How To Learn From Your Errors:

by on April 7th, 2010

errorWhen I make an error, I get excited. Seriously – you should be excited when you make errors, too. I know that I’m about to learn something and get better, and that’s definitely worth getting excited!

Errors can come in several different forms: careless errors, content errors, and technique errors. We’re going to discuss something critical today: how to learn from your errors so that you don’t continue to make the same mistakes over and over again. First, let’s define these different error types.

Careless errors

Remember those times when you were sure you got the answer right, only to find out that you got it wrong? For a moment, you even think that there must be a mistake in the answer key. Then, you take a look at the problem again, you check your work, and you want to slap yourself on the side of the head. You knew exactly how to do this problem and you should have gotten it right, but you made a careless mistake!

By definition, a careless mistake occurs when we did actually know all of the necessary info and we did actually possess all of the necessary skills, but we made a mistake anyway. We all make careless mistakes (yes, even the experts!); over 3.5 hours, it’s not reasonable to assume that we can completely avoid making careless mistakes. Our goal is to learn how to minimize careless mistakes as much as possible.

Content Errors

“Content” is the actual knowledge we need to know in order to answer a question. What’s the formula for the area of a circle? What are the rules for noun modifiers? Content errors typically come in two forms: knowledge you did know but forgot, and knowledge that you didn’t know, or didn’t know well enough, in the first place.

Technique Errors

Beyond the content itself, we can typically work through any quant or sentence correction problem in multiple ways; the particular method we choose to use is the technique. For reading comprehension and critical reasoning, of course, all we have is technique; no actual knowledge is being tested on these question types. We also need to employ timing techniques, in terms of both individual questions and the overall section.

The Error Log

Your first step is to create an error log. You can do this in a notebook or an electronic file, but have one consistent place where you can record your errors. I typically record careless mistakes separately from all other mistakes, but you can organize things however you want, as long as the organization is consistent. Then, you can use the error log to learn from your errors!

For each problem, keep track of this data:

1) The basics: where the problem can be found again in your materials, the question type to which the problem belongs (as specifically as possible), the content category being tested (if applicable), the time you spent, and the current date.

2) The error: describe the error in specific detail; if applicable, actually copy into your file the part of the work where you made the error. (Note: one problem could have multiple mistakes; include them all.)

3) The reason: figure out WHY you made this error and write that down; if there are multiple reasons, note them all. The next step hinges on this step, so make sure you really dig deep to figure out why! If you can’t figure out why, then you can’t figure out how to fix the problem. (See more on this, below.)

4) To Do: figure out what habits you need to make or break in order to minimize the chances of making that particular mistake again. For example, you might:

  • create flashcards to help you memorize some content or technique that you didn’t know or messed up
  • re-write your work for this problem in its entirety and try the problem again in a week
  • do several problems of the same type, or drill certain skills, in order to build a new, good habit
  • decide that whenever you see a certain type of hard and relatively infrequent problem, you’re just going to make an educated guess and move on – so learn how to make an educated guess and practice moving on!

Whatever it is, do the necessary work to create good habits and destroy bad ones.

5) Review and reinforce: at least once a week, review your log. Are there certain types of mistakes you tend to make repeatedly? Are you continuing to make mistakes that you’ve made in the past and already tried to fix? Go back to steps 3 and 4 again.

The simple fact that you’re now aware of your tendencies will allow you to notice when those kinds of problems pop up on the test. When you’re already aware, then it’s easier for you to double-check the parts of your work where you’re most likely to make a mistake – or, if necessary, to let the problem go.

WHY did I make that mistake?

Let’s talk more about figuring out why you made a mistake. Careless mistakes will usually be pretty obvious. When you’re looking through your work, something will jump right out at you. You added when you should have subtracted. You thought something out in your head instead of writing it down. You calculated area instead of circumference. You missed the word “not,” which negated the entire answer choice.

Quant content errors also tend to be more straightforward, but quant technique errors can sometimes be tricky to fix. Don’t assume that the first technique you tried is the one you have to use. Read the explanation, check out some online forums, and try to find different, better ways of tackling the problem.

Verbal errors can be even trickier to understand. Whenever you pick a wrong answer (or you guessed and got lucky), ask yourself several things:

1) Why did I pick the wrong answer? Something about it looked good; something about it made me think it was right. What was that thing (or those things)? Now I know those aren’t good reasons to choose an answer.

2) Why did I eliminate the right answer? Something about it looked wrong. What was that thing (or those things)? Now I know those aren’t good reasons to eliminate an answer.

3) Why is each wrong answer wrong? (As specifically as possible!) Why is the right answer right? (Sometimes, the answer to that is: it’s the only one left!)

