 I'm just waiting for it to start. Okay. So it's recording already. Pag narinigniyong sound na yan, please tell me because the recording actually stopped in my previous class. That's why I have to record it again. Okay. So let's go now to method evaluation and quality control. So when we were in, when you were in quality, this is very important. Okay. Some of this are actually been discussed to you before. So we'll be doing some of the review and some refinement and reinforcement of the thing that you know already. So for this morning we will be talking about basic concepts, method evaluation, quality control, reference interval studies, and diagnostic efficiency. I hope you still remember quality control down to diagnostic efficiency because that was discussed during your quale lecture during second semester of your second year. In addition to that, basic concept is also been taught to you from quale and even in your biostatistics before. Your method of evaluation will be one important addition to this discussion. So please listen intently. But for this morning in particular we will only be discussing basic concept. This basic concept will be the foundation for your method evaluation, for your quality control, for your reference interval study, for your diagnostic efficiency. You will not be able to accurately compute for the reference value or the reference interval of a particular analyte if you did not understand the basic concept. You will have a hard time understanding what the method evaluation is all about if you did not understand the basic concept. In short, you need to listen very carefully. Not because I don't have a quiz today that doesn't mean you can just review during my time with the other classes that will be coming in. You should have reviewed last night and not utilize my time to review for any of your upcoming quiz this afternoon. Having said that, let's have a quick story first. Take for example, there is a sick patient. I believe that most of you had a hard time appreciating quality management and quality control last semester because you hardly realize the importance of it in the laboratory medicine. So here now, am I clear on your side, Daniel? Clear yung audio ko? Yes, po. Okay. So moving forward, take for example, there is a sick patient that came in into the hospital. A sick patient with difficulty in breathing, with fever, with high blood pressure, name it. Signing symptoms, that patient has it. So your doctor now, when your patient go to the hospital, an attending physician will check their vitals, will check their signs and symptoms, but eventually will not make a conclusive diagnosis without diagnostic examination. And those diagnostic examination is the part now where laboratory medicine comes in. But maybe one question that should be playing around your mind is that how sure are we that those laboratory results, that those laboratory tests are accurate, are precise and reliable, that after having those tests, I will be able to really identify and diagnose a patient of a particular disease. How sure are we natama yun? How sure are we that that particular test will really be giving me the diagnosis for this patient? And that is now the time where quality management, where quality control comes in. Because your quality control, make sure or just go in general nalang muna, your quality assurance makes sure that all your laboratory data are accurate, reliable and that will give you a strong evidence in concluding for a particular diagnosis. In this particular part, in this particular part, I would also want to emphasize that maybe some of you should be wondering, Sir, why is it, okay, tanong ko lang ha, in hematology, what is the reagent used for the measurement of your hemoglobin? What is the reagent? Correct. And the method to be exact is your cyan met hemoglobin method. And the question you will be asking, sir, why are we using cyan met hemoglobin? And that is another answer that will be given by your quality control, your quality assurance to be more specific, your method selection. And all of those things will be covered in this topic, method selection or method evaluation and quality control. So in a nutshell now, this particular topic that we will be discussing is very important because this is our tool, this is our evidence, sabihin natin in the future. One of your patients sued your laboratory because of giving you, because of the laboratory results that you released and that patient is claiming that that is erroneous. You can win the case by showing your quality control charts, you can win by showing them your quality management protocol. Okay, so bakit na nun, sir? Because your quality management protocol and your QC result, the method evaluation are your evidence that the tests that you are doing are reliable and accurate within the laboratory. Are we clear? So moving forward now, okay, so from the pre-analytic to analytic, the post-analytic quality is very important. Quality is what sets the ordinary from the extraordinary. And if you want to be an ordinary, well kapatid na dito kami to remind you that you should be aiming to become an extraordinary person. And quality is very important if our goal in the board exam is to top the board exam. Ang laging kong ipapapalala sa inyo, hindi mabagsak kayo kundi laging kong ipapapalala sa inyo that we are doing this to top the board exam. And mind you, kagabing na nung naguhugasa ko ng pinggan sabi ko, how can I actually awaken you at times na parang you feel like giving up and you feel like ayaw nyo ng parang tinatamat na kayo. And then I realized just think about your future patients. Just think about treating thousand and million of people when you study hard. And of course think about the hard work of your parents which also would lead me to another homily that hopefully this coming pre-leam none of you will really cheat. For the record, I've been with