 Before I start the second lecture, I will like to share with you some of my personal experiences in some of the aspects that he mentions about communication in general. But more importantly the fact is that he while he is talking about this in the context of research, you will all agree that all of it is important for every teacher who teaches any subject. I realized pretty early in life, when I joined as a teacher that we are not taught at all how to teach. It is presumed that if you are appointed as a lecturer or a student professor, you are able to teach. If you have done a postgraduate degree or PhD, you are able to teach. This is nonsense. The only people who are taught how to teach are actually school teachers who are required to do BA ed and something like. We do not have BA ed means Bachelor of Engineering Education. But you will agree that something of that sort is essential because something which you are going to do for 40 years should all of us be compelled to discover good teaching methods only as we go along. Meaning that first 3, 4, 5 years of our life what we teach is perhaps not as effective as it could have been. So, this is something which is now getting corrected by AICT. Remember I mentioned that research methodologies is one of the core courses that is being made compulsory across all postgraduate programs in engineering. Communication and ethics is another core course which is being made compulsory. Sadly, however the communication course that exists even in IIT Bombay has not been very effective so far because it is traditionally taught by the humanities department and communication is not a course which you can study while listening to lectures. It involves lot of practice whether it is oral or written communication and so on. I will tell you my own experience when I became a teacher and I felt that I must improve my teaching. I did 2 things which helped me tremendously. I am talking about 1973, 74. I purchased a tape recorder. In fact, I remember a big fight with my wife because the tape recorder that time costed 750 rupees and my salary was a grand 350 rupees per month. So, I would have wives of course would love saris or some such useful things in life. In fact, first she thought I have brought it for her to listen to music and then she found out that in the night I was standing in before the mirror and was giving a lecture to myself. In fact, now I realize that my neighbors must have called me a mad fellow that type as he mentioned. I did not realize but I tell you when I replayed that lecture I thought that I was quite fluent but when I replayed that lecture I noticed several who break wrong choice of words etcetera. I did that consistently almost for a month. It helped me to improve my skills. The second thing I did is I selected some colleague teachers each one of whom was known for a specific attribute. For example professor Lemay used to be famous for his time management. So, he would start with a lecture in the morning as a maths professor some theorem let us say at 8.30 sharp and at 9.25 sharp he would have written QED. Nothing artificial about it. The lecture is free flow there is lot of interaction. There is perfect time management. There is a senior teacher of ours Professor Bedford who was well known for extremely decent handling of the participating students. So, I had realized it when I was a student and he had taught us one course and there was one of our colleagues in the M Tech program here who was asking the same question again and again and when he asked the third time I thought the professor should blast him for being a total idiot. But the way professor Bedford very patiently replied you have a point. However, if you once again look at that point from this point of when he spent three minutes laboriously explaining. I did not react at that time, but later on when I became a teacher and I was attending his class primarily to understand how he can maintain that patience. And then one of the evenings I asked him, sir do not you get angry do not you get frustrated when people ask such stupid questions. He told me something which is very vital I will never forget in my life. He says Fata you and I know that this question is stupid question, but please understand that he is not stupid enough to ask that question had he known that it is a stupid question. The question is genuinely in the minds of that person that is why that person is saying this question. Now if I brush him aside saying that it is a stupid question he will probably never ask a question in his life and which means he would curtail his own curiosity and the learning will stop. A very important lesson for a teacher. So, these are the things which although we will be speaking in the context of research methodologies you will agree that it should benefit all ten thousand teachers in terms of improving their teaching as well. He talked about humor. So, let me end this monologue by my own experience of humor. When I was studying my bachelor's in J. S. I. T. S. in though this is 65 I think 65, 1966, 1965, 1966. There was a very renowned professor of Mathematics Professor Paria who used to teach us maths. In spite of being extremely knowledgeable his lectures were completely drab. He would write down theorem and prove and very systematic voice, but completely drab. The lectures used to be boring. One day he came he wrote the theorem there and then suddenly stopped and they turn to us and says, now I will tell