Ended UW with a bang

[osli]

---

Members don't see this ad.

My UW performance throughout has been fairly consistent, with a steady improvement in my "last 5" or "last 7" averages. I haven't suffered from the wild swings that some people seem to have, going from 80 on one test to 50 or worse on the next. In fact, my standard deviation on the whole question bank was only 8.5, and my average of "last 7 blocks" standard deviations was just 5.

But I did have a good positive upward trend... first 7 blocks averaging 51.5%, last few 7 block runs averaging just over 70%. It did bother me though that I couldn't seem to break 80, ever. I thought surely I'd have a lucky run towards the end, but no. With exactly 48 questions left, I had one more shot... and pulled a 90. 🙂 OK, so it was really an 89.58. And the national average on that block was 60%, one greater than the next highest (a few blocks had 59). And it was 50% pathology, which I have been doing quite well in. But still... what a great confidence booster going in to the exams!!

For you guys that are averaging in the 80's (and better!?!), hats off. 👍 It took a convergence of factors and a lot of luck for me to get one block out of 40 that high.

BTW, for those wondering if UWorld really gets harder toward the end, I took a look. "National" average score over the last 5 and 7 block runs dropped about 1% from beginning to end for me. However, I did notice that the variation seemed to go way up towards the end (from 48% on one block to 60% on the next), and all of the "highs" and "lows" were definitely in the second half of my 40 tests. I also noticed that subjects became more clumped towards the end, as well as actual subject areas. A couple of tests ago I had four questions dealing with a pituitary tumor of some sort in a row. I kept going back and forth to check, not believing that they were actually asking almost the same thing.

 

[R_C_Hutchinson]

my Uw Performance Throughout Has Been Fairly Consistent, With A Steady Improvement In My "last 5" Or "last 7" Averages. I Haven't Suffered From The Wild Swings That Some People Seem To Have, Going From 80 On One Test To 50 Or Worse On The Next. In Fact, My Standard Deviation On The Whole Question Bank Was Only 8.5, And My Average Of "last 7 Blocks" Standard Deviations Was Just 5.

But I Did Have A Good Positive Upward Trend... First 7 Blocks Averaging 51.5%, Last Few 7 Block Runs Averaging Just Over 70%. It Did Bother Me Though That I Couldn't Seem To Break 80, Ever. I Thought Surely I'd Have A Lucky Run Towards The End, But No. With Exactly 48 Questions Left, I Had One More Shot... And Pulled A 90. 🙂 Ok, So It Was Really An 89.58. And The National Average On That Block Was 60%, One Greater Than The Next Highest (a Few Blocks Had 59). And It Was 50% Pathology, Which I Have Been Doing Quite Well In. But Still... What A Great Confidence Booster Going In To The Exams!!

For You Guys That Are Averaging In The 80's (and Better!?!), Hats Off. 👍 It Took A Convergence Of Factors And A Lot Of Luck For Me To Get one Block Out Of 40 That High.

Btw, For Those Wondering If Uworld Really Gets Harder Toward The End, I Took A Look. "national" Average Score Over The Last 5 And 7 Block Runs Dropped About 1% From Beginning To End For Me. However, I Did Notice That The Variation Seemed To Go Way Up Towards The End (from 48% On One Block To 60% On The Next), And All Of The "highs" And "lows" Were Definitely In The Second Half Of My 40 Tests. I Also Noticed That Subjects Became More Clumped Towards The End, As Well As Actual Subject Areas. A Couple Of Tests Ago I Had four Questions Dealing With A Pituitary Tumor Of Some Sort in A Row. I Kept Going Back And Forth To Check, Not Believing That They Were Actually Asking Almost The Same Thing.

Gj!

PS- I feel your pain on the wanting to do better part, I'm pretty deep in the bank now and I'm still just barely sloping upward...the spreadsheet gives me an r^2 value of 0 🙂

 

[doctorFred]

BTW, for those wondering if UWorld really gets harder toward the end, I took a look. "National" average score over the last 5 and 7 block runs dropped about 1% from beginning to end for me. However, I did notice that the variation seemed to go way up towards the end (from 48% on one block to 60% on the next), and all of the "highs" and "lows" were definitely in the second half of my 40 tests. I also noticed that subjects became more clumped towards the end, as well as actual subject areas. A couple of tests ago I had four questions dealing with a pituitary tumor of some sort in a row. I kept going back and forth to check, not believing that they were actually asking almost the same thing.

i'm not entirely sure what you're asserting here, but i'm fairly sure that the questions are totally randomized. not only that, but many people (including myself) have taken smaller 16 and 24 question blocks, meaning that even if the questions were dispensed in the exact same order, our question blocks would be different (i.e. frameshift!)

 

[osli]

i'm not entirely sure what you're asserting here, but i'm fairly sure that the questions are totally randomized. not only that, but many people (including myself) have taken smaller 16 and 24 question blocks, meaning that even if the questions were dispensed in the exact same order, our question blocks would be different (i.e. frameshift!)

I'm asserting that like many other "random" sequences generated from simple algorithms, this one too may be subject to a little bit of non-randomness. They're probably just using a random number generator coupled to a simple selection algorithm, and while that should be enough, in my experience it isn't perfect.

I'm not asserting that there is any intentional bias to the question distribution. The 1% drop from beginning to end can easily be explained by chance over that interval, and I posted that to put other people's suspicions to rest that the qbank intentionally gets harder as you go. It doesn't seem so.

However, the odds of the test to test variation going up so predictably as I moved through the bank are getting a little uncomfortable to explain through chance. And getting four questions in a row on one single subject... yeah, those are some pretty stout odds! Drawing four of a kind from a 52 card deck is hard enough. Try a 2000 card deck.

I don't know if it is the random number generation, or part of the selection algorithm (programming glitches can be quirky... it could be something as simple as which questions are edited last, or the furthest in the past, or have a certain attribute tag in the database, or a certain range of ID numbers, that skews the randomness of the selection). And I can't say for sure. But it does seem likely.

And it doesn't matter. Up until the very last test, the breakdown of subjects was still plenty good enough to provide for a solid learning experience. And even that one was a great review of path!

 

[mudphudwannabe]

PS- I feel your pain on the wanting to do better part, I'm pretty deep in the bank now and I'm still just barely sloping upward...the spreadsheet gives me an r^2 value of 0 🙂

My graph looked a lot like that too. 😉