most of you since first year. I know performance from first year to third year. Yes, it is usual that there will be some students that will be having their awakening moment in the third year. But your quiz scores are also evident kapag nagsushoot up yan out of the blue. So may chikahan moment kami ng faculty and if really nakapag-pop-up ang ano, it will be consistent kasi across all ano. So sir ano naman pung sinasabi mo? So integrity po tayo ha, pagdating sa exam this coming preleams. Again, do your parents a favor. Study hard. Time is really hard at this time. Money is not just being picked anywhere. You have friends who are not able to continue with their studies. And you are the few people who are really, really blessed to be able to continue. And I hope that you will be able to cherish that. May gosh na record pa yung aking homily. Moving forward now, sa lahat ng nanulohod lang video nato para sa inyong lahat yan. So let's go now to the descriptive statistics. Are you still there? Let's go now to the descriptive statistics. The descriptive statistic as defined by Bishop is now the foundation for the monitoring of performance and quality control within the laboratory. Sir, why descriptive statistics? Scriptive statistics enable us to describe laboratory data patterns using your center by describing their center, describing their spread or dispersion, and describing their shape, whether it is normally distributed or not normally distributed. So here, the assessment now of the dispersion or the spread is a powerful tool because it now enables the laboratorians to predict the laboratory measurement, whether it is good or bad, whether it is acceptable or already unacceptable. So that is descriptive statistic, the foundation of monitoring performance and quality control. So let us go now to the first descriptive statistic that we are using to describe quality in the laboratory, and that is your measures of central tendency or your center, which are your mean, median, and your mode. Your mean, median, and mode are actually the most commonly used descriptor for your central tendency. For your central tendency. So we have your mean, we have your median, and we have your mode. So how different is one from the other? Your mean is calculated by getting the sum of all individual values and dividing that sum by the total number of data points. Your mean is also known as your average. And among those three, your mean is the most commonly used. Your median, on the other hand, is a measure for your medial. By the way, I will not be teaching you how to compute mean, not teaching you how to compute median or mode because I am assuming that you already know that. And if you do not know that, okay, if you do not know that, if you do not know that, that is part of you being the student of reviewing the notes that you had in biostat and in oale before. Okay? So moving forward now, your median is, gives you now the medial value of the data. If the data are odd, take for example, we have five data or seven values or nine values, it's easy to look for the median. But take for example, okay, I'll give you an example. What is the median of this, what is the median of this value? One, two, three, four, five, six, seven. What is the median? Four. Correct, the median is four. But what if one, two, three, four, and five? What is the median? Three. What is the median if the numbers are one, two, three, four, five, six? Three point five. Three point five, correct. So what are we trying to, so I'm assuming na alam nyo nga kung paano mag-comput ng median, okay? So I will save my voice for that. So mode on the other hand, wait lang, median, by the way, your median are used when the data are skewed. What do we mean by data are skewed, okay? So take for example, if this are your value, it's easy for you to use your mean, you can use your mean. But it's a different story if the value are already like this, okay? If your value are already like this, okay? Then this is the time now that you will be needing to make use. You will be needing to use your median, okay? So your median is being used if the value are skewed. Sir, when I say skewed, when I say skewed, when I say not, normally distributed, when I say the shape is not motion, all of this pertains to the same thing that the value are skewed. So when I say skewed, meron siyang outliers. We have like 10, 20, 30, 40, and then the next value would be 1000. So when you compute for the mean, when you compute for the average, the average will be affected by the outlier. But if you make use of your median, your median will now be eradicating your outlier. Am I clear? So moving forward now, okay, that is for your median. Your mode on the other hand, among the three, is rarely used, okay? Your mode is rarely used. But your mode is described to be the most frequently appearing, okay? The most frequently appearing value in the data set, okay? The most frequently value in the data set. So your mode is usually being used. Your mode is usually being used if the data that we are trying to describe are bimodal or it has two centers, okay? It has two centers. So please refer, if you want to check how does it look like, you may actually go to figure 3.6. So we have there an example of the Gaussian. We have there an example of, wait lang, let me just try to share my screen na lang. So we have an example there of your Gaussian. You're normally distributed. You're, sorry, para lang magising ka lang. You're skewed. And you're skewed and also you're, there, can you see my screen right now? Hello. So on this part, okay, nakikita nyo tuloy ko kung paano ko mag-notes. So this one is your Gaussian curve, okay? Sorry, I don't, ayok ko burahin na. Pwede ko pa burahin pero kailangan ko yan. So this is your normal distribution or your Gaussian distribution. You can use your mean here, okay? Am I clear? You can use your mean. If the value are now skewed, so