1(self.webpackChunk_N_E=self.webpackChunk_N_E||[]).push([[5082],{57931:function(e,t,i){"use strict";i.r(t),i.d(t,{default:function(){return ee}});var a=i(2784),n=i(97729),r=i.n(n),s=i(36977),o=i(35505),h=i(8614),c=i(76621),l=i(66324),d=i(51883),p=i(67049),m=(i(34268),i(84225),i(58371),i(5478),i(83308),i(6456),i(62280),i(18666),i(54705),i(69346),i(66147)),g=i(24647),x=i.n(g),A=i(73083),u=i(99447),w=i(51647),f=i(65078),y=i(35496),b=(i(67888),i(97536),{src:"/_next/static/media/heritage.96c37ff0.png",height:945,width:1445,blurDataURL:"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAgAAAAFCAMAAABPT11nAAAABlBMVEX7+vn+/v3VsV2MAAAAAXRSTlP+GuMHfQAAAAlwSFlzAAAXEgAAFxIBZ5/SUgAAAB1JREFUeJxjYGBkgAJGRkYIh5ERwmFgBFEgUSgBAAGyABguzyQaAAAAAElFTkSuQmCC",blurWidth:8,blurHeight:5}),k=i(78188),j=o.ZP.section.withConfig({displayName:"communitystyle__Section",componentId:"sc-2sa8qb-0"})(["padding-top:60px;padding-bottom:60px;@media (max-width:1024px){padding-bottom:50px;}@media (max-width:768px){padding-bottom:40px;}.container{@media only screen and (max-width:1366px){max-width:1170px;}@media only screen and (max-width:1024px){max-width:960px;}}"]),v=o.ZP.div.withConfig({displayName:"communitystyle__SectionHeading",componentId:"sc-2sa8qb-1"})(["max-width:880px;margin:0 auto;text-align:left;.slogan{color:",";font-weight:700;line-height:22px;letter-spacing:-0.005em;margin-bottom:1.25rem;}h2{font-weight:800;font-size:26px;letter-spacing:-0.5px;margin-bottom:1.875rem;@media (min-width:1024px){font-size:26px;}@media (max-width:768px){font-size:32px;}@media (max-width:480px){font-size:15px;}}.description{margin-bottom:2rem;@media (max-width:480px){font-size:14px;line-height:1.8;margin-bottom:1.5rem;}}ul{padding-left:20px;list-style-type:disc;li{margin-bottom:1rem;line-height:1.6;&:last-child{margin-bottom:0;}}}}h3{margin-top:2rem;}"],(0,k.R)("colors.secondary")),E=(o.ZP.figure.withConfig({displayName:"communitystyle__Figure",componentId:"sc-2sa8qb-2"})(["margin:60px 0 0;text-align:center;@media (max-width:768px){margin-top:40px;}@media (max-width:480px){margin-top:35px;}img{margin:0 auto;}"]),j),I=i(52322),C=function(){return(0,I.jsx)(E,{children:(0,I.jsx)(A.Z,{width:"1400px",children:(0,I.jsxs)(v,{children:[(0,I.jsx)(w.Z,{content:"Top Marks AI Correlates with Edexcel Better Than Humans: IGCSE English Literature"}),(0,I.jsx)(f.Z,{className:"description",content:(0,I.jsxs)(I.Fragment,{children:[(0,I.jsxs)("p",{children:["Time and again, we\u2019re asked this crucial question:"," ",(0,I.jsx)("strong",{children:"how accurate are Top Marks\u2019 GCSE English AI marking tools?"})]}),(0,I.jsx)("p",{children:"So we have been conducting a series of experiments to help provide answers."}),(0,I.jsx)("p",{children:"In today\u2019s experiment, we will be looking at the Edexcel IGCSE - specifically, the 30-mark Literary Heritage question."}),(0,I.jsx)("p",{children:"On their website, Edexcel have published 85 exemplar essays for the Literary Heritage question. These exemplars represent a broad range of quality of answers. These essays are made available for standardisation purposes - so teachers can see what various levels of responses actually look like in the wild."}),(0,I.jsx)("p",{children:"We downloaded all 85 of these essays \u2013 all handwritten \u2013 and put them through our Literary marking tool. Then we measured the correlation between the official marks the board gave the essay, and the marks Top Marks AI gave those essays."}),(0,I.jsxs)("p",{children:["We used a measurement called the ",(0,I.jsx)("strong",{children:"Pearson correlation coefficient"}),". In short:"]}),(0,I.jsxs)("ul",{children:[(0,I.jsxs)("li",{children:["\u2022 A value of 1 w
