Score reports: wide band vs low, narrow band

[johndoe3344]

What's the difference between a low narrow band versus a really wide band?

They say the width of a band measures consistency, but I don't understand that. There is only one variable here.. the number of questions of a particular topic you answer correctly...

So is it like if you get 50% correct, you get a really wide band, but if you get 45% then it starts narrowing again (but on the low side)?

In other words does it go from high+narrow --> wide --> wider --> less wide --> low+narrow?? That sounds like a pretty stupid system to me...

---

Members don't see this ad.

 

[AndyRSC]

The width of the bands indicates statistical precision of measurement/estimate.

 

[johndoe3344]

The width of the bands indicates statistical precision of measurement/estimate.

I have no idea what that means. At all. ELI5?

Okay so suppose there were 10 questions on pathology. How many do I get right to get a high, narrow band? How many do I get right for a wide band? And how many do I get right for a low, narrow band?

 

[shan564]

Okay so suppose there were 10 questions on pathology. How many do I get right to get a high, narrow band? How many do I get right for a wide band? And how many do I get right for a low, narrow band?

It depends on how difficult the questions were. If you get 100%, you'll get a high, narrow band... unless a lot of other people also get 100%, in which case your band will be high and wide. If you get 0%, your band will probably be low and narrow. If you get 80%, it'll probably be on the high side and the wide side, since there are probably more people who get that many right.

The width of the band doesn't measure your performance, it reports the accuracy with which they can measure your performance. There's no way for you to make your band narrower. Just like how you can't change the standard deviation reported with your final USMLE score because it's based on other people's performance.

 

[johndoe3344]

There's no way for you to make your band narrower. Just like how you can't change the standard deviation reported with your final USMLE score because it's based on other people's performance.

Wait, then, how do you explain "A narrow band indicates consistent performance, while a wide band is suggestive of inconsistent performance." That suggests that you can change your band width if you were somehow more consistent with your performance.

 

[shan564]

I think that's just poor wording in their description. I'm pretty sure they mean consistent performance across all test-takers, not just you.

 

[RedSoxSuck]

It depends on how difficult the questions were. If you get 100%, you'll get a high, narrow band... unless a lot of other people also get 100%, in which case your band will be high and wide. If you get 0%, your band will probably be low and narrow. If you get 80%, it'll probably be on the high side and the wide side, since there are probably more people who get that many right.

The width of the band doesn't measure your performance, it reports the accuracy with which they can measure your performance. There's no way for you to make your band narrower. Just like how you can't change the standard deviation reported with your final USMLE score because it's based on other people's performance.

So just to make sure that I understand correctly, if the band is on the right or left side of the borderline then it's due to whether you got the questions right or wrong. Width of band is determined by how everyone who answered those questions performed. If you answered it correctly , band will be to the right, and if other people answered it incorrectly then ur band will narrow. Making your band to appear narrow on the right side. Walk off grand slam?

 

[johndoe3344]

Okay, so let's say I have a band that's so wide that it ranges from borderline to all the way to the right. Did I do well? Poorly? "Inconsistently"?

Where is the mark that says "You are here"?

 

[DrBowtie]

I always interpreted the width to be based upon block to block performance in each topic. Might be wrong.

 

[johndoe3344]

I always interpreted the width to be based upon block to block performance in each topic. Might be wrong.

Hm. That would actually make more sense...

 

[RedSoxSuck]

Okay, so let's say I have a band that's so wide that it ranges from borderline to all the way to the right. Did I do well? Poorly? "Inconsistently"?

Where is the mark that says "You are here"?

Well in that case, I think it would mean that u have gotten more questions correct than incorrect and same applies to everyone else who answered those questions. That's why it can't pinpoint " you are here" bc it can't differentiate your performance from everyone else.

May be Shan can help!!

 

[RedSoxSuck]

I always interpreted the width to be based upon block to block performance in each topic. Might be wrong.

If that was the case then wouldn't you need 4 different measurements showing all the parameters. In others 4 different sheets of performance profile. One for each block.

 

[DrBowtie]

If that was the case then wouldn't you need 4 different measurements showing all the parameters. In others 4 different sheets of performance profile. One for each block.

No. Each block would be a data point within the band.
Say for cardiac you got 4/4, 5/6, 3/4, 4/4. That would be a fairly tight band.

 

[RedSoxSuck]

No. Each block would be a data point within the band.
Say for cardiac you got 4/4, 5/6, 3/4, 4/4. That would be a fairly tight band.

Oh ok. Could be. Lol

 

[shan564]

Huh, I guess that could be it too. I hadn't thought about that possibility. But I don't know if 46 is a big enough sample size to say anything... if you have 1 question on every section about biostatistics, and your scores are 1/1, 0/1, 1/1, etc, then it seems pretty strange to say that your performance is inconsistent.

On the other hand, it doesn't matter anyway. All that matters is the 3-digit score...