when I say skewed, it is non parametric, it is not normally distributed. Are we clear? Okay. And this is how a bimodal distribution look like. It has two centers. Okay. So resuming now on my discussion, okay, we are using your mean median mode depending on the need, okay? Hindi kung ano lang yung trip mo, yun yung gagamitin mo. It depends on the data, okay? So that is for the measures of center. Let's go now to the measure of spread. The measure of spread here this time, okay? For the measure of spread, this allow a laboratory to assess the predictability of the laboratory test. Later on, we will be discussing linear regression. Yes, maagagang linear regression para po sa ating lahat. So it represents the relationship of all data points from the mean. That is the function of your spread. It describes the distance of a particular point from the mean, okay? And in measuring the spread, we have four major parameters. We have your range, standard deviation, your coefficient of variation and your standard deviation index. Maybe some of you are wondering, sir nasan po si variance, okay? Variance is not included here, okay? Variance will just be under your standard deviation. So moving forward now, okay? Moving forward, when we are talking about range, range is the simplest expression because your range is simply the difference between the highest and the lowest data score in your dataset. So yung pinakamatas at pinakamababa, subtract them, that is the range, okay? And that is the simplest expression of your spread. So meaning to say, you are able now, your mean is 100 and your data is 60. The data that you gather is 60. So 100 minus 60, that is 40. So you can say that your value is 40 units away from your mean, ganon, the range. When it comes to your standard deviation on the other hand, it gives you now a clearer picture of the measure of dispersion, okay? It is a measure of dispersion, your standard deviation, dispersion of values from the mean. So it helps us describe the normal curve. Why? It helps us describe the normal curve as we go to one, two, three sd. You would realize that what we are using to describe the distribution, the distance of a value from the mean is actually your sd. Your variance on the other hand is also known as your standard deviation squared, okay? Your standard deviation squared. So if standard deviation is a measure of dispersion, your variance is a measure of your variability, okay? It is a measure of your variability and it is determined, okay? It is determined, it determines significant differences between data group. Remember ANOVA, analysis of variance. So you are using there your variance, okay? We are using there your variance. So please take note when it comes to standard deviation, everybody look at your chat box. When we are talking about standard deviation, your precision and standard deviation are inversely proportional. So meaning to say, if the value of my standard deviation is high, okay? The precision is low. And if the value of my standard deviation is low, the precision is high. Meaning to say, tagalugin natin, let us say it in the vernacular, kapag ang standard deviation mo mababa, ibig sabihin noon, yung values mo magkakalapit. And yun yung gusto natin kapag precision because precision is the closeness of one value from the other. Paano naman kapag mataas yung standard deviation mo? If your standard deviation is high, that only goes to show na ang distance sa paggitan ng mga values mo ay malaki, that's why they are not precise, okay? So take for example, for this particular one, okay? For this particular one, if your SD is high, your precision is low. What do we mean by that? Putting it into picture, this is how it look like, okay? This is how it look like. Your value will be like this. These are your values. They are far from your mean. Far from your mean and far from one another. So the distance here, okay? The distance here, this is your SD, okay? This is your SD. Are we clear? On the other hand, if your SD is low, meaning to say, okay? Meaning to say the distance from one another is only small, the precision is high. Okay? So taking it into a drawing, this is how it look like. So your data are all concentrated here. Okay? Are we clear? Okay. Closeness of the value to one another. Okay. Closeness of your value to one another. Moving forward now. Let me go back. So that is your standard deviation. On the other hand, we also have here your coefficient of variation which is a percentile expression of your mean. And it is also known as the index of precision and it is used to compare your standard deviation. And last but not the least is your standard deviation index of value of a data point and the mean all over or divided by the group's standard deviation. Okay? The group's standard deviation. So among those four, what is the most, what is the simplest? Your range. What is the most commonly used? Your standard deviation. What is the index of precision? Your coefficient of variation. We will not be talking about circles, square, rectangle or triangle but we will be talking about the bell curve, the Gaussian curve or the normal the Gaussian distribution which represents now a normally distributed data. So if your distribution is a Gaussian distribution, it is synonymous to normal distribution and to bell curve. Okay. So when we are saying Gaussian distribution this is how it looks like. This is a Gaussian distribution and a Gaussian distribution sabi natin is normally distributed meaning to say your shape, why do we use your shape? We identify errors by statistical method but there are actually statistical data but it's no longer in normal control. Remember your 1, 2S, your 2, 2S your shape, your trend. If we're going only to base it in their value we cannot identify an error but if you plot it in your chart now and you see its distribution you see the shape that is now the time that you are