you a joke. So, suddenly there was a furrow because Professor Paria telling us a joke and then he says in a very monotonous voice exactly how he describes the theorem he says a table has four legs a cow has four legs therefore table is a cow. And then he turned back now we will solve this theorem. That is the funniest kind of humor we had heard. We did not know whether to laugh or whatever, whatever, whatever. Three or four days later we had another professor Kaili of English who used to help us prepare for our debates intercollegiate debates and I was in the debating team and I was reciting this very humorous incidence to some other colleague members of the team and Professor Kaili was listening to this. And then he said oh Paria did this. We all looked at him. He says you know Professor Paria came to me last week and asked me my lectures are not getting interesting. So, what should I do? And I told him tell a few jokes. So, you see the notion of humor itself has to be understood and imbibed properly. It can be done. Believe me I mean that professor was so around and he was a great maths professor and there is absolutely no lack of any intellectual capability to understand it. But it simply so happens that he never in his life gave importance to these aspects. He thought that he is dealing with maths and that is good. So, anyway life is very interesting and we should attempt to make it more interesting to our students who all the teachers. There are a total of 12 lectures which we have given here. Each one of them is actually important. Unfortunately, we may not have time in the mega workshop to conduct all the lectures. By the way all these lectures even in those days when I was running the Eklavya project like we do now in these workshops all the lectures, video, audio, recorded lectures, slides everything is released in open source. Even these are available on the website. However, it is not well known that they are available. So, what I propose to do is I will let him choose which lectures he actually wants to deliver physically. And in addition to those recorded lectures we will also tell people that these other lectures are also available. You can perhaps listen to them. Let me play one of the experimental skills. Now let us begin our discussion of the experimental skills. Here we will cover the following points shown on the slide. First we will illustrate what is the meaning of the scientific method with an example. Then we will see variables and errors in experiments. Then what is meant by control, precision and accuracy. And finally we will mention in passing how to fit a straight line to measure data. Let us look at the scientific method. What does it involve? One can broadly identify four steps in this method. First is the observation of phenomena. Then based on the observation and our thinking we come up with a hypothesis about some happening in the world around us. The hypothesis is some sort of an imaginative explanation for things which we observe. However it is not verified to be correct. So hypothesis is a sort of statement of explanation for the phenomena which is not yet verified to be correct. So it is mostly based on intuition. Then comes the important step of verification of the hypothesis. So here broadly there are three steps experiment, observation and inference. If one wants one can club this entire thing, step of verification all the three steps of experiment, observation and inference in just one word that is experiment. And after the experiment and the inference drawn from it you make a generalization to all situations which may be similar to that considered in the experiment. Now let us look at the various elements of this method in little bit more detail. See observation. Observation is not a passive acquisition of sensory information but a critical purposeful process needing a high level of awareness. This high level of awareness is very very important for being a good observer. Because when you say that you should be observant, what does it mean? What is it that you are supposed to observe? So when you ask a student if the student is doing some chemical reaction and then you tell him you be observant and see how variables in the reaction are changing. The point is you are not telling him what are all the variables that the student should observe. It is not possible to tell also because after all you are doing a new experiment it is your duty to be aware as much as possible to identify as many variables which may be affecting the experiment. So it is in this sense a very high level of awareness is what is expected if you want to be a good observer. Because you cannot tell beforehand what are the things that you should observe. So for example you are doing a chemical reaction and you are concentrating on measuring the temperature. But at the same time the reaction is being done in an environment where there is light because otherwise you cannot see what is going on. It is possible that lighting may have an effect. So like this it all depends on a person how much aware he is. Depends on the awareness whether you are able to observe the presence of a variable which is affecting the reaction. So there is a reason why we choose to observe a specific portion of the whole domain of possible objects of observation. So it is not as though something that is very evident to your eye or something that can be heard very clearly alone is chosen for observation. You are aware of several things