@OHToCollege, don’t really think it will end at 300 with 218 SI. The data is skewed. We should expect more lower/average numbers in the remaining students. Still could be 10-12 times larger, so may 150-160 at 218.
@OHToCollege, good point. I’ve always thought they went with 210 and that was way to conservative. Using the midpoint of the concorded ranges for Wyoming would yield around a 204. Close to the 205 you are estimating for commended.
@Speedy2019 your post#2946----how do you get 6% from 60 out of 10,000? That is .006??? What am I missing. Thanks 
Who knows how representative this graph is of typical national scores, but this is the best eyeball count I get of their data at 200 and above:
228 - 0
227 - 2
226 - 5
225 - 6
224 - 8
223 - 3
222 - 8
221 - 7
220 - 13
219 - 7
218 - 14
217 - 10
216 - 13
215 - 12
214 - 13
213 - 9
212 - 12
211 - 7
210 - 13
209 - 6
208 - 11
207 - 11
206 - 16
205 - 7
204 - 17
203 - 14
202 - 16
201 - 16
200 - 15
@hcmom65, yep, I was too fast with the fingers (and brain). @DoyleB has it right.
I was getting myself confused between percentage of SFs nationally and percentage of SFs within the 10,000 sample. What you get from me when I try to quickly compute, analyze and blog all at the same time.
@candjsdad, I tried to compute percentiles based on your data list. @hcmom65, used a spreadsheet this time! Hopefully got it right.
220 = 99.5%
216 = 99%
206 = 98%
200 = 97%
Based on data, 200 could even be commended. Closer to what @dallaspiano was predicting and way lower than others.
My conclusion is that the 10,000 sample won’t be representative of Texas. Testmasters data, I’m sure is factual , but maybe too many high performing students/schools in the sample (and yet not many Perfect/NearPerfects). Others may disagree.
I like my current cutoff predictions. But thinking commended could very well be lower.
Time to jump on the treadmill and watch CNBC. Predicting stock movements has to be easier than SF cutoffs!
@candjsdad and @Speedy2019 Based on the count (0+2+5+6+8+3+8+7+13+9 = 59), the number <= SI 219 is 59.9. What happens to the one SF space (60-59). Gets allotted to another state?. SI 218 had 14 students so cannot make it 59+14 = 73. Suppose Texas had (in this sample calc of 10K students) 70 students, would the entire SI 218 of 14 be included making the cutoff as 218?.
I am not from Texas but was curious of the process. Any thoughts…?
correction.
the number <= SI 219 is 59 (not 59.9)
@Speedy2019 I agree with your conclusion that 10,000 sample probably represents a higher end due to the fact that the surveyed SI scored included those students who prepared using their services. However, it would be “safe” to say that TX cutoff is going to be 217 ~ 219 range. I am more interested in what this bodes for CA. Since their previous prediction for CA cutoff was 219, does this mean their new CA cutoff estimate is likely to be 221 (adding 2 points to the previous cutoff of 219?). This new estimate corresponds to what I predicted early on for CA cutoff: 220, 221 or even 222. IMO, the CA cutoff is almost certain to be either 220, 221 or 222.
I am glad that Testmasters finally shared their data - though the answer is not music to many of our ears. I wonder if a dominant test prep co and if PSAT prep is very common in the Texas schools that have turned out many NMSFs. They do say that the cut off may be as high as 219 - imagine they do not want to create expectations. But they are not yet altering their estimates for other states. But it does call into question the estimates for MD, NY, GA, VA, CA, MA etc. Wonder if Prep Scholar or Applerouth or others will come out and say anything more.
It looks like Testmasters may have the data for HISD (Houston Independent School District). Based on internet research, HISD had 63 NMSF in the recent past (I can’t remember the year). I can’t seem to find the total number of graduating seniors. I did find an SAT saturation number, and I think they listed the total number of students as 9,400, but I don’t know if that is just seniors AND they did not have a year attached to it.
Unless they have a lot of other data points, I don’t think that six high schools would have 10,000 juniors. The high schools here are large, but more than 1,200 per grade would be very unusual, especially for all six high schools.
Testmasters states:
“Using new data from roughly 10,000 Testmasters students and other students who took the PSAT in Texas, we reproduced this process.”
If they surveyed a significant portion of 10,000 students who were Testmasters students, that means the new cutoff estimate is likely to be higher. However, there is a strong argument that the cutoff will not be lower than 218 for TX. Therefore, my best guess for TX cutoff is 218 or 219, not 217.
I guess there are more than 6 high schools, but maybe the NMSF mostly come from the six listed in the post #2972.
Here are some schools that are reported to typically do well: http://www.thesismag.com/2014/10/02/national-merit-semi-finalists-comparing-houston-area-schools/
OK. Got hung up on email - but going to do treadmill soon.
Wanted to post details from an email sent to me. There is a large tutoring company in Austin, called More Than A Teacher. Think of them as our alternative to Testmasters in Houston.
Email to me says More Than A Teacher is predicting the Texas cutoff will be 212-215. Email was sent by PSAT contact person at our high school.
@websensation, I know you and others are tempted to add 2 points to your states testmasters prediction. But may the Texas range should be thought of as 212-219. I’m sticking with 214. To each their own prediction.
Treadmill now, no turning back. Got to lose weight. Sitting at the computer is the enemy.
By stating that their sample has 60 SF, Testmasters is indicating their sample has no bias. It is representative of the state of TX.
Testmasters is taking a good stab at estimates, but I think others on this thread have probably made a more careful analysis. They stated in their update they derived their estimates with linear regression. My limited knowledge tells me that the Concordance data is not linear (blue line) but the Testmasters estimates (purple line) becomes more linear (because they used linear regression) so we see bad fits around 210 and above 220.
What really confuses me is this: Based on the graphs how did they estimate CT down to 215 from 220 (typo?) and how did FL go from 214 to 215 but HI stays at 214.
I like @Speedy2019 estimates best, I would add three points to cover the inflated percentages though.
http://collegeadmissions.testmasters.com/update-psat-scores-cut-national-merit-2016/