How effective are teachers' unions at making life better for teachers?

[xiphias]

Someone on my Facebook linked to a study by "Wallethub.com" on 2014's Best and Worst States For Teachers.

I thought that was interesting, and got curious. So I found a study, How Strong Are U.S. Teachers' Unions? A State-By-State Comparison, from the Thomas J. Fordham Institute, in 2012.

I wanted to see how well unions do at making life better for their members. Here's what I found.

State name	How good for teachers	How powerful teachers union
Alabama	31	20
Alaska	13	15
Arizona	46	51
Arkansas	37	48
California	32	6
Colorado	40	35
Connecticut	25	17
Delaware	35	19
District of Columbia	20	33
Florida	44	50
Georgia	33	45
Hawaii	47	1
Idaho	12	36
Illinois	22	8
Indiana	17	31
Iowa	10	27
Kansas	16	32
Kentucky	42	28
Louisiana	26	42
Maine	43	22
Maryland	19	23
Massachusetts	4	21
Michigan	15	16
Minnesota	3	14
Mississippi	50	46
Missouri	34	38
Montana	23	3
Nebraska	14	26
Nevada	36	25
New Hampshire	27	30
New Jersey	11	7
New Mexico	28	37
New York	7	9
North Carolina	51	40
North Dakota	38	24
Ohio	8	12
Oklahoma	39	43
Oregon	24	2
Pennsylvania	2	4
Rhode Island	21	5
South Carolina	45	49
South Dakota	48	34
Tennessee	41	41
Texas	29	44
Utah	9	39
Vermont	18	11
Virginia	5	47
Washington	30	10
West Virginia	49	13
Wisconsin	6	18
Wyoming	1	29

Here's it graphed out:

https://drive.google.com/file/d/0B2kcYWXBdJDMUkptaUkwcVFNekU/view?usp=sharing

I think I see a SLIGHT tendency that stronger unions are better for teachers, but not a big one. I don't know how to calculate correlation values, or to tell if they're statistically significant, but to my layperson's, uneducated eyes, I don't see that teachers' unions actually help teachers very much.

(Oh -- for what it's worth, I realized that that graph is kind of flipped upside down and backward from what I was originally intending. At the moment, the top right corner is WORST places to work and the WEAKEST unions, and the bottom left is the best and strongest, which was the opposite of how I intended to put it together, but it shouldn't make that much difference, really.)

Comments

[tylik.livejournal.com]

Hm. That the numbers are rankings rather than values might substantially distort any correlations...

[xiphias.livejournal.com]

There are all sorts of things that might distort it -- for instance, who collected the data and for whom? How are they counting unions' power; how are they counting conditions for teachers? That's why I made sure to link to the articles, which explain some of that. But I didn't really read through them.

Still, if the overall patterns are at all related to reality, it's suggestive.

Looking at some of the extremes, you find just weird stuff. Hawaii has the most powerful teachers' union, and almost the worst conditions for teachers. Wyoming, best conditions, union in the bottom half. Virginia, one of the weakest unions, some of the best conditions.

Again -- I don't know who's collecting the data and how. But if it relates to reality, it looks like teachers' unions aren't terribly helpful.

[goljerp]

Well, the thing is that things are complicated. Take New York State. We've got a fairly strong Teacher's Union, and a "democratic" governor, but he's got delusions that he will be able to be president someday, so he's trying to prove his "moderate" street cred by "standing up" to the union -- i.e. making the teachers' life worse. Would their life be even worse without a strong union? My guess is yes, since the union does have some influence over other parts of state government... but it doesn't really matter how strong the union is if the governor has decided that he's gonna screw teachers over.

[boxofdelights]

You can have a lot of fun with statistical calculators here: http://www.socscistatistics.com/tests/pearson/

I typed in your data and got:

The value of R is 0.3546. Although technically a positive correlation, the relationship between your variables is weak (nb. the nearer the value is to zero, the weaker the relationship).

As Tylik points out, this is a silly thing to do to rankings (as opposed to a numerical value that purports to measure how good it is to be a teacher). We know that there is a correlation between height and weight, but if we took a group of fifty people, ranked them according to height, then ranked them according to weight, there might not be much of a correlation between the rankings even though there is a correlation between absolute height and weight.