There’s one type of careless error I want to address specifically: when we meant to choose one answer (the right one!) but accidentally chose another. It’s especially disheartening when this happens, and it often happens because of sloppy scrap paper technique.

On quant, it is critical to write down what the problem asks for you to solve. On problem solving questions, I leave a little space for me to do the work, and then I write what I want to find and circle it. Then I go back and do the work in the space I left above. When I’m done with the work, I run right into my “x=___?” circle and I’m much less likely to, for example, pick the answer that actually represents y. On data sufficiency questions, I write the question at the top and the two statements below, and I’ve made it a habit to check the question after each step.

On verbal, it is critical to keep track of your thinking for every answer choice. First, write down “ABCDE” vertically, just as the answer bubbles appear on the screen. Next, you need three consistent symbols. One means “definitely wrong,” one means “maybe…” and one means “right!” As you think through each answer, make the corresponding symbol on your scratch paper. You can use any symbols you want, as long as you always use the same symbol for each category. When you’re ready to choose an answer, circle that letter on your scrap paper, then immediately look up and select the corresponding bubble on the screen.

Okay, you’re ready to learn from your mistakes. Go start that error log right now!


  • Thanks a lot for this article! It'll really help me with organizing my error log in a better manner...and now I'll have a source of definition for each error type! :)

  • Dear Stacey,

    Wow, Its great to see a timely post dealing with "Errors". Very much insightful.

    It would be nice to hear how to "reuse " the tracked data of errors so that I am not repeating it again?

    Plz do mention some strategies which I can use to collect "data" from the error log & how effectively I can leverage the knowledge of known errors so that I am NOT repeating them again.

    In my case, invariably I mess up with tenses in SC.How to go about it? I am pretty clear of the concept of Tenses, but once I see them in a test set up, I get excited & and make of fuss of it & mess it up. So how to avoid these kind of scenarios.

    Stacey, On apersonal note, I would be happy if I can see a post dealing with elite strategies that a candidate should apply for 770 levels.

    Please share us the Gyan of that elite methodology to crack 770 figures.

    • Good questions. Part of the issue with errors is simply to be aware of them. If you know where you tend to mess up, you can be extra careful and "alert" when doing a problem, or part of a problem, that you know is problematic for you.

      If you see that you keep making the same type of error over and over again, then you also have to make a very concerted effort to fix that error. For example, I once had a student who constantly mixed up the formulas for area and circumference of a circle. Eventually, she developed this habit: at the start of quant, she drew two circles at the top of her scrap paper. Next to one, she wrote the area formula and she shaded in the circle. Next to the other, she wrote the circumference formula and she drew a couple of arrows pointing to the outside (or circumference "line") of the circle. She also wrote "area" and "circumference" next to the correct circles. When she got a problem dealing with either of those, she checked her little diagrams against the word "area" or "circumference" up on the screen, and made sure she was using the right formula.

      By the time she got to the real test, she was able to simplify the above to writing down the two formulas and matching the "A=" to "area" on the screen and "C=" to "circumference" on the screen - but she still made herself double-check every time.

      So, you basically set up even more elaborate work and checks to hammer this idea home. Most of the time, by the time you get to the real test, you will be able to simplify whatever your "check" process is, because you will really have internalized it.

      For tenses, for example, you may want to make some flash cards with parts of OG sentences on them, and then you drill exactly HOW to recognize particular tenses in the context of actual OG problems (not just isolated examples from a grammar chapter). What is it that tells you this problem is testing present vs. past perfect? Etc.

      I like your idea for a "770 strategies" article. I'll put that on my list. :)

  • Excellent article that I have come across on how to make your "errors" as swords in the GMAT war....!!!!

  • I don't get excited per se when I make an error, but I don't get frustrated over it any more. It's sort of bitter sweet. It just means that I need to focus on that topic some more.

    • Dear Richard,

      Nice to hear that..very optimistic way of seeing things. Actually, it makes sense. I have devised the same methodology so that I can fix all that bugs. As a Business Analyst, when i tend to do some UAT, i do all the ways of functional testing to identify a bug so that the client doesnt scream at me or my team .SO same technique I am gonna apply to fix all the weak links.


      Thx a ton! I am eager to c ur 770 elite prep methodology!

  • I learn from my errors and i find it the best way , because every one know their weakness.

    • Stacey,
      Nice article! I am getting started on practice questions from the OG.
      An error log is a wonderful tool but I would like to know how do I juggle timing limit and entering information in the error log when I am working through questions? Recently, I read a debrief from somebody who did not enter detailed information but kept codes (For example, 1 - Concept Error and so on) and entered this code in the error log.
      What are your thoughts on this "information overload" in the error log and ways to get around this problem during timed practice?