1ould mean ",(0,I.jsx)("strong",{children:"perfect correlation"})," -- when one marker gives a high score, the other always does too, and when one gives a low score, the other always does too."]}),(0,I.jsxs)("li",{children:["\u2022 A value of 0 means ",(0,I.jsx)("strong",{children:"no correlation whatsoever"})," -- knowing one marker's score tells you nothing about what the other marker gave."]}),(0,I.jsx)("li",{children:"\u2022 Negative values would mean the markers systematically disagree - when one gives high scores, the other gives low scores."})]}),(0,I.jsx)(w.Z,{as:"h3",content:"For context, how do humans perform?"}),(0,I.jsx)("p",{children:"What sort of correlation do experienced human markers achieve when marking essays that have also been marked by a lead examiner?"}),(0,I.jsx)("p",{children:"In an authoritative academic paper that AQA has cited before Parliament, a set of scripts were marked, first, by AQA human markers, and then, second, by the chief examiner. The chief examiner was not told what the original examiners had given these scripts. Nor were they shown any annotations."}),(0,I.jsxs)("p",{children:["The Pearson correlation coefficient between the scores these experienced examiners gave and the chief examiner was"," ",(0,I.jsx)("strong",{children:"0.67."})," This indicated a positive correlation, though far from perfect. If you are interested, you can find the study"," ",(0,I.jsx)("a",{href:"https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1754-8845.2009.01032.x",target:"_blank",rel:"noopener noreferrer",children:"here"}),"."]}),(0,I.jsxs)("p",{children:["And this study is not a fluke. Cambridge Assessment took 200 scripts that had been given an initial score by OCR markers, then sent these same scripts for blind marking to experienced human examiners. The correlation between the two sets of scores was below 0.7. You can find the study"," ",(0,I.jsx)("a",{href:"https://drive.google.com/file/d/1iiKzBW25PUT9BGeQ2PuLNGSERdzdJFxl/view",target:"_blank",rel:"noopener noreferrer",children:"here"}),"."]}),(0,I.jsx)(w.Z,{as:"h3",content:"How did Top Marks AI perform?"}),(0,I.jsxs)("p",{children:["Top Marks, across the 85 essays, achieved a correlation of ",(0,I.jsx)("strong",{children:"above 0.91"})," -- an incredibly strong positive correlation that far outperforms the experienced human markers in the Cambridge study. (Top Marks AI was also not privy to the \u201ccorrect marks\u201d or any annotations)."]}),(0,I.jsxs)("p",{children:["Moreover, ",(0,I.jsx)("strong",{children:"80% of the marks"})," we gave were within a 10% tolerance of the mark given by the chief examiner. In other words, 80% of all the marks we gave were within 3 marks of the board\u2019s grade."]}),(0,I.jsxs)("p",{children:["Another interesting metric is the ",(0,I.jsx)("strong",{children:"Mean Absolute Error"}),", for which it scored a 2.4. On average, the AI differed from the board by 2.4 marks, which is comfortably within tolerance. As a percentage, that's an average of"," ",(0,I.jsx)("strong",{children:"8% different"}),"."]}),(0,I.jsxs)("p",{children:["In contrast, in that same Cambridge study, experienced examiners marking a 40-mark question showed a Mean Absolute Error of 5.64 marks. This equates to ",(0,I.jsx)("strong",{children:"14.1% difference"}),". On a 30-mark scale, this would proportionally adjust to a Mean Absolute Error of 4.23 marks, or the same 14.1%. These results highlight the exceptional accuracy of Top Marks AI compared to traditional marking practices."]