DAT percentile

[Halitosis]

hello! for those who took the test last year, can you please tell me what percentile is a 20, 21, 22, or 23? thanx.

---

Members don't see this ad.

 

[panama]

The percentile given for each scaled score varies on individual tests. A 20 PAT on one test may be an 85% and another person?s 20 PAT may be 90%. There have been several previous posts discussing this.
That said, a 21 AA / 21 PAT showed around 93% in each category when I took the test early last fall. 25 on Gen Chem. was a 97%, while 25 in Organic was 98%. Hope this helps.

 

[Biogirl361]

i got a 21 in march and that was 92 percentile at the time.

 

[Calculus1]

Percentiles don't mean anything, the schools don't receive them because they are not constant. The DAT is not like the MCAT where everyone takes it the same day. I got a 21 and my percentile was around the 97th percentile, but Biogirl's is worth the same as mine. I've heard that the percentiles are from a one month period that the test is taken(that amount of time could be wrong, I also heard 3 months). Anyway, I wouldn't worry about the percentiles as much as the raw score.

 

[sajjy]

Percentiles don't mean anything, the schools don't receive them because they are not constant. The DAT is not like the MCAT where everyone takes it the same day. I got a 21 and my percentile was around the 97th percentile, but Biogirl's is worth the same as mine. I've heard that the percentiles are from a one month period that the test is taken(that amount of time could be wrong, I also heard 3 months). Anyway, I wouldn't worry about the percentiles as much as the raw score.

Schools dont receive the percentile? So, what is the use of those percentages?

 

[Ron Jeremy]

I strongly believe that percentiles do matter!

Although the percentile numbers for a given DAT score do vary as stated earlier, in general, the three most important DAT scores are AA, TS, and PAT.

From looking at past DAT score percentiles published by the ADA, a 21 in those three scores means that you scored in the top 10% in the nation. That is why it is so desirable to hit 21. All the schools know this.

22 is approximately the top 5%. That is why Harvard's DAT class ave for the last five years is 22.

Here's the numbers published from the ADA's "Dental Admission Testing Program User's Manual". This document can be downloaded from the ADA.

Here's the percentiles for 1994, 1999, and 2004 respectively.

AA
20, 93.5, 86.8, 89.2
21, 97.0, 92.9, 94.4
22, 98.8, 96.8, 97.3

TS
20, 95.0, 88.2, 90.1
21, 97.4, 93.3, 94.6
22, 99.0, 96.9, 97.2

PAT
20, 94.2, 90.6, 96.7
21, 97.5, 95.0, 98.7
22, 98.4, 97.7, 99.6

When I took the DAT in Sept of 2006, I scored AA - 21 (92%) and PAT -21 (93%).

Yes, the percentiles do vary, but you can still see a clear trend. 21 is consistently in the top 10%, and 22 is consistently in the top 5%.

Therefore, when the ADCOMMS review 3000 applications, only about 200-250 those applicants will have numbers 21 and above.

Most schools only interview about 250 students. Humm....

 

[doc toothache]

hello! for those who took the test last year, can you please tell me what percentile is a 20, 21, 22, or 23? thanx.

www.ada.org/prof/ed/testing/dat/dat_score_frequency.pdf

 

[curiosteph]

A 23 is the 97.7% in AA

 

[USU24]

Schools dont receive the percentile? So, what is the use of those percentages?

I think mostly so that when people (lay people) ask you what your score was and you tell them you got a 20 and they ask what it was out of and you say 30 and they look at you like you got 67% then you can say, "that was in the 90th percentile though." and they still won't get it. But you will feel better about yourself.

 

[jigabodo]

I think mostly so that when people (lay people) ask you what your score was and you tell them you got a 20 and they ask what it was out of and you say 30 and they look at you like you got 67% then you can say, "that was in the 90th percentile though." and they still won't get it. But you will feel better about yourself.

Haha yeah. When I tell my friends I got an AA of 24 out of 30, some will just look at me very sympathetically and say "I am very sorry..."

 

[amalgamgrillz]

Haha yeah. When I tell my friends I got an AA of 24 out of 30, some will just look at me very sympathetically and say "I am very sorry..."

Is the American DAT scaled/curved according to other applicant scores?

 

[achilles2]

The scores have increased alot this year.

Last year I think from the DAT score freq

a 20 = 94.1
21 = 97.6
22 = 99.2
23 = 99.7
24 = 99.9
25 = 100


This year people are scoring high so the freq has changed

I think a 21 now breaks into the 90th%ile
21 = 90
22 = ~96
23 = ~98
24 = ~99
25 = 99.9
26 = 99.9
27 = 100

those are the scores I've seen posted so far this year

In short people are stepping up their game and its getting harder for a score to be more competitive....

 

[doc toothache]

This year people are scoring high so the freq has changed

I think a 21 now breaks into the 90th%ile
21 = 90
22 = ~96
23 = ~98
24 = ~99
25 = 99.9
26 = 99.9
27 = 100

Highly unlikely, but until the ADA releases the figures for 2008 we are dealing with pure speculation.

 

[achilles2]

Highly unlikely, but until the ADA releases the figures for 2008 we are dealing with pure speculation.

I could be wrong but I feel I remember people posting scores of 21 and saying it was between 90-92%ile....


the only real data I have is from last year, so I agree with you time will tell

 

[UCfan]

I think Achilles can be right:

The scores have increased alot this year.