able to identify a particular shape and a particular error referencing it to the West Guard rules. Are we clear? So kaya siya important. Are we clear? Yes sir. Okay. So most of the things that I am saying here are terminologies but the reason why I am still saying this because I am emphasizing this because I want you guys to understand its significance. Differentiating whether will actually distinguish if you will understand the lesson on the coming parts of this chapter or not. So moving forward now the total area of your Gaussian curve your Gaussian distribution or your normal distribution is 1 or in percent that is 100% and here we can your Gaussian curve is subdivided into three distinct parts your 1SD, your 2SD and your 3SD which I call the 689599 hindi po yan vital statistics nang sumasale sa isang pageant konde that is the normally the distribution in your Gaussian curve 1SD means 68.3% are accepted 95% 95.4% is your 2SD and your 3SD is 99.7% sure what do we mean ba kasi by that 68% that 95% that 95% memorize ko na yan sir 1SD, 68, 95, 99 1, 2, 3SD but what is its significance what is the significance when we say 68% when we say 68% meaning to say if we're going to subtract the 200 the answer is 32% sure what is the 32 kanina 68 what is the relationship of the 32 sir the 32% represent the error na tinatanggap mo ok take for example you have your joa may pagkakamali siya if you are a 1SD person yung 32% of the time tatanggapin mo pag umabot na yan ng 33% ayawan mo na siya well you would say ok siya but in reality sometimes the things that you accept becomes your standard and in laboratory if you keep on accepting errors then I could say that you are your laboratory test are errors and referencing it back to our storytelling a while back we don't want errors in the laboratory we want to be precise we want to be accurate in all the things that we are measuring in the laboratory so that is the meaning of 32% the 95% meaning to say there will be 5% chance that a person or what person humuhugot na there will be 5% chance that the laboratory results will be out of range there will be outliers and that 5% is ok you will be accepting that 5% for me this is a perfect this is a perfect scenario if that person really loves you you will not actually be making a lot of mistakes the 5% is an occasional mistake and that's fine in the laboratory we accept 95% confidence interval sir if I am traumatized by errors in the laboratory I am traumatized with people betraying me I want to go 3 SD well you can that is 99.7% 0.3% or just let me say 1% error that you will be accepting and if that happens you will be alone most of the time because there will be no people who will be perfect enough to reach the 3 SD same thing with laboratory methods there are yes there are methodologies that will reach the 3 SD actually remember your 6 sigma pushes for the elimination of your error that's why we have your 6 sigma the total elimination of error but in reality sooner or later laboratories will move through that ok I don't know if you still remember an analogy in Kuala before 95% confidence interval meant 5% error and that 5% error is a total of 500 planes crashing over 100 500 planes crashing all over 100 planes that that flew it also would mean take for example 5 people dying over 100 people who enter the hospital and sometimes ok 5% is just small but if we put it in a larger scale when we put it in a larger scale that 5% is also big enough that 5% is also big enough but for the sake of discussion for this time in the laboratory we are mostly using your plus minus 2 SD which is 95% confidence interval allowing a total allowable error of 5% 5% we will be rejecting the method ok and that is important as we go along 689599 people of the Philippines are you still there are you still awake ok moving forward now it's already 946 I need to finish this let's move forward so in comparative statistics we can actually compare your mean we want to identify if the difference between the mean is significant we use your t-test if we want to determine if the difference between standard deviations of two group are significant what are we going to use your f-test ok your f-test t per mean f for standard deviation ok so hingan ang malalim because we are entering your descriptive statistics and descriptive statistics of groups of prepared observation and here this is actually use in comparison of method or your COM studies sir what are COM studies going back kanina why is it that we are using glucose oxidase instead of other tests why is it that we we are calling hexokinase as the reference method so as we go along when we go to your your carbohydrate, your cholesterol, triglyceride, proteins bilero bin you will be hearing me say there is a common method and there is a reference method sir what is the difference reference method is the one approved to have highest accuracy accuracy kung baga ito na yung trusted reliable subok at ano nang subok subok at sigurado subok at sigurado para hindi yun yun tried and tested subok at sigurado subok at sigurado these are try and tested methods ok sir bakit po ako gagawa ang comparison of method take for example in the coming years you discovered or you develop a particular method of identifying SARS-CoV-2 sabihin na lang natin na yun in COVID-19 you discovered a method oh i can identify SARS-CoV-2 in just 5 seconds are you sure that is the question of the people of the Philippines but the entire planet that will be the question and they will be asking you now to compare that to a method and what method would that be that is your reverse transcriptase polymerase chain reaction or your RT-PCR so what I'm trying to say here is that in comparison of method we are trying to compare the new method from the reference method why that is for us to make sure that new method if not better from the reference method who doesn't want that an RT-PCR that will