HPAT

mikl

ya id have to agree with almost all of what statsman has said.. iv been running the figures over a few times also, heres what i think:
considering as statsman put it the top 15 percent will get in(rough estimate) that equates to a score of just above 170 on the hpat..
Also going by previous leaving certs the so top 15 percent have been achieving 570 points or above and gaining entry into medecine..this amounts to 554 on the new scale..
Add the 2 together and you have a combined score of 724..lets say 725..

so there you have it, a combined score of 725-730 will be the entry score for this year in the cao...id put the house on it!!

A Neurotic

I preferred JammyC's logic, he gave me a place on 530 points, Statsman tells me I need 550...:(

mikl

I apologise for any distress caused.;) it was my cousin who did the calculations though..3rd year actuary in UCD!

sarahbrennan

i notice some people have written up scores from each section as well as an overalll score, can anyone tell me where ye got this as i only got an overall score??

MrPain

sarahbrennan wrote: »

i notice some people have written up scores from each section as well as an overalll score, can anyone tell me where ye got this as i only got an overall score??

Log into Cao website, its at the bottom of 'my application'

A Neurotic

If you log into cao.ie, and view your application you can see your HPAT results at the bottom of the page, broken down into 3.

sarahbrennan

thanks

Where are people going saying each percentile rank contains the same number of people? Does it say somewhere on the HPAT website that this is the case? I doubt it is; would the results not fit a normal distribution curve?

Take IQ for an example: it follows a normal distribution. It certainly doesn't contain the same number of people in each percentile rank. My guess is that the HPAT would be no different.

It's impossible to work out exactly how many people are in each percentile range without knowing what standard deviation they've used; as far as I know they haven't revealed any of their methods.

If it does follow a normal distribution (or something similar) then the vast majority of people will have scored around the 50% percentile (it would be something like 68% of people within one standard deviation i.e. from 35% - 65%, or something similar - can't work it out without knowing their standard deviation). Again, something like 95% of people will fall within two standard deviations of the mean, which would be something like 20% - 80%. My guess would be that at an absolute max, only 7-8% of people will have scored above ~80%. People say that 2900 people took the exam, so that means (conservatively), that 232 people will have scored above 80%; which isn't an awful lot, really. (There's a bias on people posting their scores on Boards, so don't let that fool you. People aren't going to post their score if they've done badly, meaning more high scores will be posted on Boards, which doesn't give an accurate representation).

Now, last year, 2.6% of Leaving Cert. candidates scored between 550-600 points in their Leaving Cert.; approx. 58k people took their LC last year, meaning ~1500 people scored between 550-600 points. Assume that the same will be true for this year, that means that 1500 people (if they ALL chose to go for medicine, which is unlikely, as there are other courses out there) will already have between 550 and 560 points in the bag. Now, what scores are these likely to get in the HPAT? Well, they should follow the normal distribution (i.e. that 68% of them will score within one standard deviation), but I'm assuming that there's some bias, and one third (just an arbitrary number) of these 1500 people score above the 75th percentile. That means that 500 people will have between 710 and 770 points. There are roughly 400 places, meaning that anybody (in my opinion) who scores above 720 will be in with a shot. (By the way, I hope I'm right, please don't get your hopes up because of me).

sarahbrennan

logged on to my cao... where my hpat results are it still only gives my overall score and says n/a under each individual score?? really annoyed now, anyone know why that is? i did my leaving cert in 2008 so i already have my results from last year , could that be anything to do with it?

TheMightyDouche

StatsMan wrote: »

Hello to you all, I have been following this thread for a while now and just want to clear up some of the facts and figures people have been throwing around to help myself and hopefully some of you have a clearer picture of what your results mean. most of this has been taken from an article in today's Irish times

http://www.irishtimes.com/newspaper/ireland/2009/0623/1224249340275.html

Firstly around 2,900 people sat the exam - therefore there are 29 people (give or take 1) in each percentile of the results.

This also means that there at most 2,900 people competing for the places on offer.

Also the article points out that 480-490 places are on offer.

