The Pierre Capital Journal makes available this fun map of South Dakota school districts and their average teacher pay and ACT scores:
Naturally, that map got me wondering: do we find any correlation between the two figures featured here, teacher pay and ACT scores?
I rush to the state Department of Education's Statistical Digest, spreadsheetify the 2013 data, and uncover the following correlations for schools with ten or more kids taking the ACT in 2013:
- Average Teacher Salary to ACT Composite Score: 0.1037
- ATS–ACT English: 0.1839
- ATS–ACT Reading: 0.1151
- ATS–ACT Science: 0.0431
- ATS–ACT Math: 0.0268
- ATS–# students taking ACT: 0.3713
- ATS–high school graduation rate: –0.2526
The ACT numbers suggest a faint association between higher teacher pay and student scores. But the only correlation with statistical significance I feel comfortable citing is the correlations to ACT reading scores, and even that number, 0.18, is so practically small that, even if it hints at some boost kids get in class from our paying their teachers more, that boost could easily be overwhelmed by a grouchy principal or one mild flu outbreak.
The last two correlations are somewhat more significant, suggesting that higher teacher pay aligns with more students taking the ACT but more students struggling to graduate. Those two correlations are still smaller than my "Holy cow!" threshold. But even if you see something worth reading into a –0.25 or +0.37 correlation, note that those two numbers on teacher pay may really just be a back-door way of reminding us that bigger school districts have slightly lower graduation rates (correlation of K-12 enrollment to HS grad rate: –0.1734).
The most confident statement one can make from these data is that within South Dakota, relatively higher teacher salaries are a very weak predictor of student performance on the ACT.
Tangentially related: The Department of Education lists 66 out of 151 districts as having opted out of the general fund levy. Interestingly, opt-outs don't mean higher teacher salaries: the correlation on those two columns is actually negative, –0.1975.
The biggest disconnect I see in South Dakota education:
Schools say "More money please".
Legislature hears "Higher wages".
I think my wife would be satisfied with her salary if she was not constantly being asked to do more than she had to do 3-4 years ago because of staffing and budget cuts. She knew she was not going to get rich, but she does something she is passionate about and is really good at! She now spends less time "teaching" and more time with stuff that has nothing to do with kids. If schools were funded so they could be fully staffed and had people using their strengths most of the time think what could be accomplished!
If large school districts have more problems and therefor have lower graduation rates and and ACT scores, then there are two variables in the comparison. This does not make for accurate research or analysis. Compair the smaller three quarters of the districts for pay to success and compair the larger districts separately. This does not affect the need to close teenytiny districts, but it does put into question whether Sioux Falls should combine elementary schools into giant elementary schools. We are tearing down three schools to build one. A school needs to be big enough to be efficient but not huge enough to have social problems or to be intimidating to the children.
The larger districts tend to pay higher and are not getting the improvement. If we could see compairison on a long term basis of same size schools with higher pay to those with lower pay could tell us more than putting everyone in the same pool.
Remember Corey - correlations go both ways. There is no "ONE-WAY" element to the nature of the relationship between the variables.
It may just as well be that more students taking the ACTs predict that teachers will be paid more, or that fewer students graduating predict that teachers will be paid more.
Want to make a case for a pay increase? Either make sure more of your students take the ACTs, or make sure more of them dropout. From a statistics standpoint - a valid approach.
Roger - these relationships are corollary (they vary together) they are NOT casual (one does not lead to the other - in fact, it is just as likely that some third unmeasured variable (avg. tenure of the teachers in the district for ex:) is the real cause for BOTH of these measures to vary in (almost no - really) synchronicity.
Why not compare test scores from Minnesota & Wyoming to S.D. to determine if higher teacher pay results in better test scores.
Maybe a bonus structure should be set up for those teachers whos students achieve higher test scores?
Bonus pay for higher test scores fails to target the real causes of improvement. Bonus pay also distracts teachers from just doing their jobs. It makes them view students as sources of revenue, not human beings.
I'd love to do the cross-state comparisons. Got a data set?
In Dec. 2010 I ran state-by-state numbers showing a tenuous positive correlation between teacher pay and math test scores: https://madvilletimes.com/2010/12/spend-more-on-kids-and-teachers-get-higher-math-scores-maybe/