  • Don't fill out your error log WHILE you're practicing. That's what you do while you're reviewing the problem. If you're doing a set, always leave yourself enough time to do the problems and then review the set immediately after - if I'm going to do a 2-hour study session, for example, I do problems for about 30m and then I spend the remaining 1.5h reviewing (part of which is reviewing and learning from my errors).

  • Hi Stacey,

    I'm starting out on my GMAT plan and in the process of creating my error log on excel around your article.

    I have a quick question from your bullet-point labeled "The basics", within "The Error Log" section. Quoted below:

    "1) The basics: where the problem can be found again in your materials, the question type to which the problem belongs (as specifically as possible), the content category being tested (if applicable), the time you spent, and the current date."

    Could you give me an example for each of the following:
    -the question type to which the problem belongs ?
    -the content category being tested ?

    I have an idea of what you meant but I would just like to be sure before I start.

    Thank you in advance for your prompt response.



    • Sure! The overall question type is one of these 5: DS, PS, SC, CR, or RC (the 5 main GMAT question types)

      Content category varies depending on the question type. So for DS or PS, you might have things like "circles," "ratios," "inequalities." For SC, you might list specific grammar concepts: parallelism, pronouns, etc.

      For CR and RC, you'd list based upon the sub-type of question, and that will depend upon whose materials you're using to study. For example, if you're using MGMAT materials, you might list a CR question as "find the assumption," "draw a conclusion," "analyze the structure" and so on, because those are the categories we use in our books.

      Good luck with your study!

  • Hi Stacey,

    Wish I had read this article earlier :(

    One question - How do you deal with the information overload?
    I frequently get into the issue that I end up committing the same mistake multiple times - Now two things can be done
    1) Do NOT proceed. Correct this error, find the root cause and eliminate the problem. Only then Proceed
    2) Prioritize the error based on the relative ranking of this error w.r.t the other errors.
    If it is an error that can HURT you really, follow step-1, else follow step-2

    Please let me know if this approach is correct or any alternative way of handling these errors.

    • Do try to figure out why you made the mistake - that will partly help you decide what to do. If you realize - oh, I was rushing because I was running out of time, well ok - the real thing to correct is wherever you spent too much time earlier.

      But if you realize something like "I solved for the wrong thing" or "I did something in my head instead of writing it down (eg, quant) or checking for proof (eg, RC)," that's a process error that can hurt you on many questions, so you definitely need to start figuring out how to fix that.

      And certainly if you realize you keep making the exact same mistake, you've got to fix it if it's happening on a problem that shows up with any frequency. If you keep making the same mistake on a combinatorics problem, I don't care - those are infrequent. But if you keep making the same algebra mistake when solving equations, then obviously you have to fix that!

    • Stacey -
      "But if you realize something like "I solved for the wrong thing" or "I did something in my head instead of writing it down (eg, quant) or checking for proof (eg, RC)," that's a process error that can hurt you on many questions, so you definitely need to start figuring out how to fix that."

      I sometimes end up with Process error (These are the worst and time-consuming ones because need to change the thought process) and try to Google it out to see expert replies. BUT in MOST of the places the replies are NOT from experts and NOT the ones that address the root cause. I realize that following those steps will NOT help me in addressing a problem following a similar PATTERN in future. 
      Reasons stated - OG questions cannot be discussed here :(
      Now that are the ones I need to learn and the experts tell this cannot be discussed.

      Any comments on this ?

    • You don't need to discuss an OG question (or any particular question) to discuss how to deal with a general process error. Just describe the error, not the problem itself.

      For example, let's say you solve for Bobby's distance instead of Susie's distance, or x instead of y. I talk about some ways to fix that in the article above, but we didn't discuss any actual problems in order to discuss it.

      Or let's say you did something in your head instead of writing it down on quant (or checking the proof on RC). You notice this after. Then you write "Write Everything Down!" on a piece of paper (or "Check the Proof!") and for the next 10 problems, you look at this piece of paper and make sure that you do whatever you wrote. After you finish a problem, you actually go back and make sure you wrote everything down and didn't do anything in your head. If you DID do something in your head, you make yourself do the entire problem again, start to finish, writing everything down or checking the proof or whatever - even if you got the problem right. It doesn't matter that you happened not to make the mistake this time - the bad habit will cause you to make random careless mistakes later, so you have to change it.

      We just discussed another process error without any mention of a specific problem. :)

    • Stacey - "You don't need to discuss an OG question (or any particular question) to discuss how to deal with a general process error. Just describe the error, not the problem itself."