}),(0,I.jsxs)("p",{children:["For transparency, you can also access"," ",(0,I.jsx)("a",{href:"https://docs.google.com/document/d/1zjTxgGGn1A4JBcJem0sOoSU5Rrw2KZhqiqXD5OYQDcQ/edit?tab=t.0",target:"_blank",rel:"noopener noreferrer",children:"85 exemplars we used here"}),", and see where we sourced them from."]}),(0,I.jsx)(x(),{up:!0,children:(0,I.jsx)("iframe",{src:"https://docs.google.com/spreadsheets/d/e/2PACX-1vSmyCYqQOg3HIrHadbeQk8CIHxvJXAecoh9BlGwOzS63Maj7aiRKIlTXkGfI2P_mlOwq-r0-X-si8lZ/pubhtml?gid=0&single=true&widget=true&headers=false",width:"100%",height:"500",style:{border:"none"},title:"Google Sheet with Updated Marks"})}),(0,I.jsx)(w.Z,{as:"h3",content:"Can I see a graph to help me visualise this?"}),(0,I.jsx)("p",{children:"Absolutely."}),(0,I.jsx)("p",{children:"First, here\u2019s a scatter graph to show you what a theoretical perfect correlation of 1 would look like:"}),(0,I.jsx)(x(),{up:!0,children:(0,I.jsx)(u.Z,{src:y.Z,alt:"Perfect Correlation Graph"})}),(0,I.jsx)("p",{children:"Now, let\u2019s look at the real-life graph, drawn from the data above:"}),(0,I.jsx)(x(),{up:!0,children:(0,I.jsx)(u.Z,{src:b,alt:"Actual Correlation Graph"})}),(0,I.jsx)("p",{children:"On the horizontal axis, we have the grade given by the exam board. On the vertical, the grade given by Top Marks AI. The individual dots are the essays -- their position tells us both the mark given by the exam board and by Top Marks AI. You can see how closely it resembles the theoretical graph depicting perfect correlation."}),(0,I.jsx)(w.Z,{as:"h3",content:"The Handwriting Factor"}),(0,I.jsx)("p",{children:"As mentioned, all the essays we downloaded were handwritten. That Top Marks was able to correlate so closely with the official board grades indicates not only its marking efficacy but also the strength of its transcription technology."}),(0,I.jsx)("p",{children:"It\u2019s worth noting that the largest outlier in our data\u2014Exemplar 29, with a difference of 10.2 marks\u2014is explicitly identified in the board materials as having been written by a dyslexic student. This suggests that the discrepancy may stem from challenges in transcribing the handwriting, rather than any shortfall in the marking capabilities of Top Marks AI."}),(0,I.jsxs)("p",{children:[(0,I.jsx)("strong",{children:"Discover how Top Marks AI can revolutionise assessment in education. Contact us at"})," ",(0,I.jsx)("a",{href:"mailto:[email protected]",children:"[email protected]"}),"."]})]})})]})})})},T=i(95235),M=i(50056),Z=i(41513),O=i(66562),S=o.ZP.section.withConfig({displayName:"testimonialsstyle__SectionWrapper",componentId:"sc-5arbhg-0"})(['padding-top:70px;background-color:#3dcab1;color:white;padding-left:190px;position:relative;&::before{background-color:#3dcab1;content:"";position:absolute;
1left:0;bottom:0;width:100%;height:calc(100% - 30px);}@media only screen and (max-width:768px){padding:50px 0;&::before{height:100%;}}']),P=o.ZP.div.withConfig({displayName:"testimonialsstyle__Content",componentId:"sc-5arbhg-1"})([".swiper-pagination{bottom:50px;left:-48px;@media only screen and (min-width:1024px) and (max-height:1366px) and (orientation:portrait) and (-webkit-min-device-pixel-ratio:1.5){left:-18px;}@media only screen and (min-device-width:768px) and (max-device-width:1024px) and (orientation:landscape){bottom:40px;left:-46px;}@media only screen and (max-width:768px){bottom:0;left:auto;position:static;margin-top:20px;}@media