# Help Creating a Formula for Largest Return on MS Application (Stats Not Included)

This forum made possible through the generous support of SDN members, donors, and sponsors. Thank you.

#### To be MD

##### Med School Or Bust
10+ Year Member

Members don't see this ad.
Over the years, @WedgeDawg and @LizzyM have made great resources for helping to calculate an individual candidate's likelihood of acceptance based of their stats. As the GOP would (now) say, "they're fantastic."

But I am trying to address a different question. I want to find out (ignoring stats because I'm assuming candidate X stats are the average for a matriculant at whatever school she applies to) which schools give you the biggest bang for your buck in the application process.

I've already calculated the OOS favorability factor or FF (i.e. a number showing how willing schools are to take OOS candidates from the pool of all applicants, IS & OOS) for 2015-16. After requesting the data in excel format from the AAMC, I calculated this value using Amherst's formula (FF = % OOS matriculants / % OOS applicants).

That's only half the battle. Do any of you number geniuses know how I could also factor in the number of applicants and the number of available seats at each school into the formula? For example, George Washington is very friendly to OOS students (FF = .95), but they received about 15,600 applications for 179 seats. Pardon my slang, but I ain't applying to GW. How do I combine these 3 numbers?

Ideas?

Last edited:

#### tskbdmnd

##### Full Member
It's so difficult man. There's no point. Any reasonable person can tell just by looking. We have wamc if you wanna make sure you aren't deluded.

1 user

#### To be MD

##### Med School Or Bust
10+ Year Member
I got one PM about trying to figure it out, but I think they just wanted my excel spreadsheet.

#### LizzyM

##### the evil queen of numbers
15+ Year Member
What about #OOS matriculants/#OOS appliants. That is the number of 00s applicants for each seat to be filled by an OOS matriculant. I know it doesn't factor in what proportion of students are IS vs OOS but it lets you know how many others are competing for the seat you'd like to be in.

#### Cawolf

##### PGY-2
7+ Year Member
You could determine the mean number of applicants to all schools and then find a way to weight variance from this into your formula?

Say if the mean # of applicants = 5,000 with a SD of 500, then for each SD from the mean you add or subtract from a multiplied factor. We could call your proposed value "X" the applicant # factor "a"; then created a modified value "aX", where a = 1 for a school receiving 5,000 applications (the mean value). Then for each SD above the mean, find some working number to subtract from a, or for each SD below the mean, add to a. In my example above I would say a = 1 - 0.1(# SD from mean).

So a school with 6,000 applications has an a value of 0.8 and a school with 4,000 applications has an a value of 1.2. In the case of GW, a would be a negative value, meaning don't apply.

That is my rambling, sorry!

Edit: I wanted to clarify that my numbers are fictitious and I have no idea about the actual mean number of applicants.

Last edited:
1 user

#### To be MD

##### Med School Or Bust
10+ Year Member
What about #OOS matriculants/#OOS appliants. That is the number of 00s applicants for each seat to be filled by an OOS matriculant. I know it doesn't factor in what proportion of students are IS vs OOS but it lets you know how many others are competing for the seat you'd like to be in.
Thank you for your input!

I may be wrong, but isn't the #OOS matrics / #OOS applicants pretty similar to (the inverse of) the FF? Or... is there a way to change my FF to include the physical # of applicants?

I'm not good at this kind of math. #_#

#### Goro

##### Full Member
Lifetime Donor
10+ Year Member
This is a pointless exercise. Just look at MSAR Online. A med school application is not an algorithm.

Over the years, @WedgeDawg and @LizzyM have made great resources for helping to calculate an individual candidate's likelihood of acceptance based of their stats. As the GOP would (now) say, "they're fantastic."

But I am trying to address a different question. I want to find out (ignoring stats because I'm assuming candidate X stats are the average for a matriculant at whatever school she applies to) which schools give you the biggest bang for your buck in the application process.

I've already calculated the OOS favorability factor or FF (i.e. a number showing how willing schools are to take OOS candidates from the pool of all applicants, IS & OOS) for 2015-16. After requesting the data in excel format from the AAMC, I calculated this value using Amherst's formula (FF = % OOS matriculants / % OOS applicants).

