Comparing average educational status of a State withi their position on Gay Marriage

Today North Carolina became the 30th state to officially ban gay marriage.  Putting aside the obviously heated rhetoric for a moment, I wanted to run a quick check and see what correlation (if any) exists between states that have passed gay marriage, those that have banned it, and those who are somewhere in the middle.. at a stalemate. 

I first located a ranking of the educational status of each state.

http://www.huffingtonpost.com/2011/07/11/state-education-rankings-_n_894528.html

From the article: 

The Science and Engineering Readiness Index (SERI) measures how high school students are performing in physics and calculus -- based on publicly available data, including Advanced Placement scores, National Assessment of Educational Progress reports, teacher certification requirements by state and physics class enrollment data.  The SERI was developed by Susan Wite from the Statistical Research Center at the American Institute of Physics and physicist Paul Cottle of Florida State University.
The SERI score given to each state is on a scale of 1 to 5 and reflects how well states perform and allow opportunities for success in physics and math education and teacher qualifications.

This seemed reasonable.

Note the Index also groups each state into one of five categories: 

• Well above national average
• Above average
• Average
• Below average
• Far below average

This will come up later. 


Next I compiled a list of each states' position on gay marriage from Wikipedia:

http://en.wikipedia.org/wiki/Same-sex_marriage_status_in_the_United_States_by_state

While it is clear which of the few states have legalized gay marriage and those that have outright banned it from their territory, there is a large swath of intermediate states where the outcome is in doubt.  Sometimes these intermediate states are because there are more Democrats in key spots (governor, legislature) that hold up reactionary measures, other times because court challenges are in progress (one way or the other). 

Coding for banned/legal is easy.  What I did was coded states that banned marriage as "0", those who have legalized it as "1".  (see table below the chart). 

It is the middle group that presents coding challenges.  Those with indeterminate status I coded as "0.5" for first-round simplicity.  In some cases it was a challenge to code as a 0 or 0.5 and one could probably go either way on a few of these.

Take California.  California passed gay marriage in the legislature, but was overturned by the voters, which was overturned by the court, then was in turn overturned by a higher court, now is possibly advancing to a higher court.  I gave California a 0 just for the hassle with coding it. 

So a state could have a value of 1 = Legalized,  0.5=indeterminate, or 0=banned. 

There are more subtle ways of coding the variants and I may do that later, but for demonstration purposes I started here.

So I have two tables:  one table contains the educational ranks and categories, and the other table contains their position on gay marriage.  I combined these tables for the analysis. 

Recall there are five categories for education, ranging from "Well above average" to "Far below average."  I aggregated these tables to get the average educational ranks and gay marriage scores, then plotted them on a scatter plot:

Each dot represents a group of states (see below table).  On the vertical axis we have the relative acceptance of gay marriage, again with 1.0 being a perfect score.  On the horizontal is the relative educational score in science.

Note that the Well-Above-Average states scored much higher in terms of acceptance of gay marriage, while the three bottom groups had the worst scores.

In fact, this chart suggests that acceptance of gay marriage really doesn't seem to take off until a state is associated with a rating of at least Above Average and higher. 

********** table *********

Education Level	Rank	State	Education Index	Gay Marriage Status
1. Well above average	1	Massachusetts	4.82	1
1. Well above average	2	Minnesota	4.06	0.5
1. Well above average	3	New Jersey	4.04	0.5
1. Well above average	4	New Hampshire	4.01	1
1. Well above average	5	New York	3.94	1
2. Above Average	6	Virginia	3.73	0
2. Above Average	7	Maryland	3.57	0.5
2. Above Average	8	Connecticut	3.28	0.5
2. Above Average	9	Indiana	3.28	0.5
2. Above Average	10	Maine	3.24	0.5
3. Average	11	Florida	3.13	0
3. Average	12	Illinois	3.08	0.5
3. Average	13	South Dakota	3.08	0
3. Average	14	Wisconsin	3.06	0
3. Average	15	Colorado	3.04	0
3. Average	16	Kansas	3	0
3. Average	17	Kentucky	3	0
3. Average	18	Vermont	2.93	1
3. Average	19	Georgia	2.88	0
3. Average	20	Washington	2.86	0.5
3. Average	21	Utah	2.85	0
3. Average	22	Pennsylvania	2.8	0.5
3. Average	23	Tennessee	2.67	0
3. Average	24	Ohio	2.64	0
3. Average	25	Delaware	2.6	0.5
3. Average	26	Michigan	2.6	0
3. Average	27	Oregon	2.58	0
3. Average	28	Wyoming	2.58	0.5
3. Average	29	Montana	2.53	0
4. Below Average	30	Idaho	2.47	0
4. Below Average	31	Texas	2.45	0
4. Below Average	32	North Dakota	2.4	0
4. Below Average	33	Missouri	2.39	0
4. Below Average	34	California	2.38	0
4. Below Average	35	Rhode Island	2.38	0.5
4. Below Average	36	North Carolina	2.34	0
4. Below Average	37	Hawaii	2.29	0.5
4. Below Average	38	Iowa	2.25	0.5
4. Below Average	39	Alaska	2.2	0
4. Below Average	40	South Carolina	2.2	0
4. Below Average	41	Arkansas	2.14	0
5. Far Below Average	42	Oklahoma	2.01	0
5. Far Below Average	43	Nebraska	1.97	0
5. Far Below Average	44	Nevada	1.93	0
5. Far Below Average	45	Arizona	1.91	0
5. Far Below Average	46	New Mexico	1.72	0.5
5. Far Below Average	47	Alabama	1.6	0
5. Far Below Average	48	Louisiana	1.59	0
5. Far Below Average	49	West Virginia	1.58	0.5
5. Far Below Average	50	Mississippi	1.11	0

Now of course the chart isn't totally definitive without a couple more tests.  Just for shits and giggles I put together a one-way ANOVA to analyze the average gay marriage score against the educational groups.

For the stat geeks below are the diagnostics.. essentially what they show is that only the WELL-ABOVE AVERAGE group is likely to have more tolerance regarding gay marriage acceptance. 

Implications.
I will be the first to state that correlations do not imply causation;  sometimes a correlation is just a random occurrence.

Lets think about it sociologically.  I can see how low education might be associated with an unaccepting environments:  more insular, dogmatic, etc. and how broad-based top-tier educational systems would be associated with communities that strive to provide opportunity. 

For proponents of gay marriage, the implications are.. support your schools' academic curricula!  For those opposed to gay marriage.. eliminate the learning!

In any event, this is all just a quickie run through the data here.  Lots of angles to examine around the socio-demographics.. wish I had more time! 

(note, above results updated to reflect a change in Delaware from 0 to 0.5 ... owing to their passing of Civil Unions despite banning gay marriage. Thx Andrew).