Therefore (480/2900) 16.5% of those applying will get in.

Taking into account the fact that a number of these places are reserved for people other then the "bog -standard" (non-mature, non- underprivileged schools) applicant - just as a very conservative guess - 400 will still be available to these applicants.

But also it must be understood that not all the 2,900 people who took the hpat were actually "bog - standard" applicants. Again - as a very rough guess maybe 300 of those sitting the test were mature/underprivileged schools applicants (excuse my lack of knowledge of the actual name for this type of entry)

Based on this 400/2600 or 15.3% of all bog standard applicants will get a place.

So where does that leave you with your hpat result in your hand. Well in my humble opinion this is how it will be

Those in the top 1% of hpat may get in once there LC is 500 (maybe even someone with 480 will get in but I have my doubts)

Those in the top 3 or 4% of hpat still require a LC of maybe 530+ to assure themselves of a place

Those in the top 4-20% hpat or so will be hoping for a LC of 550 or more. Then they should be confident of a place.

Those just outside this (maybe 20- 35, at most really) will need a near perfect leaving cert (550 as an absolute minimum) and then will be hoping that a large number those above them in the hpat do poorly in the leaving cert (which is by no means impossible)

Those below the top 35% of the HPAT - to be honest I don't see how you will get in this year.

Great post Statsman. Just hope the Irish Times got their figures right! Where were people getting the 3356 applicants i heard yesterday?

MrPain

TheMightyDouche wrote: »

Great post Statsman. Just hope the Irish Times got their figures right! Where were people getting the 3356 applicants i heard yesterday?

Here:
http://www2.cao.ie/dir_report/pdf/AppStats01Feb2009.pdf
I think it also contains grad entries

Is it 2900 or 3356? If it's 3356, then adjust what I wrote. I'd guess the cut-off point will be around 720-730. And depending on whatever bias the HPAT has towards high-points scorers, it could even be 740.

dog_pig

-JammyDodger- wrote: »

Where are people going saying each percentile rank contains the same number of people? Does it say somewhere on the HPAT website that this is the case? I doubt it is; would the results not fit a normal distribution curve?

Take IQ for an example: it follows a normal distribution. It certainly doesn't contain the same number of people in each percentile rank. My guess is that the HPAT would be no different.

It's impossible to work out exactly how many people are in each percentile range without knowing what standard deviation they've used; as far as I know they haven't revealed any of their methods.

If it does follow a normal distribution (or something similar) then the vast majority of people will have scored around the 50% percentile (it would be something like 68% of people within one standard deviation i.e. from 35% - 65%, or something similar - can't work it out without knowing their standard deviation). Again, something like 95% of people will fall within two standard deviations of the mean, which would be something like 20% - 80%. My guess would be that at an absolute max, only 7-8% of people will have scored above ~80%. People say that 2900 people took the exam, so that means (conservatively), that 232 people will have scored above 80%; which isn't an awful lot, really. (There's a bias on people posting their scores on Boards, so don't let that fool you. People aren't going to post their score if they've done badly, meaning more high scores will be posted on Boards, which doesn't give an accurate representation).

Now, last year, 2.6% of Leaving Cert. candidates scored between 550-600 points in their Leaving Cert.; approx. 58k people took their LC last year, meaning ~1500 people scored between 550-600 points. Assume that the same will be true for this year, that means that 1500 people (if they ALL chose to go for medicine, which is unlikely, as there are other courses out there) will already have between 550 and 560 points in the bag. Now, what scores are these likely to get in the HPAT? Well, they should follow the normal distribution (i.e. that 68% of them will score within one standard deviation), but I'm assuming that there's some bias, and one third (just an arbitrary number) of these 1500 people score above the 75th percentile. That means that 500 people will have between 710 and 770 points. There are roughly 400 places, meaning that anybody (in my opinion) who scores above 720 will be in with a shot. (By the way, I hope I'm right, please don't get your hopes up because of me).

The percentiles refer to the total number who took part

dog_pig

dog_pig wrote: »

The percentiles refer to the total number who took part

I'm not sure I get you...