take minutes and hours being replaced by a 5 second method people would die for that people would go for that but you have to make sure first that is comparable that is even better or comparable at the very least to the reference method and how do we do that comparison of method studies are we clear? are we clear? yes take for example I'll give you an analogy take for example you consider your dad to be an ideal man you consider your dad as an ideal person that you want to marry in the future you want to have the value that you are going to marry in the future or your mom for the guys as for me I don't care because I'm not getting married so as for the for if you want to get married you will look for a person comparable if not better than my dad not better than my mother so how are you going to do that you compare you compare that person ang datimong ex whether you like it or not means you will be comparing your present from your ex kino-comparaka but you are here to say I am a new method I am different I am not just comparable to that I am better than that so nagigas natin ngayon yung point when it comes to comparison of methods are we connecting? yes so sana hindi lang yung hugot ko yung natatandaan ok so moving forward now our reference method pag-usapan kasi natin ito your reference method is on the x-axis your x-axis is your independent variable also known as your what your other term for x-axis anyone other term for x-axis correct that is your aksi aksi means absisa your x-axis that is your independent variable we also have your y-axis ok so before I move on sabi natin dan your aksi is where your x-axis or your absisa is where the independent variable is that is where your reference method is y hindi kasi ito yung independent kasi na independent so what about the y-axis other term for y-axis ordinate correct your y-axis your ordinate where you can see your dependent variable which is your test method in this case and did you know how we compare test method yan correlation and when we have correlation we have we have your linear regression so kala ko sir tapos na ako sa linear regression wait there's more linear regression is used to represent the relationship between your reference method and your new method so the correlation coefficient so please read it on bishop correlation coefficient it measures the strength of the relationship between two methods if it is positive if the correlation coefficient is positive meaning to say they are positively related or they are directly proportional if negative na man para ito yung para ito yung kwento nung hulimong x you're inversely proportional iba na yung direction na tinatahak ninyong dalawa or perhaps you're me the correlation coefficient is zero which has no relationship okay so in comparing your reference and your new method it's very important to know if they are positively related or inversely related wiser kasi dito nipalang malalaman na natin kapag ba ang isang test ay comparable o hindi sa isang reference method okay and talking about linear regression are you ready guys your linear regression okay is a method of determining the relationship between the dependent variable which is your y and your independent variable which is your x your y which is your test method your x which is your reference method and this is the formula sir mag-compet ba kami hindi po okay hindi po enables us to see the relationship between the reference method and the test method and the question is that sir paano po kapag comparable and acceptable and then meaning to say we can make use of that new method in determining the disease so example natin kanina di ba dun sa sa SARS-CoV-2 what if na find out natin na comparable parayas lang po sila ng performance then we can use that nagigets nyo ako so that is how we establish new method hindi yung naisip mo lang gusto mo na itry agad hindi po siya ganun you need to perform comparison of method studies on the other hand moving forward now to the possibility of difference kasi method yung new method from the reference method the difference between the reference and the test method we call them your errors and errors can be systematic or random your random errors and your systematic errors has a lot of factors that play around let's go now to your random error first your random error let's go to the random error first is present in all measurement and can be positive or negative but typically both positive and negative errors are present most of the time they are present most of the time random errors can be a result of many factors like your instrument your operator your agent and some environmental variation guys are you still there so here's the thing I want you to unearth your quaili notes okay I want you to unearth your quaili notes na dun po nakalagay yung tinasabi natin yung mga causes or factors that causes your random error and I want you to study that as well so those are the the causes of your random error so your random error sabi natin random error are because of chances okay because of chances like take for example minsan naman ok magpipit minsan hindi okay it might actually be just because of chances so your error now that random error can be present but take for example I have 5% there is a 5% error in the laboratory that is tolerable as we say it allowable as we say it on the other hand we also have your systematic error can be further classified into constant error and your proportional error but let us first define what are systematic error so systematic error influences observation consistently in one direction either going high or going low this is being measured the systematic error are being measured by using your slope and your y-intercept sir sounds familiar of course sounds familiar because that is linear regression that is linear regression so going back now your slope and your y-intercept provides you an estimate of the systematic error