Visualization of estimated ideologies of Supreme Court justices
Alex Lundy writes:
I’m an avid reader of the Monkey Cage, and as such, I wanted to point you towards a data-visualization I’ve just completed based upon Martin-Quinn scores, the numeric ideological estimates for Supreme Court justices developed by political scientists Andrew Martin and Kevin Quinn.
This is beautiful. My main comment is that I don’t know whether to believe the numbers. Is the Anthony Kennedy of 2007 (ideology score 0.14) really so close to Hugo Black in 1970 (ideology score 0.06)? To look at it another way, according to these numbers, in 1973 (the year of Roe v. Wade), six of the justices are colored red and the median justice is listed at 0.67. In 2007, only five are red and the median is at 0.14. In fact, in 2005 the median is listed as -0.07, or slightly to the left of center. Is it really plausible that the court was more liberal in 2005 than in 1973? Maybe so, but something looks fishy to me here.
Spaeth’s numbers (for each justice, the proportion of cases for which he or she voted on the conservative side; see Aleks’s graph here) look a bit more reasonable to me, although I know these numbers have issues too. See here for further discussion.
Beyond this, I have a few comments on the display:
1. I’d prefer some sort of continuous color labeling rather than the red-blue labeling that Lundry is using. When you’re mapping votes, it makes sense to use votes that are more than 50% in one color and less than 50% in another. But for these judicial ideology scores, the 50% point is pretty arbitrary, and I think it’s misleading to jump from red to blue.
2. I think it’s misleading to line up the justices in positions #2, 3, 4, etc, thus implying some sort of continuity between retiring judges and their replacements. When a new judge comes in, he or she is really coming in anew, and I think it would make sense to add a new line for him or her rather than creating an illusion of continuity. In particular, the values on the far right column of the table (with the exception of the chief Justice) are basically meaningless.
3. Supreme Court years are of the form 1970-71, 1971-72, etc. In the current labeling, it is unclear whether “1972” refers to 1971-72 or 1972-73. It would be easy to fix this by shifting the year labels to be between the blocks.
4. Justice Hughes’s middle name is misspelled. More generally, I can’t figure out what his rule is for when to include middle names, when to include middle initials, and when to include neither.
Comments
Andrew - thanks for posting the link and for the critique. A few notes:
1) Regarding the numbers themselves, I would point you to Martin & Quinn’s website: http://mqscores.wustl.edu/ where they deal with methodological issues and have links to their published works on these numbers.
2) Regarding color - this is a great recommendation and perhaps we’ll work that option into our first update. We are already working on anchoring the color white specifically to zero, as M-Q say 0 is the ideal ideological midpoint.
3) Regarding lining up the justices - that, to me, is one of the more interesting components of this dataviz. The genesis of this idea came from: “Souter is retiring - who will replace them and how different will they be ideologically…wait, who did Souter replace and how did his introduction to the court change the ideological dynamic?” I had never thought of SCOTUS seats as being specific posts that can be traced backwards through specific justices. Is it the best way to view the court? Probably not, but it’s a new way of looking at it that hopefully is informative.
4) To the best of my knowledge Supreme Court terms are labeled by the year in which they began. Thus the current term, which began in Oct of 2008 is the 2008 term.
5) The Hughes typos has been replaced (thanks for the heads up!). The use of middle names was fairly arbitrary in that I used them when that individual seemed to be commonly identified with a middle name (and I have to admit that Wikipedia was the primary resource on that!).
Again, thanks for the feedback. I look forward to any other commenters’ feedback and incorporating them into the visual in future iterations.
Posted by: Alex Lundry | June 3, 2009 09:48 AM
Michael Bailey has a set of common ideology scores (for the Court, Congress, and the president) that appear more facially valid during the 1970s, ala Roe. See http://www9.georgetown.edu/faculty/baileyma/Data.htm. To be fair to Martin and Quinn, I’ve yet to run an analysis where choosing different scores led to substantively divergent results (though I haven’t run that many).
I’m fairly certain that Democrats had lost control of Congress by 1996. I’m also not sure how the colors are “averaged” in the rows; why is the median justice in 1963 (fairly liberal at -.8) filled with white, while the median justice in 1961 (barely conservative at .15) gets a slight red fill? I note that this problem doesn’t seem to occur on the excel file.
I don’t have any problem running the seats concurrently; after all, much of the nomination debate depends on how different, ideologically, the new justice will be from the old one. Sotomayor replaces Souter, yawn. Diane Wood replaces Scalia…watch out!
All in all, a good visualization, particularly for showing “drift” in the ideology of individual justices. I’ll be using this for that purpose in my undergraduate courses.
Posted by: Rob Robinson | June 3, 2009 09:58 AM
Rob - thanks for the comments. I’ll take a look at those Bailey numbers and see how we might be able to work them…it would be pretty cool to see the same visualization with different ideological estimates.
Regarding the colors - this is the hardest to explain. The colors are chosen based upon the Min, Max, and Median of the area we are comparing. So, in the first view, the “overall” view, the darkest Red is anchored to the maximum ideology number across all justices and all terms, the darkest Blue is anchored to the minimum score, and the purest white is anchored to the actual median number (The Location of the Median Justice is NOT necessarily the actual median, as it is calculated via a Bayesian estimate).
The second “compare” option, “within each seat, row” calculates separate color anchors for each row.
Similarly, the third compare option, “within each year, column” calculates separate color anchors for each column.
The Location of Median Justice and Court Average are not included in these calculations and their color values are set to what they would be in the overall comparison.
And yes! Thank you for pointing out the mistake with control of congress - it will be fixed later today. And I expect we’ll find other typos as we go along - thanks for everyone’s patience on this!
Posted by: Alex Lundry | June 3, 2009 10:19 AM
Ruth Bader Ginsburg’s name is misspelled, too.
Posted by: Kieran | June 3, 2009 01:05 PM
Alex—thanks for the response.
Posted by: Rob Robinson | June 3, 2009 01:27 PM
Ginsburg not Ginsberg. That one always gets me. Thanks Kieran.
Also, a little more on Andrew’s initial concern about the numbers, especially regarding Roe v Wade:
First, we should look at the 1972 term, as this is the term during which Roe v. Wade was decided (the ruling itself came out in 1973, but the terms are labeled by the year they begin in).
Second, it’s important to know that the location of the median justice actually denotes the ideological score of the justice most likely to be the deciding vote. So, comparing the 1972 term to the 2007 term, the justice most likely to be decisive is more liberal in 2007 (0.14) than 1972 (1.03), but you’ll note the overall average of each justice’s ideology scores is actually more conservative in 2007 (0.43) than it is in 1972 (0.14).
Finally, it makes some sense that location of the median justice in 1972 was conservative: consider that Justice Blackmun, a Nixon appointee and generally conservative jurist, was the writer of the majority opinion on Roe v. Wade.
Posted by: Alex Lundry | June 3, 2009 03:38 PM
Simple is the beauty, thanks!
Posted by: John Smith | June 5, 2009 06:10 AM