I completely understand the fact that there is a UMS scale at which raw marks are converted to reflecting overall student's performances on a paper. But is there any sort of algorithm to how they come to be? Is it something as simple as the top 10% get an an A* for example? Or do they work out what they assume and A and an E to be and then divide by 5 to get the difference between grades? If anyone has any inclination to their actual process of generating the UMS scale please let me know!
I completely understand the fact that there is a UMS scale at which raw marks are converted to reflecting overall student's performances on a paper. But is there any sort of algorithm to how they come to be? Is it something as simple as the top 10% get an an A* for example? Or do they work out what they assume and A and an E to be and then divide by 5 to get the difference between grades? If anyone has any inclination to their actual process of generating the UMS scale please let me know!
I believe they mark all of the countries papers for a set exam first. They then look at the results and calculate the grade boundaries so that, from the people who took the paper, the top certain % of people get A*s As Bs etc.
So the grade boundaries are set afterwards to ensure the correct percentage of people get the grades exam board wants them too.
For AQA, and I imagine for the other exam boards as well, UMS is awarded almost proportionately to your raw mark, and often comes to a plateau/ grade cap where a raw percentage that is less than 100% will give you full UMS. It's not exactly linear, but the curve on the graph line is so small that it has a tiny effect on your UMS in reality (as long as you don't get like 15% raw marks or something stupid).
Take a look at the UMS Calculator I linked above - put in some random subjects and dates, and then take a look at the graph it produces. For example: image
I get the original idea of how UMS is calculated. But how do they decide the raw mark boundaries that contribute to a particular grade in a subject? I know that raw mark boundaries are calculated in relation to how the rest of the country. But what is the method used to do so?
I heard the boundaries were calculated using a normal distribution curve. If so, then I assume that the mean mark would be a C grade. And how would skewness affect results?