If it fits to a normal distribution, then the number who scored 150 and the number who scored 210 aren't the same. If it was, the graph of score vrs. the % who got that score, would just be a horizontal straight line.

Edit: For example, it should fit a bell curve (I'm assuming it fits to a normal distribution), similar to this:

http://en.wikipedia.org/wiki/File:IQ_curve.svg

The only difference would be that the mean would be 150 instead of 100. As you can see the vast majority (68%) score within one standard deviation. This is true for all normal distributions, so it should be true for the HPAT (assuming it fits to a normal distribution, which it very well may not, in that case I'm wrong).

TheMightyDouche

MrPain wrote: »

Here:
http://www2.cao.ie/dir_report/pdf/AppStats01Feb2009.pdf
I think it also contains grad entries

Is it possible that 456 people put down medicine as 1st preference but didnt sit the hpat?

pathway33

-JammyDodger- wrote: »

Is it 2900 or 3356? If it's 3356, then adjust what I wrote. I'd guess the cut-off point will be around 720-730. And depending on whatever bias the HPAT has towards high-points scorers, it could even be 740.

3356 applied for medicine as their first choice in January. Another 10,000 applied for medicine on their CAO as a second or lower preference in January. 2900 did HPAT. Therefore up to 456 people who had medicine as their first choice either

1. didn't know about HPAT or
2. missed the exam through unfortunate circumstances or
3. simply changed their mind about medicine between January and the HPAT.

We also have to assume that some of the other 10,000 who had medicine as a lower than first preference did do HPAT and possibly did quite well but will not change their first preference thereby giving medicine hopefuls....more hope.

btw the 2900 figure seems too perfect. is that just rounded up or down?

MrPain

TheMightyDouche wrote: »

Is it possible that 456 people put down medicine as 1st preference but didnt sit the hpat?

Well I was in the trinity test centre and in the rows on either side of me their was two empty seats with numbers on the table there could have been more but thats just what I saw.

dog_pig

-JammyDodger- wrote: »

I'm not sure I get you...

If it fits to a normal distribution, then the number who scored 150 and the number who scored 210 aren't the same. If it was, the graph of score vrs. the % who got that score, would just be a horizontal straight line.

Edit: For example, it should fit a bell curve (I'm assuming it fits to a normal distribution), similar to this:

http://en.wikipedia.org/wiki/File:IQ_curve.svg

The only difference would be that the mean would be 150 instead of 100. As you can see the vast majority (68%) score within one standard deviation. This is true for all normal distributions, so it should be true for the HPAT (assuming it fits to a normal distribution, which it very well may not, in that case I'm wrong).

What you are saying now is exactly what I thought when I was trying to grasp how it worked.

If we take the 99th percentile as an example and say that the total cohort is 3000 people. Then this percentile will be the top scoring 300 people who took the tests?

Now what I am confused about is what happens if more than 300 people get the same score (which is likely around the average scores) which is sort of similar to what you are saying I think.

MrPain

dog_pig wrote: »

What you are saying now is exactly what I thought when I was trying to grasp how it worked.

If we take the 99th percentile as an example and say that the total cohort is 3000 people. Then this percentile will be the top scoring 300 people who took the tests?

Now what I am confused about is what happens if more than 300 people get the same score (which is likely around the average scores) which is sort of similar to what you are saying I think.

I think you mean 30, Im not sure if this is what your asking, but if its an equal score for the last few places its given to people by random selection.