only screen and (max-width:767px){bottom:0;left:0;}}.swiper-custom-pagination{background-color:",";border:0;border-radius:10px;padding:0;height:6px;width:15px;transition:0.3s ease 0s;}.swiper-pagination-bullet-active{background-color:",";width:30px;}"],(0,O.m4)("#3DCAB1",.6),(0,k.R)("colors.tertiary")),B=o.ZP.div.withConfig({displayName:"testimonialsstyle__Testimonial",componentId:"sc-5arbhg-2"})(["display:flex;align-items:center;@media only screen and (max-width:767px){margin-bottom:40px;}"]),_=o.ZP.figure.withConfig({displayName:"testimonialsstyle__Figure",componentId:"sc-5arbhg-3"})(["margin:0 80px 0 0;margin-top:20px;width:25%;@media only screen and (max-width:768px){display:none;}"]),G=o.ZP.div.withConfig({displayName:"testimonialsstyle__Caption",componentId:"sc-5arbhg-4"})(["background-color:#3dcab1;width:50%;@media only screen and (min-device-width:768px) and (max-device-width:1024px) and (orientation:landscape){width:65%;}@media only screen and (max-width:768px){margin-left:5%;width:61%;}@media only screen and (max-width:768px){margin-left:0;width:100%;}h2{font-size:36px;font-weight:400;line-height:55px;letter-spacing:-1px;@media only screen and (max-width:1366px) and (max-height:768px){font-size:32px;line-height:50px;}@media only screen and (min-width:1024px) and (max-height:1366px) and (orientation:portrait) and (-webkit-min-device-pixel-ratio:1.5){font-size:28px;line-height:48px;letter-spacing:0;}@media only screen and (min-device-width:768px) and (max-device-width:1024px) and (orientation:landscape){font-size:28px;}@media only screen and (max-width:768px){font-size:24px;line-height:40px;text-align:center;}}p{margin:30px 0 0;@media only screen and (max-width:768px){text-align:center;}}"]),F=S,N=i(72853),R=i(14205),U=i(12322),D=i(98213),z=i(1165),H=i(94466),X=i(44046),Q=i(79396),L=i(99348),W=i(25787),q=(N.Z,R.Z,U.Z,D.Z,z.Z,H.Z,[{id:1,image:X.Z,name:"Richard Davis, CEO: 80-essay analysis shows Top Marks AI achieving a Pearson correlation coefficient of 0.91, besting experienced markers",designation:"January 13, 2025",quote:"Top Marks Correlates With Edexcel Better Than Humans: IGCSE English Literature - Literary Heritage Question"}]);Q.Z,L.Z,W.Z;function V(e,t){var i=Object.keys(e);if(Object.getOwnPropertySymbols){var a=Object.getOwnPropertySymbols(e);t&&(a=a.filter((function(t){return Object.getOwnPropertyDescriptor(e,t).enumerable}))),i.push.apply(i,a)}return i}function J(e){for(var t=1;t<arguments.length;t++){var i=null!=arguments[t]?arguments[t]:{};t%2?V(Object(i),!0).forEach((function(t){(0,T.Z)(e,t,i[t])})):Object.getOwnPropertyDescriptors?Object.defineProperties(e,Object.getOwnPropertyDescriptors(i)):V(Object(i)).forEach((function(t){Object.defineProperty(e,t,Object.getOwnPropertyDescriptor(i,t))}))}return e}var Y={effect:"fade",pagination:{el:".swiper-pagination",clickable:!0,renderBullet:function(e,t){return'<button class="'.concat(t,' swiper-custom-pagination"></button>')}}},K=function(){return(0,I.jsx)(F,{id:"testimonials",children:(0,I.jsx)(Z.Z,{children:(0,I.jsx)(P,{children:(0,I.jsx)(M.tq,J(J({className:"slider"},Y),{},{children:q.map((function(e){return(0,I.jsx)(M.o5,{children:(0,I.jsxs)(B,{children:[(0,I.jsx)(_,{children:(0,I.jsx)(u.Z,{src:e.image,alt:e.name})}),(0,I.jsxs)(G,{children:[(0,I.jsx)(w.Z,{content:e.quote}),(0,I.jsxs)("p",{children:[(0,I.jsx)("strong",{children:e.name}),", ",e.designation]})]})]})},e.id)}))}))})})})},$=i(55351),ee=function(){return(0,I.jsx)(o.f6,{theme:h.r,children:(0,I.jsxs)(a.Fragment,{children:[(0,I.jsx)(r(),{children:(0,I.jsx)("script",{type:"application/ld+json",dangerouslySetInnerHTML:{__html:JSON.stringify({"@context":"https://schema.org","@type":"Article",headline:"Top