again can be further classified into your constant error and your proportional error this is the part now where I need your focus to be on me let us go now to your constant error constant error exists when there is a continual difference between the test method and the comparative comparative method values regardless of their concentration during the last discussion last week some of the students were confused but I'm here to tell you stick on the definition stick on the definition but the only way for me to further explain it is to actually use your linear regression formula later on which we will be doing so again your constant systematic error is being reflected by your y-intercept okay y-intercept which leads me now to how I will going how am I going to show you how am I going to show you the error so anyan pa ba kayo nakikita po yung screen yes sir okay first going to discuss your constant error your constant error I want you to take note that this is always the difference pag kakaiba okay to take for example I have the reference method here I have the test method or the new method and then I have the error okay saktan natin yung mga sarili natin pumunta tayo sa timbang take for example nakitimbang ako sa reference method muna tayo nakitimbang ako sa reference method take for example ang weight ko yung 70 take for example si miliya miliya nung weight mo maya masaktan kitay kuna estimate lang to ha kuna re ang weight ni miliya na sa 51 tapos si si Daphne yung weight ni Daphne 30 lang and then si tingan natin take for example si si Teddy yung weight ni Teddy kuna rin nag-work out si Teddy naging 41 that is their original original that is their weight according to the reference method but when you try to use it on the new weighing scale take for example ang weight ko naging 75 ang weight ni miliya naging 56 ang weight ni Daphne naging 35 ang weight ni Teddy naging 46 what I'm trying to say here is this the error is the same it is just five five five and five meaning to say error is constant are we clear yes sir what do you mean you don't understand kilograms close the joke sir I do I was a joke I thought okay sorry ha sir you sako kasi I need to explain this to you very well okay so moving forward I'll give you an example kani napakasito magulod di ba suburahin ko ulit paulit-ulit tayo dito sa paulit-ulit ako dito sa nai na salam niyo ba every Tuesday Monday Monday Thursday from 7.30 to 12.00 direct siya tayong nagdadaldal okay so take for example this is your reference method this is your reference method so take for example I'll take the formula here I'll focus more on the I'll focus more on the y-intercept which is your six take for example six daw yong y-intercept niya if the reference here okay if the reference here is one the test will be what the test will be seven if this is two this is eight okay if this is five this is eleven if this is fifteen then this is twenty one the difference between the two okay if I'm going to compute the difference between the two the difference remains six six in six and that is what constant error is all about are you getting me guys yes sir okay okay hold your horses let's go now to your proportional error the proportional error became a problem last time because unlike sabi ko kanina di ba that regardless of the regardless of the concentration regardless of the weight the error still remains five pa din naging seventy five man naging iba man seventy five milligrams per deciliter and then eighty milligrams per deciliter the difference remains five milligrams per deciliter are you getting me the difference the error is still the same are you getting me guys yes okay so as for the proportional error what is the proportional error naman proportional error okay in this part proportional error is now also an error a difference between the reference method and your comparative method it's just that it is proportional to the analyte concentration meaning to say okay pakitandaan to ha if the slope is one proportional error is present if the slope is one the proportional error is present in addition to that okay in addition to that the error is proportional because it will increase with the analyte concentration okay it will increase with the analyte concentration so hopefully you're still there let me try to share my screen once again guys laban okay if you don't ang constant error bago ko magumpisak claro tayo sa constant error clear po tayo sa constant error yes so linear regression tayo ha y is equal to mx plus b this is your slope this is your y intercept in the formula in the let me try if I can open my okay nakikita po nakasplit ay hindi na nakita okay hindi pala ko pwede mag multitask pagana sorry okay one moment there can you see my screen ko na there can you see my screen now yes po sir okay so in the particular in the particular topic that we were talking about kanina let me just check okay in this particular topic the slope okay the slope is 0.85 0.89 and y intercept is 6 so remember that moving forward now let's go to this part okay on this part so m is equal to 0.68 and your b which is your y intercept is equal to 6 okay so take for example I have dataset here na 5 10 15 20 25 these are the measurement of my reference these are my y okay these are my y take for example trying to predict if there will be proportional error obviously there will be kasi merong 0.68 which is equivalent to approximately 1 okay so sabayan niya ko ha so 5 substitute niyo yung this is x pala sorry this is x so substitute niyo yung x niyo sa m sa x so let's try to compute 5 times 0.68 plus 6 that is 9.4 okay 10 times 6 10 times 6. 