That's only half the battle. Do any of you number geniuses know how I could also factor in the number of applicants and the number of available seats at each school into the formula? For example, George Washington is very friendly to OOS students (FF = .95), but they received about 15,600 applications for 179 seats. Pardon my slang, but I ain't applying to GW. How do I combine these 3 numbers?

Ideas?

#### Lawpy

##### 39% Full Member
7+ Year Member
SDN Ambassador
After requesting the data in excel format from the AAMC, I calculated this value using Amherst's formula (FF = % OOS matriculants / % OOS applicants).
What about #OOS matriculants/#OOS appliants. That is the number of 00s applicants for each seat to be filled by an OOS matriculant. I know it doesn't factor in what proportion of students are IS vs OOS but it lets you know how many others are competing for the seat you'd like to be in.
Thank you for your input!

I may be wrong, but isn't the #OOS matrics / #OOS applicants pretty similar to (the inverse of) the FF? Or... is there a way to change my FF to include the physical # of applicants?

I'm not good at this kind of math. #_#

If I understand the formula correctly,

% OOS matrics = # OOS matrics / # total matrics of that school; and % OOS applicants = # OOS applicants / total applicants to that school

So:

FF = % OOS matrics / % OOS applicants = (# OOS matrics / total matrics) / (# OOS applicants / total applicants) -->
FF = (# OOS matrics /# OOS applicants) * (total applicants / total matrics)

Where I would think that (total applicants / total matrics) is the inverse of the matriculation rate at that school.

Please correct any math mistakes I made!

#### To be MD

##### Med School Or Bust
10+ Year Member
This is a pointless exercise. Just look at MSAR Online. A med school application is not an algorithm.

Ok, Goro. I will heed your advice. I'm the wrong person to try coming up with a formula, anyway.

(But if any statistician out there wants to take on the MSAR challenge of the century and come up with the ultimate formula, holler at me )

#### makennacompton

##### Membership Revoked
Removed
2+ Year Member
Exempt from the formula OOS from Texas. Only the top ten ( and a very random fragment of lesser schools) even offer interviews to Texas applicants. Acceptances even more rare.

Most OOS acceptances are high quality apps from California. California is simply way oversupplied with great applicants that the other states need to boost their stats.

MSAR obscures the fact that Texans aren't very welcome outside Texas

#### tskbdmnd

##### Full Member
If I understand the formula correctly,

% OOS matrics = # OOS matrics / # total matrics of that school; and % OOS applicants = # OOS applicants / total applicants to that school

So:

FF = % OOS matrics / % OOS applicants = (# OOS matrics / total matrics) / (# OOS applicants / total applicants) -->
FF = (# OOS matrics /# OOS applicants) * (total applicants / total matrics)

Where I would think that (total applicants / total matrics) is the inverse of the matriculation rate at that school.

Please correct any math mistakes I made!

It's just a formula they used to figure out the ratio of the % of the matriculants that come from OOS and the % of applicants that come from OOS.
It seems a bit worthless, but it does give you a scale from 0-1. 0 is a lack of OOS matrics. 1 means the % of matrics that were OOS equaled the % of applicants that were OOS.

It's just a ratio you can look at to get a sense of something. Whether it's significant or not depends. It's true that by getting closer to 1, you could say that the school is more open to OOS school. There are so many factors to this though. I don't really think it's too useful.

#### Lawpy

##### 39% Full Member
7+ Year Member
SDN Ambassador
It's just a formula they used to figure out the ratio of the % of the matriculants that come from OOS and the % of applicants that come from OOS.
It seems a bit worthless, but it does give you a scale from 0-1. 0 is a lack of OOS matrics. 1 means the % of matrics that were OOS equaled the % of applicants that were OOS.

It's just a ratio you can look at to get a sense of something. Whether it's significant or not depends. It's true that by getting closer to 1, you could say that the school is more open to OOS school. There are so many factors to this though. I don't really think it's too useful.

I know, all i did was relating two different factors here (that formula and LizzyM's suggestion). I personally don't think such a metric can tell us much.

1 user