ThehPatroniser

Here's my letter to The Irish Times:
Many thanks for your interesting an informative article in today’s Irish Times. There is intense interest among “the applicant community” in the question of the relative influence of the H-pat result. In order to make the necessary calculations it is necessary to know the number of persons who have applied for Medicine, and also to know the number of “places”. Lively discussion on the “Boards” has been based on the premise that 3,356 persons put Medicine as their first preference on their CAO application (see CAO stat http://www2.cao.ie/dir_report/pdf/AppStats01Feb2009.pdf ). Secondly, it has been assumed that there are 405 places in Medicine available (I don’t know where this statistic comes from) . On this basis the general feeling was that those on the 88th percentile (and upwards) could be reasonably confident of a place, if everything else were equal. Accordingly to your article less than 3,000 persons sat the H-pat. (This seems incongruous if the CAO figures are correct). You state that there are between 480 and 490 “places”: this is at odds with the figure of 405 which has been mentioned on the Boards. I very much hope that you will be writing about this matter again, and if you do it will be very helpful if you would look at the disparities in question. I have a number of more general criticisms to make of the H-pat process. In the first place the nature of the subject matter of test is cloaked in secrecy. Secondly, it appears that some of what might be described as the “emotional intelligence” questions are capable of more than one correct answer. Thirdly, not all those sitting the test last February suffered the same physical difficulties. Those who sat the test at the RDS were left standing around in the cold for more than an hour after the appointed time. The atmosphere was very much “crisis” – even before the test began. Persons sitting the test in other venues, apparently, were seated reasonably promptly. My fourth point is that communication with H-pat is extremely difficult. For example, the test results were supposed to be delivered by e-mail. I did not receive the necessary e-mail. Despite my best efforts I could find no telephone number to contact H-pat. Their website gives no telephone number and nobody else seems to have a telephone number for them. My fifth point is that there is no Appeal whatsoever. An applicant is not entitled to view his or her paper. The suggestion is that the system is “fool proof” and is incapable of human error. This reminds me of the Computer Voting debacle. Finally there is a more general point to be made. The real solution to the “crisis in medicine” is to produce more doctors, by increasing the number of places. The H-pat simply increases the “lottery element” in the selection process. It is an interesting and unpredictable variable, but, fundamentally, the system will not be improved until the number of medical places is increased.

StatsMan

To clear up some more if I can

1. If you dont do HPAT you cant qualify for undergraduate medicine.
Therefore only max 2900 (approx, figure from this mornings Irish Times) are in the race

2. there are 480-490 places (IT this morning)

3. Some of these are kept for mature applicants and applicants from underprivileged backgrounds - guess this may take up 80 of the places (probably less)

4. People sitting the HPAT can be broken down into
(a) People under 23 for med- not from underprivileged backgrounds
(b) People under 23 for med- from underprivileged backgrounds
(c) Mature applicants


A very unscientific guess would say group (a) contains 2,600 applicants
(b) contains 150 and (c) contains 150

Therefore for the majority using this fourm - you 400 places are on offer for 2600 applicants - 15% will get a place.

Finally I am a conservative man - I think I have probably overestimated the number of places for people in groups (b) and (c) and I have underestimated the numbers going for these places - so maybe the chances are slightly higher for those people in group (a).

Good luck to you all

MrPain

ThehPatroniser wrote: »

Here's my letter to The Irish Times:
Many thanks for your interesting an informative article in today’s Irish Times. There is intense interest among “the applicant community” in the question of the relative influence of the H-pat result. In order to make the necessary calculations it is necessary to know the number of persons who have applied for Medicine, and also to know the number of “places”. Lively discussion on the “Boards” has been based on the premise that 3,356 persons put Medicine as their first preference on their CAO application (see CAO stat http://www2.cao.ie/dir_report/pdf/AppStats01Feb2009.pdf ). Secondly, it has been assumed that there are 405 places in Medicine available (I don’t know where this statistic comes from) . On this basis the general feeling was that those on the 88th percentile (and upwards) could be reasonably confident of a place, if everything else were equal. Accordingly to your article less than 3,000 persons sat the H-pat. (This seems incongruous if the CAO figures are correct). You state that there are between 480 and 490 “places”: this is at odds with the figure of 405 which has been mentioned on the Boards. I very much hope that you will be writing about this matter again, and if you do it will be very helpful if you would look at the disparities in question. I have a number of more general criticisms to make of the H-pat process. In the first place the nature of the subject matter of test is cloaked in secrecy. Secondly, it appears that some of what might be described as the “emotional intelligence” questions are capable of more than one correct answer. Thirdly, not all those sitting the test last February suffered the same physical difficulties. Those who sat the test at the RDS were left standing around in the cold for more than an hour after the appointed time. The atmosphere was very much “crisis” – even before the test began. Persons sitting the test in other venues, apparently, were seated reasonably promptly. My fourth point is that communication with H-pat is extremely difficult. For example, the test results were supposed to be delivered by e-mail. I did not receive the necessary e-mail. Despite my best efforts I could find no telephone number to contact H-pat. Their website gives no telephone number and nobody else seems to have a telephone number for them. My fifth point is that there is no Appeal whatsoever. An applicant is not entitled to view his or her paper. The suggestion is that the system is “fool proof” and is incapable of human error. This reminds me of the Computer Voting debacle. Finally there is a more general point to be made. The real solution to the “crisis in medicine” is to produce more doctors, by increasing the number of places. The H-pat simply increases the “lottery element” in the selection process. It is an interesting and unpredictable variable, but, fundamentally, the system will not be improved until the number of medical places is increased.