that are happening around you but then even though you are aware of all the things you choose only a part of those things that you are aware of for critical inquiry and careful observation. Hypothesis. Hypothesis is an imaginative preconception or an inspired guess about some particularly interesting aspect of the world. Every discovery begins as a hypothesis. Now hypothesis is based on intuition it may or may not be correct. We will illustrate this with examples. What is experiment? Experiment is an act undertaken to verify a hypothesis. It discriminates between possibilities and gives direction for further thought. This is another role of experiment. It discriminates between possibilities. It is a happening device to apprehend the truth. So it is a set of conditions that you ensure in order to observe a particular thing. So for example if you want to detect the wave nature of matter then you have to devise conditions, special conditions so that this wave nature will become evident. If you want to detect the particle nature of light you have to devise special conditions to detect that particular aspect. For knowing the wave nature of light you may not have to devise special conditions because already the conditions which are existing in the environment may be sufficient for you to observe easily the wave nature of light. So here you are not devising special conditions. It is a method of discovering causal interconnections. So what is experiment? Another way of looking at experiment is it is a method of discovering causal interconnections. Hypothesis. Devising a hypothesis that can be tested by practical experiments is an art. So every hypothesis may not be possible to test every hypothesis in a scientific way or to establish whether it is true or not. For example hypothesis all men are mortal. It is not something that you can test. It is not possible to scientifically examine this hypothesis. All men are mortal. Then how do you frame a hypothesis? That is also important. So there is a method of framing the hypothesis. It is called null hypothesis. That is hypothesis stated in negative terms. I will leave it as an assignment to you to see this point in detail. What is the meaning of a null hypothesis or framing a hypothesis in negative terms? In a very simple way I could put it as follows. Suppose you make a hypothesis, some conditions a result in effect b. So a implies b. That is your hypothesis suppose. Now an equivalent statement of the same hypothesis would be if there is no b there is no a. a implies b is equivalent to saying b bar implies a bar. b bar means not b. Supposing you are making a hypothesis some variable a results in an effect b. It is same as saying if you do not see the effect b then it means a is not present. Not b implies not a or no b implies no a. Now this statement and any of these statements are equivalent. But when you frame the statement in this form not b implies not a. No b implies no a. That is if you do not find the effect b then you can be sure that the condition a is absent. This form of framing the hypothesis is called a null hypothesis. Normally people say that to test a hypothesis it is better if you test its null form of a hypothesis. So more about this I leave it as an assignment because framing of hypothesis itself is a important and there are several details about this. Now we will consider an example to illustrate this particular aspect of the scientific method. West one observes a high level of prevalence of obesity. So there is a large number of people who are obese. In fact about 50% of the population is overweight. This is an observation. So naturally people are concerned about it. So some scientist working on the issue of obesity and removal of obesity. They hypothesized that the overweight people are overweight because they have a higher level of hunger than the average weight people. Because they have a higher level of hunger they eat more and therefore they become overweight. So the fact that people are overweight is one observation. The fact that overweight people are seen to eat more is also an observation and then you are trying to frame a hypothesis by intuition. What is the cause of overeating? What is the cause of overeating? So the cause of overeating is more hunger. Now some scientist came up and said look I do not accept this hypothesis that the hunger, the higher level of hunger is the cause of overeating. I want to test whether it is true. So he suggested an experiment or rather he carried out an experiment. Now his experiment was as follows. He took 100 overweight people and 100 average weight people. This group of 200 people was divided into two groups of 100 people each. Each group of 100 people had 50 overweight people and 50 average weight people. Now each of the two groups were given sandwiches to eat. So one group of 100 was put in one room and each person was given one sandwich to eat. The other group of 100 people was put in another room and they were given three sandwiches to eat. People in both the rooms were told that if they do not want they need not eat or if one sandwich is not sufficient to satisfy their hunger they can go and fetch another or more sandwiches from room which was kept about 50 meters away. So they can bring sandwiches from there and eat. If one sandwich is not sufficient. Now after carrying out this experiment they measured what is the average number of sandwiches eaten by the overweight people and the average weight