      There is a real problem that is the MOST COMMON if I do not post the exact question and my thought process. MANY TIMES we are committing a Process error and NOT identifying the correct process error. This is especially TRUE in Verbal.
      example - In a SC, I feel that I am committing an IDIOM error whereas really there is nothing wrong with the IDIOM, but something else.
      Without the question, no one is able to see my thought process.

      In Quant, I can at least replicate the process by quoting a similar variant. But how do you project the errors in your thought process for Verbal without the actual question ? I have NOT become an SME, so asking these questions !!!

      My thought process for the Process error that has a GAP put in bracket :
      I need to identify the error (HOW TO Project this to the experts/SMEs in that area for Verbals)--> Correct the error --> Ensure the error does not repeat again

      Stacey - Help me plug the gaps.

    • Ah, ok - you're talking about a slightly different thing. First, you're talking about identifying the error in the first place (as opposed to asking for help fixing an error that you already have identified). Second, it sounds like you might be talking about a content error, not a process error. A process error has to do with HOW you are working through something. A content error has to do with WHAT you're supposed to know in order to work through it.

      I thin that's why you want/need to quote the problem - because you're really talking about content errors. And I don't have a good answer for you re: how to get *free* help with OG problems on content errors. Unfortunately, the makers of the official test have announced that they are invoking their copyright rights and they don't want OG problems posted online. A lot of sites do post these problems anyway, but most of the reputable companies / teachers avoid discussing them because we all want to stay on GMAC's good side.

      For our class and tutoring students, we discuss these problems in class / during tutoring and during office hours (which is legal, because everyone in the class has purchased a copy of the book, so we can use the material together under fair-use laws). But that, of course, involves spending money to take a class or work with a tutor.

      I wish I had a better answer for you. We did used to answer OG questions on our own forums up until a few years ago when GMAC expressed their displeasure with the posting of their copyrighted material online. (Note that they do allow GMATPrep questions to be posted online, because GMATPrep is a free product already.) :(

  • Stacey -
    " how to get *free* help with OG problems on content errors."
    I feel it is free only in monetary sense. But really it is NOT FREE because I am spending my time and giving back to the community by sharing knowledge. Ultimate goal is that everyone should become an expert - So helping others does NOT HURT. We can learn a lot when helping by getting OOB (Out Of the Box) approaches when helping a hugely diverse community. It is really GIVE and TAKE, though unfortunately we have to GIVE a lot at the beginning.

    "Unfortunately, the makers of the official test have announced that they are invoking their copyright rights and they don't want OG problems posted online. "
    ok - I am NOT posting the content. But can I put the OG version and the question number ? example : "OG#12 - CR-103". 
    After that put my thought process without actually putting the content.
    Does posting like this violate any copyright issues of GMAC ?
    I want to know the best way to communicate my thought process without HURTING others.

    "(Note that they do allow GMATPrep questions to be posted online, because GMATPrep is a free product already.) "
    Wow- That is great.....So, can we post GMATPrep questions online without any violation of GMAC copyright issues ?
    For the OG, I can at least look into the official solution. But for GMATPrep questions, there are no official Answers - So that would be great.

    P.S. Anyway I have bought an official copy of GMAC.

    • When I said free help, I meant free help from experts (that is, people who normally charge for their expertise). :) If you're talking about being able to discuss this with fellow students, just follow the rules of whatever forum you're on. If that forum doesn't prevent you from posting the text of an OG problem, then you may be okay to do so - after all, GMAC hasn't told you specifically that they don't want you to post the text. They've told me (and other experts like me / companies like mine), but you're a student - you're different.

      Yes, according to GMAC, GMATPrep questions can be posted and discussed freely. We post those on our own forums and the experts discuss them.

      For OG, according to GMAC, you are allowed to discuss the questions as long as you don't post the problem in full. I think it's even okay to quote a small part of the text as long as you don't reproduce all or most of the full problem. Beyond that, GMAC will technically be mad at you. (They don't seem to be suing anybody or doing anything bad to anyone who does post the full text, but as I said, the reputable companies / teachers generally want to keep on GMAC's good side, so as long as they say they don't want us doing something, we're not going to do it.)

      Discussing a problem without posting the text is a lot more cumbersome, though - we don't do that on our own forums because it's harder and more confusing to discuss something when you can't just put the problem right there.

  • I am taking the GMAT again after scoring a dismal 580. In my quest to find the elusive 720, I have first tried to identify the mistakes done by me in my preparation. Here are some of my post in this regard. However, I sincerely wish to know if I can achieve the score within one month. Madam, I know you are busy person. However, I sincerely and humbly request you to identify some broad patterns of errors in my preparation.
    I will be most obliged.

