the conclusion drawn from this post regarding "stand your ground" prosecutions?
White defenders: 12 of 21 were exonerated, a rate of 57%.

Black defenders: 8 of 11 were exonerated, a rate of 72%

Hispanic defenders: 2 or 2 were exonerated, a rate of 100%

So in the end, no retreat or stand your ground laws actually benefit minorities more than they do whites!

8 comments:

Anonymous said...

Your sample sizes are too small for those statistics to be significant. No conclusions can be drawn.

Darren said...

Yes!

The confidence intervals would be *quite* large.

I haven't done a chi-square analysis to see what it would tell us yet....

PeggyU said...

You might want to look at their raw data again, as several victims with Hispanic surnames are classified as "white". They may have classified the shooters (and stabbers, etc.) similarly, so the data may be open to a somewhat different interpretation.

Also, this information seems to come just from Florida. Are there other "stand your ground" states that could provide additional data?

Anonymous said...

PeggyU: "You might want to look at their raw data again, as several victims with Hispanic surnames are classified as 'white'. They may have classified the shooters (and stabbers, etc.) similarly, so the data may be open to a somewhat different interpretation."

Any "raw" data with Hispanic as an option alongside black and white is likely to have lots of problems.

As an example ... one of my co-workers is a Mexican citizen. He was born in Mexico, grew up in Mexico, speaks Spanish fluently, etc.


He looks "white" ... because his parents were both Jews from Belarus.

He now also has US citizenship.

Our former (also Jewish) CEO referred to him as a "Russian Jew from Mexico" [that is/was a joke, for those who are humor impaired].

So ... is he white or Hispanic?

Beats me.


Black <--> White is a fairly distinct axis with the only real problem that it is a continuum, not binary. Asian works pretty well, too.

Hispanic adds a *geographic* component to this. The upshot is that we have folks like Enrique Pena Nieto (current president of Mexico) who can be reasonably scored as white (look at the Wikipedia article picture) or Hispanic (hey, he's from Mexico!). Pablo Sandoval can also be scored as Hispanic (he's from Venezuela), but if you didn't know that, you'd think he was black.

My inclination is to look at the white/black data for something like this and mostly ignore the Hispanic bits ... but that doesn't fix the fundamental problem that the data is tainted due to bad partitioning.

-Mark Roulo

Steve USMA '85 said...

Mark, you are confusing race with ethnicity. Under current regulations you can be classified by asking two questions:

1) Are you Hispanic? Ans: Yes/No

2) What is your race? Ans: White/Black/American Indian/Asian/Native Hawaiian and other Pacific Islander

In the parlance of Government, you are classified as your race/ethnicity combination. Your Russian Jew from Mexico would be classified as White, Hispanic as is George Zimmerman for instance. President Obama is Black, Non-Hispanic. And so on.

Anonymous said...

I'm not confusing race with ethnicity. The stats that Darren posted do!

If my co-worker shot someone and claimed self defense, would he be a white defender or a Hispanic defender? Or both?

-Mark Roulo

maxutils said...

I read the link ... and it appeared to be a sample of 77. Your numbers 34 ... not sure which is right, but neither would be a good sample size. I think a better measure of your stats is that most people who get charged, regardless of race, are exonerated ...that means that the law is probably serving its purpose.

Steve USMA '85 said...

Mark, both. Your co-worker is white, Hispanic as I said before. I think your REAL question is, "What will the media portray him as?"

To that I can say merely, "Whatever gets the media the best headline and attention to sell more advertising."