What do you think is the cut score for PE Civil Oct 2010? (Raw ?/80)

[benbo]

Any ***** who doesn't think that is possible, must not of faired to well in Engineering Statistics.

Nice. Any ***** who makes all these assumptions obviously knows very little about how NCEES sets the cut score. THere is no "curve" on the entire exam, the process uses a subset of questions. It may have some predictive value for the exam as a whole, but we have no data on how the groups fared on that subset. So you need to make a big assumption here.

We have nowhere near enough data to train and utilize any reasonably accurate predictive method for future exams. Any ***** who believes we do knows even less about data mining techniques.

The best we could do, after making fairly radical assumptions, is get a reasonable range for cut scores on old exams. Which in itself tells us very little about the current exam or any particular future exams (other than a possible range of scores, which as I wrote, is probably little better than a guess).

And your stellar prediction of 52 is completely wrong, based on actual past data.

The insult in CAPS was a nice touch though. It really adds to your analysis.

I already passed the test, so I knew enough statistics for that.

BUT I WOULDN'T WAIST MY TIME DOING SUCH A THING.

Interesting statement. How does one WAIST their time?

 

[Sschell]

ooohhh.... "*****" in all caps! Where can we go from here?

 

[pmblair]

Any ***** who doesn't think that is possible, must not of faired to well in Engineering Statistics.

Nice. Any ***** who makes all these assumptions obviously knows very little about how NCEES sets the cut score.

We have nowhere near enough data to train and utilize any reasonably accurate predictive method for future exams. Any ***** who believes we do knows even less about data mining techniques.

The best we could do, after making fairly radical assumptions, is get a reasonable range for cut scores on old exams. Which in itself tells us very little about the current or future exams.

And your stellar prediction of 52 is completely wrong, based on actual past data.

The insult in CAPS was a nice touch though. It really adds to your analysis.

24 sets of data isnt enough for you? http://www.pels.ca.gov/applicants/exam_statistics.shtml Are you Serious??????? You know ncees has published past percentages too????

I hope you are kidding!!! You think that statisticians dont use and have equations for equating past to future results, with good degree of accuracy?? Stock market, price indexing, i can go on and on

There is only one big assumption and that 56/80 is the base score, and it isnt that radical.

That 52 is a 10 second calculation like i said a quick calculation but the +/- could be 5 and you would be hard pressed to prove me wrong.

But whatever Homie.... I apologize and have a happy turkey day everyone.

 

[benbo]

24 sets of data isnt enough for you? http://www.pels.ca.gov/applicants/exam_statistics.shtml Are you Serious??????? You know ncees has published past percentages too????
But whatever Homie.... I apologize and have a happy turkey day everyone.

Now you've switched to question marks??????????????????????????? That's much better.

As I think I've written repeatedly, it is possible that, after making one big assumption (and yes, that's a big assumption because we are talking about only a few points here), you could get the scores of past exams within a range. I believe you would probably have to accurately know at least two complete data points, including actual cut scores, to accurately train teh statitis. But could probably make some other assumptions to get some sort of approximations.

After making further assumptions, you could then predict the current cut score, or future cut scores, to an even less accurate range. I bet if you applied hypothesis testing you would find it not much better than the guesses people put on here.

If you think you can look at 24 past stock prices and predict the stock price accurately for every day in the future (without any other significant information), well, good luck.

That 52 is a 10 second calculation like i said a quick calculation but the +/- could be 5 and you would be hard pressed to prove me wrong.

Your prediction is that the cutscore is between 47 and 57? I wouldn't even try to prove you wrong. Congratulations on pin-pointing it for us.

 

[Relvinim]

NCEES must be laughing their ***** off when they read these posts.

 

[pmblair]

Benbo,

I agree with everything you just wrote, my original answer was to the question of this post "What do you think is the cut score for PE Civil Oct 2010? (Raw ?/80)"

That if some one wanted to try to calculate the cut-score to some degree of accuracy, it could be done. But it would be futile, due to the fact no one has a clue how many they got right.

The only reason i fired of the ***** comment, was because people jump on someones post and start flaming the person and people that respond. I get the frustration of seeing the same question over and over, but then the same people start flaming those people as if them flaming isn't just as futile of an act.

Brokers and others use those formulas to make purchases, but we know how easy they can FUBAR it.

 

[Sschell]

NCEES must be laughing their ***** off when they read these posts.

they're not the only ones!

 

[benbo]

Benbo,
I agree with everything you just wrote, my original answer was to the question of this post "What do you think is the cut score for PE Civil Oct 2010? (Raw ?/80)"

That if some one wanted to try to calculate the cut-score to some degree of accuracy, it could be done. But it would be futile, due to the fact no one has a clue how many they got right.

The only reason i fired of the ***** comment, was because people jump on someones post and start flaming the person and people that respond. I get the frustration of seeing the same question over and over, but then the same people start flaming those people as if them flaming isn't just as futile of an act.

Brokers and others use those formulas to make purchases, but we know how easy they can FUBAR it.

Okay. If that is what you are trying to say, I'm in general agreement. Your analysis is probably fairly reasonable.