Good letter. This is constuctive critism, it could do with paragraphs (just makes it easier to read), the 405 comes from www.qualifax.ie. I think they are right to have no appeals since they adjust their marking scheme in very fancy ways and they would only have people complaining and would never get everyone happy and since its marked by computer the only human error is the candidate not marking the box correctly.

dog_pig

MrPain wrote: »

I think you mean 30, Im not sure if this is what your asking, but if its an equal score for the last few places its given to people by random selection.

Sorry yea I meant 30, typo.

Why would it be done by random selection?

hpat 123

i did my lc this year n it says n/a on my hpat score 2.......so damn annoyd......first thy hav such a vague scoring system.... all under covr...now thy wont evn tel me wat my individual scores r .....apparently they r not availabl.......hate this:mad:

TheMightyDouche

yeah good letter hPatroniser, hope they print it! Theres none in todays IT....and yeah RDS was ****, freezing and disorganised

dontlikeurmam

-JammyDodger- wrote: »

I

This is true for all normal distributions, so it should be true for the HPAT (assuming it fits to a normal distribution, which it very well may not, in that case I'm wrong).

Aren't those IQ graphs built up over years to produce a standard measure of the IQ of the population?...I would expect that the percentiles for the Hpat are more of a straight forward indication of your score in comparison to others, as in all the results are divided up or split up into one hundred roughly similar groups and then percentiles are allocated to the groups.I think it would explain why the percentiles for the 50%ish mark are fairly set (around 150?) and the percentiles for the high marks are over a range, e.g 99% between 195 and 202.

I know i'm very late but i got 182 (91%) so i reckon I need 550 to get in
550+182=732

Waaahhh

The boys over at the CAO have removed the breakdown of the results.... Must've been a mistake putting them up in the first place. From what I can remember from reading HPAT information a long time ago we were told that a breakdown of the results wouldn't be available. I'm guessing an angry phone call was made from ACER (I'm guessing they have a phone that makes calls and doesn't receive them, we all know how bloody impossible it is to get through to them:D:D) to the CAO telling them to sort it out quick!!

amacachi

pathway33 wrote: »

3356 applied for medicine as their first choice in January. Another 10,000 applied for medicine on their CAO as a second or lower preference in January. 2900 did HPAT. Therefore up to 456 people who had medicine as their first choice either

1. didn't know about HPAT or
2. missed the exam through unfortunate circumstances or
3. simply changed their mind about medicine between January and the HPAT.

We also have to assume that some of the other 10,000 who had medicine as a lower than first preference did do HPAT and possibly did quite well but will not change their first preference thereby giving medicine hopefuls....more hope.

btw the 2900 figure seems too perfect. is that just rounded up or down?

There will be the same number of people in each percentile, but the range for each percentile can be different. The 50th percentile could be a score of 149-150.5 while the 99th percentile could be 212+ or something similar. Range of scores is different, but the number of people is the same.

hpat 123

wat sorta percentile hav chances 2 get in???

Ormo

Great letter ThehPatroniser.

Conditions at Blackrock weren't ideal as well, poor directions as well as the exam not starting till over an hour after it was supposed to.