Thank you. I hope we are correct.
Any idea what the cutoff is for commended?
I have not done an analysis for the commended cutoff, but I would be willing to undertake it if someone can point me to a national summary of stats that is similar to the state summaries.
Amazed by all the stats but don’t have the Excel skills. If anyone does a prediction for NJ cutoff please share it here…thank you!
As you know, we always get slammed with the high numbers…!
Just hoping that a 223 will make the cutoff this year.
I went ahead and ran the numbers on NJ. For some reason the 2009 test point was well away from the others on my linear regression. I will have to look into it. Throwing that point out, I get the cutoff for the 2014 test being 222.7. So, my best guess is a 223 cutoff number. The results definitely point to a lower number than the past two years. Whether it is enough for a full integer lower is the question. I think it is.
If I were to keep the “bad” 2009 point in, the result would have been an even lower number.
I hope I am right and I hope that helps your piece of mind since your son/daughter needs a lower number than the previous two years.
Thank you! I appreciate your effort. Hopefully, it will be 223 this year.
Seems to me that our kids have higher scores than we did. They should worry about us, not the other way around.
I’d like to have confidence in your methodology, but 220 is not the record high for NY. We hit 221 in both 2006 and 2007.
Consider my confidence shaken…
@96redse5sp I think NY had one year (over last 12 years) of 221, not sure we had two straight 221. There is still hope that 220 is enough for class of 2016.
Yep, 2006 and 2007. My daughter got a 220, so I’m a little nervous.
My son too. I told him to forget about it. Not sure he has. I know I am still thinking about it.
NY scores shake out alot like they did in 2012. My guess is 219. Hope I’m right for both @96redse5sp and @nycuws!
So, @WRUAustin’s TX prediction looks spot on to me!
Due to @Barfly 's comment about the numbers of test takers varying or going up, I decided to perform the calculation based on the raw numbers, but modify the raw numbers based on a states population growth. This assumes that the change in the states population will reflect equally in the %change of graduating seniors. That may or may not be exactly true, but I think it is better than not correcting for the change in population, especially for states like Texas where the growth has been high.
After doing this, I found that the R2 values generally went up and also that the results typically (but not always) didn’t change. Here are the results for the states that I calculated:
Oops, I accidentally hit the save button. I will repost this in it’s entirety.
Due to @Barfly 's excellent comment about the numbers of test takers varying or going up, I decided to perform the calculation based on the raw numbers, but modify the raw numbers based on a states population growth. This assumes that the change in the states population will reflect equally in the %change of graduating seniors. That may or may not be exactly true, but I think it is better than not correcting for the change in population, especially for states like Texas where the growth has been high.
After doing this, I found that the R2 values generally went up and also that the results typically (but not always) didn’t change. Here are the results for the states that I calculated:
****First Method**** ***Revised Method*****
State,2015 prediction, 2015 actual, 2016 Prediction,2015P,2015A,2016P
TX ,217.6 ,218,217.5,217.9,218,218.0
CA ,221.8 ,222,221.9,222.4, 222,222.4
DE ,216.2 ,215,216.0,216.3,215,215.9
NY ,219.4 ,218,219.0,218.7,218,218.6
FL ,211.8 ,211,212.3,211.6,211,212.1
NJ ,222.5 ,224,222.3,223.2,224,223.3
NJ* ,223.1,224,222.7,224.4,224,224.0
A few notes: “A” stands for Actual and “P” stands for Predicted. When I put the prediction of 2015 compared to the 2015 actual, I am not using the 2015 data in the linear regression (because that would be cheating. 
). The linear regression is recalculated with the 2015 data when making the 2016 prediction. Overall, the revised method appears better. The actual predictions round off to be the same number for every state except for NJ. I calculated NJ twice. There is one year where the data really doesn’t fit the line. I don’t know why. I calculated the NJ result with all the data and the NJ* result by throwing out the bad point. I know that one person wants the NJ number at 223. It might happen, but it could easily be 224 also. On average, the 2015 predictions were off by 0.6. They also typically predicted high (in the revised model).
What I am seeing is that across these states there appears to be no big changes from last year. Perhaps a slight bias upwards, but not much.
And the most important thing: caveat emptor
We’ll see how it all turns out. It will also be interesting to see how the commended cutoff turns out.
Good luck.
@WRUAustin Thank you so much for your passionate effort here. I especially love the fact for NY it’s still below 220. @96redse5sp and I TY again!
Thank you so much @WRUAustin
@WRUAustin Wow! Thanks so much for going through all the effort to calculate the NJ predictions! Hopefully 223 will make the cutoff.
Will the score for the commended cutoff make a difference or shed some light on this selection process? If so, how?
@Thinkbig2 isn’t commended a national score? If so, it is an indication of how the group of Juniors did nationally. Your particular state could go up while the nation went down. Also if your state’s population fluctuated, it seems your state cut off will be affected even if the national commended score did not change. Unless you are in one of the really low scoring states, I am not sure you can judge your state’s cutoff by the commended score.