I think our disagreement is in how accurate the process would be as compared to mere guessing (given the amount of data we have). I know this because I am working on a data analysis project right now on transmission line outages using Kalman filtering. We have thousands of data points, and multiple factors, and it still only predicts things to around 5% accuracy. But we may be screwing it up - and probably are.

I don't necessarily get annoyed when people repeat the question. I mainly get annoyed when people (not you) post "facts" that they heard from a buddy,or an instructor, or got off a Ouija Board someplace. It just annoys me for some reason.

And you are right, it isn't worth doing the work.

 

[pmblair]

"Okay. If that is what you are trying to say, I'm in general agreement. Your analysis is probably fairly reasonable.

I think our disagreement is in how accurate the process would be as compared to mere guessing (given the amount of data we have). I know this because I am working on a data analysis project right now on transmission line outages using Kalman filtering. We have thousands of data points, and multiple factors, and it still only predicts things to around 5% accuracy. But we may be screwing it up - and probably are.

I don't necessarily get annoyed when people repeat the question. I mainly get annoyed when people (not you) post "facts" that they heard from a buddy,or an instructor, or got off a Ouija Board someplace. It just annoys me for some reason.

And you are right, it isn't worth doing the work."

I agree with you completely.

5% percent is extremely accurate compared to my daily calculations.

I have seen live load tests that show the code values for live load are off as much as 60%, or wind loads based on calculations compared to wind tunnel testing off as much as 40%. When the contractors give you the old " you engineers over-design everything" its hard not think there on to something.

Happy Thanksgiving. I am out of the office in 15 minutes

 

[Dark Knight]

By the way, a friend of a friend of mine told me that the instructor of a buddy of his, told his buddy that the Quija Board told him the cut score was...

Just messing with you Benbo.

 

[FLBuff PE]

I usually get cut off after 4 drinks. Does that count?

 

[Flyer_PE]

I'm such a lightweight these days that I'd be cut off at 2.

 

[FusionWhite]

I usually get cut off after 4 drinks. Does that count?

4? Sounds a little low to me. I would think there would need to be a sliding scale or formula:

Cut off = (0.66*Beer^time)-tequila^2

 

[bigray76]

I usually get cut off after 4 drinks. Does that count?

4? Sounds a little low to me. I would think there would need to be a sliding scale or formula:

Cut off = (0.66*Beer^time)-tequila^2

That makes more sense... really, if you had 4 beers in an hour they might cut you off, but if you are a 'dairy farmer' and milk thoses beers for 4 hours, no one is going to cut you off (and chances are that you are still fairly sober)...

 

[snickerd3]

I usually get cut off after 4 drinks. Does that count?

4? Sounds a little low to me. I would think there would need to be a sliding scale or formula:

Cut off = (0.66*Beer^time)-tequila^2

i'd be cut off before I started then as the beer part would be zero so a minus tequilla number

 

[bigray76]

I usually get cut off after 4 drinks. Does that count?

4? Sounds a little low to me. I would think there would need to be a sliding scale or formula:

Cut off = (0.66*Beer^time)-tequila^2

i'd be cut off before I started then as the beer part would be zero so a minus tequilla number

You can drink all the tequila you want, I won't cut you off!

 

[snickerd3]

I usually get cut off after 4 drinks. Does that count?

4? Sounds a little low to me. I would think there would need to be a sliding scale or formula:

Cut off = (0.66*Beer^time)-tequila^2

i'd be cut off before I started then as the beer part would be zero so a minus tequilla number

You can drink all the tequila you want, I won't cut you off!

How kind of you. Will you provide a pillow and blanket when I pass out?

 

[bigray76]

I usually get cut off after 4 drinks. Does that count?

4? Sounds a little low to me. I would think there would need to be a sliding scale or formula:

Cut off = (0.66*Beer^time)-tequila^2

i'd be cut off before I started then as the beer part would be zero so a minus tequilla number

You can drink all the tequila you want, I won't cut you off!

How kind of you. Will you provide a pillow and blanket when I pass out?

Of course! Anything else?

 

[roadwreck]

I usually get cut off after 4 drinks. Does that count?

4? Sounds a little low to me. I would think there would need to be a sliding scale or formula:

Cut off = (0.66*Beer^time)-tequila^2

i'd be cut off before I started then as the beer part would be zero so a minus tequilla number

You can drink all the tequila you want, I won't cut you off!

How kind of you. Will you provide a pillow and blanket when I pass out?

Of course! Anything else?

probably going to want to provide a bucket too.

 

[snickerd3]

I usually get cut off after 4 drinks. Does that count?

4? Sounds a little low to me. I would think there would need to be a sliding scale or formula:

Cut off = (0.66*Beer^time)-tequila^2

i'd be cut off before I started then as the beer part would be zero so a minus tequilla number

You can drink all the tequila you want, I won't cut you off!

How kind of you. Will you provide a pillow and blanket when I pass out?

Of course! Anything else?

probably going to want to provide a bucket too.

no bucket required...I fall asleep long before I get to that point. One of the main reasons I was not a big bar goer in college.