10 times 0.68 plus 6 is ano ba yan ayos umaman ng calculator ko 10 times 0.68 12.8 tulungan niyo na lang 12.8 12.8 15 times 0.68 plus 6 16.2 tama 12 times 0.68 plus 6 is 19.6 25 times 0.68 plus 6 is 23 and if you're going to subtract ilan to this is 4.4 this is 2.8 this is 1.2 this is ilan 0.4 0.4 and this is 2 so makikita niyo dito di ba in this particular part you cannot appreciate yung sinasabi ko na as the concentration increases the error increases why? why? I want you to I want you to realize in this particular part both the constant and proportional error are present both the proportional error are present let me take you into another story whereby I will go directly only to your pro proportional same thing y is equal to mx plus b but this time m will still be 0.68 but pag sinabi natin there is no constant error the b will be what? 0 sir, is that possible? yes ladies and gentlemen it's possible because according to bishop ayan according to bishop we have here your your bland atman method in this particular part okay there is only proportional error okay there is only proportional error and the constant error there is no constant error if the y-intercept is 0 okay no error if the y-intercept is 0 okay so meaning to say zero dao yung zero dao yung ating y-intercept so if we're going to compute now as the concentration increases as defined what will now be okay what will now be the value of my errors okay tignanin nyo to ha try to multiply it again by 0.68 alone 5 times 0.68 is what? 5 times 0.68 is ilan 10.4 10.4 10 times 6.8 0.68 is 6.8 6.8 6.8 times 15 10.2 20 times 0.68 that is 13.6 13.6 25 times 0.68 11.17 let's now compute the difference the error in this particular part the proportional error ilan ilan siya 5 is minus 3.4 1.6 1.6 1.6 this one 10 minus 6.8 3.2 3.2 15 minus 10.2 4.8 4.8 20 10.4 6.4 ha 6.4 6.4 6.4 25 minus 17 that is 8 8 you see it now na kita ni can you see the proportional error now as the concentration increases the error also increases did you get it? yes sir yes sir are we clear? are we clear? oh po sir sir question po ah go dary ko nag-exist ng po yung proportional error kapag nipag zero po yung constant error po no in this particular part Mr. Bolaños mas nakikita mo lang mas nakikita mo lang yung proportional error kapag wala pag zero yung y intercept kasi tignan mo here here kasi sabi natin di ba pag 0.68 pag 1 yung slope meaning to say there is proportional error did you na gets back ko sa part na yan? yes sir yes po ok eto ha this is my assumption in why they had to put in writings na if the slope is one proportional error is present why? kasi tignan mo yung ginawa natin tignan yung ginawa natin dito guys ha the slope is 0.68 which is one meaning to say there is proportional error but when we computed it alongside with the constant error alongside with the y intercept nakikita nyo ba na may amininin nyo was there proportional here nakita nyo ba yung proportional wala sir wala kore wala hindi natin siya nakita so ang only basis natin for that is yung value ng slope are you getting me? unlike kaya ko parang dini construct ok kaya ko dini construct kasi when you when your y intercept is 0 which is possible ok which is possible mas nakikita natin yung definition ni proportional error are you getting me guys? yes ok so ang itchuran di ba sabi natin kapag linear regression di ba may line na pagan yan so meaning to say ok if the line is going upwards meaning to say ano yan tawag dito um positively related sila kapag ganyan naman di ba pagpababa ok ibig sabihin they are negatively parang parayas lang sir wait lang natawa ko sa hinawa ko napansin nyo ba parang iisa lang din di ba inibako lang yung inibako lang yung yung position parang pagan nito pa lang ayan so ganyan yung yung negatively negatively correlated sila pagdating natin dito sa ano this is possible pagdating natin sa ano pagdating natin sa yung formula ko na y is equal to 0.68 x plus 0 ang line natin is like this y kasi ito kasi ko na re um ito yung ginagawa natin pag may constant error this is 5 ok plus 5 ok it becomes 10 plus 5 it becomes 15 it will increase nagigets nyo ko sana nagigets ayun sir question po ago everyone po possible mag 0 yung B when kamo oh po pa anong in what instance that is actually the very same question of O2 kanina gole and gole and gab so ito kasi siya did you know that linear regression when we are talking about linear regression the one that I am doing kasi in this formula are already the prediction ok I am predicting if um if I give if I have a value of 80 what would be the what would be the what would be the like what would be the value unlike here let me just share my entire screen ok ok how will I be able to get a constant um a value of my y-intercept as 0 as was miss um that was miss andaya correct yes sir ok if you are looking for an answer like ko nare sa sabihin ko kapag ganito yung pipiting if the pipiting is like this if the machine is like this you will not get the answer from me that way it's not a conditional na physical ko nare pag bilog ang buwan when the value is um when the value is like this or like that wala tayong ganong basis to get a zero value of your y-intercept how are we going to get a value of y-intercept when we compute for the y-intercept which is this one in your linear regression you have this formula for your linear line your linear regression formula but you also have a formula for your slope and a formula for your a formula for your y-intercept so meaning to say miss andaya take for example these are the values ok ipagpalagay natin ok na this value on your right are the reference value and these are the value coming from the test how will I be able to compute for their slope and their y-intercept I will go through the entire formula of linear regression ok and if I find out if overall the value of my slope is zero then that's the time now that's the time now that I can say that there is no constant error are we here na sagot ko kami na nagsayong nag-upayong point ko ayan and in the same manner konwari yung ginagawa ko kasi kanina everyone what I was doing is already predicting the value like for example pag 5 yung value what would be the error just for the sake of you guys