Marks AI: Accuracy in Edexcel IGCSE Literary Heritage Question",description:"Discover how Top Marks AI achieves groundbreaking accuracy in marking Edexcel IGCSE English essays, surpassing human markers with a Pearson correlation coefficient of 0.91.",author:{"@type":"Person",name:"Richard Davis",jobTitle:"Founder & CEO",worksFor:{"@type":"Organization",name:"Top Marks AI",url:"https://www.topmarks.ai"}},publisher:{"@type":"Organization",name:"Top Marks AI",url:"https://www.topmarks.ai",logo:{"@type":"ImageObject",url:"https://www.topmarks.ai/_next/static/media/logo-white.e24574d9.png"}},datePublished:"2024-02-01",dateModified:"2025-01-13",mainEntityOfPage:{"@type":"WebPage","@id":"https://www.topmarks.ai/ai-marking-top-marks-edexcel-igcse-literary-heritage"}})}})}),(0,I.jsx)($.PB,{canonical:"https://www.topmarks.ai/ai-marking-top-marks-edexcel-igcse-literary-heritage",title:"Top Marks AI: Accuracy in Edexcel IGCSE Literary Heritage Question",description:"Discover how Top Marks AI achieves groundbreaking accuracy in marking Edexcel IGCSE English essays, surpassing human markers with a Pearson correlation coefficient of 0.91.",openGraph:{title:"Top Marks AI: Accuracy in Edexcel IGCSE Literary Heritage Question",description:"Discover how Top Marks AI achieves groundbreaking accuracy in marking Edexcel IGCSE English essays, surpassing human markers with a Pearson correlation coefficient of 0.91."}}),(0,I.jsx)("link",{rel:"stylesheet",href:"https://fonts.googleapis.com/css2?family=DM+Sans:ital,wght@0,400;0,500;0,700;1,400;1,500;1,700&family=Work+Sans:ital,wght@0,700;1,700&display=swap"}),(0,I.jsx)(c.Z,{}),(0,I.jsx)(l.Z,{}),(0,I.jsxs)(l.v,{children:[(0,I.jsx)(s.Z,{top:0,innerZ:100,activeClass:"is-sticky",children:(0,I.jsx)(d.D,{children:(0,I.jsx)(p.Z,{})})}),(0,I.jsx)(K,{}),(0,I.jsx)(C,{}),(0,I.jsx)(m.Z,{})]})]})})}},45187:function(e,t,i){(window.__NEXT_P=window.__NEXT_P||[]).push(["/ai-marking-top-marks-edexcel-igcse-literary-heritage",function(){return i(57931)}])},67888:function(e,t){"use strict";t.Z={src:"/_next/static/media/2study.47e0efa5.png",height:446,width:2420,blurDataURL:"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAgAAAABCAMAAADU3h9xAAAAElBMVEXP1LDf37vN1aPM0MHh24L48SFU+IcrAAAACXBIWXMAABYlAAAWJQFJUiTwAAAAEUlEQVR4nGNgZWFmYmBgZAQAAGwAEb+zJ3oAAAAASUVORK5CYII=",blurWidth:8,blurHeight:1}},97536:function(e,t){"use strict";t.Z={src:"/_next/static/media/Picture1.4d5d5c04.png",height:1235,width:2793,blurDataURL:"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAgAAAAECAMAAACEE47CAAAABlBMVEX9/Pz18vESIFfFAAAACXBIWXMAABcSAAAXEgFnn9JSAAAAFElEQVR4nGNgQAGMjIwgzMiILAoAAKsACeWlOLkAAAAASUVORK5CYII=",blurWidth:8,blurHeight:4}},35496:function(e,t){"use strict";t.Z={src:"/_next/static/media/hypothetical.cbaf8fff.png",height:768,width:1283,blurDataURL:"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAgAAAAFCAMAAABPT11nAAAABlBMVEX9/f74+PgOonWDAAAACXBIWXMAABcSAAAXEgFnn9JSAAAAFElEQVR4nGNgYGBkZGRABqg8ZAAAAMIABSbyD1UAAAAASUVORK5CYII=",blurWidth:8,blurHeight:5}}},function(e){e.O(0,[4160,5075,5351,6562,1319,3229,6769,1316,6189,6528,7690,9774,2888,179],(function(){return t=45187,e(e.s=t);var t}));var t=e.O();_N_E=t}]);
Line numbers count LF bytes from the start of the resource, as the search results do. Vendor segments are library code the classifier recognised; they are stored but not indexed. Bytes are shown as Latin1 characters, one per byte.