Last year I think from the DAT score freq

a 20 = 94.1
21 = 97.6
22 = 99.2
23 = 99.7
24 = 99.9
25 = 100


This year people are scoring high so the freq has changed

I think a 21 now breaks into the 90th%ile
21 = 90
22 = ~96
23 = ~98
24 = ~99
25 = 99.9
26 = 99.9
27 = 100

those are the scores I've seen posted so far this year

because let me tell you what happened last june (08). I and four more chemical engineers took the test in the same week. Let me tell our scores:

Mine aa: 21- 91 percent
friend 1 aa: 22- 95.2 percent
friend 2 aa: 22- 95.2 percent
friend 3 aa: 24- 99.1 percent
friend 4 aa: 26- 99.9 percent

which is very well matched with his statement.

 

[doc toothache]

I think Achilles can be right:

The scores have increased alot this year.

About 2 years ago, the ADA frequency table from Jan to June reported only about 2K applicants having taken the test. From July to Dec an additional ~10K took the DAT. Such a dramatic increase in the frequency would indicate a great deviation from the premise of a standardized test, and that is not likely to have happened.

 

[water88]

I think Achilles can be right:

The scores have increased alot this year.

Last year I think from the DAT score freq

a 20 = 94.1
21 = 97.6
22 = 99.2
23 = 99.7
24 = 99.9
25 = 100


This year people are scoring high so the freq has changed

I think a 21 now breaks into the 90th%ile
21 = 90
22 = ~96
23 = ~98
24 = ~99
25 = 99.9
26 = 99.9
27 = 100

those are the scores I've seen posted so far this year

because let me tell you what happened last june (08). I and four more chemical engineers took the test in the same week. Let me tell our scores:

Mine aa: 21- 91 percent
friend 1 aa: 22- 95.2 percent
friend 2 aa: 22- 95.2 percent
friend 3 aa: 24- 99.1 percent
friend 4 aa: 26- 99.9 percent

which is very well matched with his statement.

The 21 AA I received in August was 93%

 

[achilles2]

About 2 years ago, the ADA frequency table from Jan to June reported only about 2K applicants having taken the test. From July to Dec an additional ~10K took the DAT. Such a dramatic increase in the frequency would indicate a great deviation from the premise of a standardized test, and that is not likely to have happened.

I'm basing my estimates off of numbers posted by people who took the test and posted scores. What could happen tho is at you stated the true volume of scores is unknown until Dec, at that point ADA might recompile the data and post correct %'s.

So essentially you take the test in the summer the computer HAS to output some value for a %ile rank so it bases that rank off factors that are not 100% accurate since the future is unknown. Thus the %iles we get on our score reports may be rough estimates.

So hypothetically we could both be right At least in terms of numbers score reports are showing vs the true parameter.

 

[supraman]

10k took the test? My god why does everyone want to do Dent!

 

[doc toothache]

I'm basing my estimates off of numbers posted by people who took the test and posted scores.

Getting an estimate off the numbers posted on SDN can lead to erroneous conclusions since they do not represent a random sample of the population that took the DAT, irrespective of the period in question. For the Jan-June and subsequently Jan-Dec periods mentioned above, there was not a significant difference in the mean scores in spite of the number of applicants. While it is well known that the percentile given to each individual after the exam is taken is a comparison of how well the individual did compared to others, it is not clear if the comparison is limited to those who took the test on the same day, the same, week, the same month, the last 6 months or the past 12 months. If, as you say, the percentages reported were indeed as high as you claim, it is quite possible that the percentages reflect the results of only a limited and proximal sample of the population. A tempting guess is that the percentiles reflect a particular version of the test so that if we are at the beginning of the cycle the numbers may tend to wonder way into left field.

 

[Contach]

You don't need a 'random sample' to garner some numbers from people. The type of people has no bearing on what percentile the prometric computers print out.. There is no opinion/confounding factor that the person with the score can/will do to affect the percentile. A number is given to said person by the computer and he/she comes online and posts it - that is all. The only thing you must assume is that every SDNer is telling the truth about percentile.

my 24AA in August was 99.4%

Getting an estimate off the numbers posted on SDN can lead to erroneous conclusions since they do not represent a random sample of the population that took the DAT, irrespective of the period in question. For the Jan-June and subsequently Jan-Dec periods mentioned above, there was not a significant difference in the mean scores in spite of the number of applicants. While it is well known that the percentile given to each individual after the exam is taken is a comparison of how well the individual did compared to others, it is not clear if the comparison is limited to those who took the test on the same day, the same, week, the same month, the last 6 months or the past 12 months. If, as you say, the percentages reported were indeed as high as you claim, it is quite possible that the percentages reflect the results of only a limited and proximal sample of the population. A tempting guess is that the percentiles reflect a particular version of the test so that if we are at the beginning of the cycle the numbers may tend to wonder way into left field.

I think what the FRAP / Piano passages have proven is that the percentiles are not unique to each test. If this were the case, then there would equal amount of people doing well on the FRAP/Piano in comparison to any other RC Passage combination. But, we see that everyone who writes the FRAP/piano gets a horrid RC score and percentile and this is attributed to the other tests being easier, and their scores being compared to these easier tests.

 

[doc toothache]

You don't need a 'random sample' to garner some numbers from people. The type of people has no bearing on what percentile the prometric computers print out.. There is no opinion/confounding factor that the person with the score can/will do to affect the percentile. A number is given to said person by the computer and he/she comes online and posts it - that is all. The only thing you must assume is that every SDNer is telling the truth about percentile.

You are right- "You don't need a 'random sample' to garner some numbers from people" especially when we are not interested in attaching any significance to the "garnered numbers".

 

[klutzy1987]

My 22 in June was in the 96.4%.
Unfortunately my 18 RC was 40%