visualizing the error kasi kahagong last time when we were just explaining it according to linear regression it was a hard time to be honest even for me I had a hard time explaining it because I can just simply say na if it's constant it's independent of the concentration so meaning to say whatever the concentration is whatever the concentration is consistent yung difference consistent yung error nagsnyong ko so if you compare take for example if you compare Andaya to Miss Armenta if you compare if you compare the body organ sabi na natin parahas na nila ng body organ if you compare one from the other parahas lang yung pagkakaiba okay pag parahas lang yung pagkakaiba unlike your proportional error like what my example was kanina na as you increase as you actually measure higher concentration the chances of having a higher proportional error is also high kasi na the higher the concentration the higher din the difference nagsnyong ko dun sa part na yun nagsnyong ba sir d po hindi okay one more time so ang point ko lang dito okay so yung question kasi ni miss Andaya is is a very valid question but what I'm trying to say here so meaning to say hindi yung naintindihan yung in-explain ko kanina ng constant at proportional error is that correct miss Alba? sir slight ano na lang po explanation in what part? in what part? take me through your dun sa inisip mo saan banda magulong? sa my proportional error okay sa proportional parin talaga lord okay, okay let me give you another example okay sabi natin kanina ha okay I'll read the definition of proportional error first sabi niya dito that it exists when the difference between the test method yung bago and the comparative method the reference method are proportional to the analyte so what do we mean by proportional? if the okay eting ibig sabihin yan we can say na proportional okay when we say proportional as the concentration ay nakikita niyo bang screen ko we say proportional define natin what is proportional pag sinabi natin proportional parang ikauto at yung jawa mo or yung best friend mo na lang sabihin natin kung single ka kapag pumunta siya sa kanan pupunta ka sa kanan that is perfect or but when we say proportional in number kapag tumasya ng five points tatas ka din okay so never mind the value okay if you increase you will also increase okay if a increases b also increases are we clear? so what I'm trying to say here is that when concentration increases okay the error also increases anong example sir begin ko na lang kayo nang bagong example 10 20 30 40 okay take for example this is from the reference ito yung nag-measure ko na rin ang sugar yung reference ito yung nag-measure na sa unang patient, sa pangalawa, sa pangatlo sa pangapat nung ginamit naman natin yung test method ang nangyari sabihin na natin na okay let's stick with let's stick with 0.68 okay 0.68 o ganun ulit 10 times 0.68 mga kapatid 10 times 0.68 guys I'm only predicting ha I'm only predicting di ano in this particular part I'm only trying to predict ano yung value nung x so this is 0.68 this is x plus your zero zero tayo nagigets so if I do that 10 minus 68 that is 6.8 okay patulong 20 times 0.68 13.6 13.6 this is 20.4 40 times 0.68 kasi malapit na magtime I need to go to the other 27.2 so if we try to subtract that 10 minus 6.8 this is 3.2 13.6 minus 20 this is 6.4 this is 9.6 27 27 26.4 minus 40 ilan? 8 lang po it's 12.4 12.4 12.8 12.8 12.8 12.8 okay I want you to look now here I want you to look saan I want you to look here okay nakikita nakikita niyo to as the concentration increases habang tumataas yung concentration habang tumataas yung concentration tumataas din yung error nagiget sa ko yun yung proportional error ibig sabihin habang tumataas o habang dumadami yung minimesure mo lalo rin dumadami yung chances ng error are you getting me now am I clear yes sir thank you sir okay so ganun yung yun yung ibig sabihin natin sa proportional habang tumataas tumataas din yung error bakit tumataas yung difference nila okay so guys pag naringig ako sa exam difference error the same are we clear nagigets na po sir so proportional error still exists when y-intercept has a value but mas well defined po siya yes Jericho that's why okay in your book and ito siya actually in this particular area of your book sinabi dito okay yung mismong libro na yung pinapakita ako di ba sabihin nya dito that proportional error exists when the difference between the test the test between asan ako the method and the comparative method value are proportional to the analyte concentration the proportional error is present when the slope okay when the slope is one okay okay so if the slope is one are you getting me now yes siya so for me ha the reason why they have to say that the slope is if the slope is one then that goes to show na there is a proportional error okay kasi nga hindi di ba kanin ang nang compute tayo hindi nyo hindi evident si hindi evident si proportional error are we clear are we clear guys yes yes sir yes sir yes sir okay thank you thank you so moving on now okay let's try to end this okay I will try to to end this now so last two slides three minutes okay so when we are using parametric okay ito lang yung gusto kong tandaan nyo dito para mabiles if it is not normally distributed sir nasan po si not normally distributed when I say not normally distributed meaning to say na ayan not normally distributed not normally distributed meaning to say sabi natin kanina pag sinabi ni sir na not normally distributed it's also synonymous to skewed okay kapag nat normally distributed if it is skewed what you will be using our non parametric test are you getting me hello tapo okay tapos kapag naman normal normal distribution Gaussian distribution we are using your parametric test are we clear yes sir okay so actually that's it thank you so much for listening so I'll just entertain a few more questions I'll just end the recording yay natapos yung recording