    • Hi, I'm sorry - I no longer participate on the BTG forums. We have two teachers, Ron Purewal and Whitney Garner, who do and you can PM them to ask for a response.

      You can also post on our own (MGMAT) forums in the General Strategy folder, where I answer questions. (Note: it can take 5 to 7 business days to get a response.)

      As a short response: most people would struggle to increase from 580 to 720 in one month. I'm not saying it's impossible - but most people would need more time!

  • Hi Stacey,
    I gave a practice test and scored 730(Q46,V45) but i am *not* satisfied because a) my target is at least 770 b) i scored worst ever on quant as my usual score is Q49-50. I am happy that i have improved on verbal and a lot of credit goes to you. You have guided me whenever i felt astray on any of my strategies. I have identified mistakes that i made and will be working on them. However, i have few queries before i work on improving myself:
    1) Analyzing the results of test, i saw that i got 1 weaken & 1 paradox question wrong. Now, i don't know what this means for me. Earlier when i was scoring 680, too many incorrect weaken questions meant that i needed to improve my strategy for solving weaken question. Please help me understand what these 2 wrong questions mean for me and what should i do to improve.
    2) one problem that i successfully identified is that as i score more correct, i get more longer, difficult passages and i took too long (5 min) to read them. which habits should i incorporate to get used to those convoluted passages that i will be encountering on my way to 770
    3) On RC questions, irrespective of difficulty level of passage, I tend to move Back-Forth among options. Though i get them correct in the end but i want to stop myself from going Back-Forth because doing so will help me reduce time for answering RC question.  
    4) I see that although i got almost equal number of incorrect on DS and PS, i generally took more time on DS question (irrespective of correct/incorrect). I mean my average time for DS is greater than average time for PS. any strategy for reducing taking too much time on DS

    Thanks !!

    • Why are you going for a 770? Are you just trying to give yourself some leeway in case your performance drops on the real test? Are you trying to work for us? :)

      Answers to your questions:
      1) You scored 45 on verbal - 99th percentile - which means there very likely won't be any big patterns in terms of strengths and weaknesses. You'll have a few mistakes here and there in almost random areas mostly because there was some very obscure rule you didn't know or because you fell into a tricky trap. If you miss a weaken CR, it doesn't (necessarily) mean that you're bad at weaken CRs in general. It might just mean that you fell into some kind of trap or made some kind of mistake on this one.

      So you really just need to go through and figure out exactly why you missed any question that you did miss. Once you figure out why, that tells you what you need to do or study or practice in order to maximize the chances that you'd get a similar question right in future.

      2) That makes sense - as you lift your score, you're presented with harder material. If you're taking 5+ minutes to read them, then you're letting yourself get caught up in the details. You do NOT want to do that on the read-through; the read-through is only for the big picture ideas and contrasts or changes of direction. Stop trying to understand so much detail on the read-through - seriously. :) You can go back and very carefully read any details you need *once you know that you have a question about that detail.* But don't read ALL of the detail carefully because you'll never need most of it.

      3) Once you've narrowed down *any* verbal question to two answers, compare the two ONCE more. That's it. Then pick and move on. This is a discipline thing - you just can't let yourself go back and forth EVER. You either do know after that first comparison or you don't; things aren't going to change after going back and forth three more times.

      4) Dig into the details. Do you have just a few DS questions that are pulling your overall average higher? Or are there many that are just a little too long (10-15 seconds)? If the former, you need to learn to identify these questions around the 60 to 90 second mark and cut yourself off at that point. (You can still take a little time to make an educated guess.) If the latter, then you need to see what specific skills are slowing you down (translating from english to math? manipulating equations? etc) and drill those things until you can speed up a little bit.

      If you *ever* find yourself thinking any of these things:
      - but I studied this!
      - but I should know how to do this!
      - if I just had some more time, I'm sure I could figure it out...
      - I've already invested so much time on this one; I don't want to waste that time by giving up now!

      Cut yourself off IMMEDIATELY. Make an educated guess if you can and haven't used too much time already. Move on.

  • Hi first of all good work,one gets clarity on important issues ones needs to focus on.I have a question can you please elaborate on Technique errors.

  • Thank you Stacey for sharing this information.
    Thought it was late, i was happy to find this now, so that I can reduce the number of mistakes.

Ask a Question or Leave a Reply to Gautam Kumar

The author Manhattan Prep gets email notifications for all questions or replies to this post.

Some HTML allowed. Keep your comments above the belt or risk having them deleted. Signup for a Gravatar to have your pictures show up by your comment.