people. Now very interestingly they found that on average the average weight people ate only one sandwich in both the rooms. So in the room where I am sorry I correct myself the average weight people ate two sandwiches in both the rooms which means in room number one where only one sandwich was provided most of them fetched one more sandwich. In the second room they left out one sandwich they did not eat all the three sandwiches. However overweight people the behavior was different. In room number one most of them ate only one sandwich whereas in room number two most of them ate all the three sandwiches. Now this experiment showed that internal factors such as hunger may not explain this observation because in this case I mean it depends on the experimental conditions sufficient care was taken to see that everyone was sufficiently hungry and so on. There are many other details of the experiment so which we will not go into here. We will assume that conditions were maintained identical in both rooms. But what it showed is that for the overweight people the amount they eat depends on the availability of food which is an external condition not an internal condition to the person. So the eating habits of overweight people are stimulated by external conditions in the environment more specifically availability of food. For the average weight people however there is an internal control mechanism independent of the amount of food available they will eat right some amount that is controlled. So this experiment therefore showed that you cannot assume hunger as the cause or higher level of hunger as a cause for over eating. Of course there have been further experiments and so on whatever I have said is a simplified form of the situation. But nevertheless it illustrates the importance of scientific approach. So what is the scientific approach? Scientific approach is you must verify the hypothesis nothing is accepted until you test and you must devise an experiment. So that is also important in science how do you think of an experiment to test your hypothesis and testing of the hypothesis does not mean that you should prove it even if you disprove a statement right even if you disprove a statement it is significant. Now the lack of scientific method however or scientific approach is very prevalent in society. Now before I give an example of a non-scientific method right I want to emphasize the fact that it is possible that not every hypothesis can be tested we have already said this okay. So there are many things which you may not be able to test by the scientific method devise an experiment and test and you cannot say that what is not possible to test by scientific approach is not true that is not the kind of thing the statement that we are making. But for our purpose of research right in a wider philosophical sense it may be true that everything cannot the truth all aspects of truth cannot be scientifically tested it may be true here we are only talking about research in science and engineering or humanities even humanities where you can test things scientifically. But when you say testing scientifically you should be clear what it means. So what it means is in a scientific method two important elements are there a hypothesis and an experiment. So both things are important an example of not using verification it happens very often in life but a significant example of this is the belief of I forget whether it was Aristotle or Socrates okay one of these two. So this philosopher believed that women had less teeth than men. Now this is a hypothesis it was based on intuition and once since the philosopher was well established and well known people also believe this statement and this happened for several hundred years. Now no one attempted to ask a woman to open the mouth and simply test check how many teeth she has this is very interesting why because the person who made the statement was very well established and regarded as a good thinker. So we must be aware that very often we might use non-scientific approach and we may have beliefs which are not tested we must always try to test the beliefs as far as possible okay. So this is an example of thinking based on only intuition very often all our hypothesis are based on intuition without intuition you cannot work. So we are not saying that you know intuition is bad intuition is very very important in fact only intuition can give you hypothesis but intuition is not sufficient in scientific working in working by scientific method intuition alone is not sufficient you must verify the verification or experiment is very very important. So now let us look at some other aspects of experiment conducting experiments types of variables independent variable stimulus or input or causes all these are synonymous terms. So this is the condition manipulated to determine its effect on an observed phenomenon then some variables can be dependent variables. So dependent variables or response or output or effect so this is the condition that appears disappears or varies as an independent variable is introduced removed or varied very important point in some cases it may be difficult to decide which of a pair of variables is the cause and which is the effect. So just because you see there is a relationship between two variables two variables are correlated does not mean you will know which one is the independent variable and which is the dependent variable then apart from independent and dependent variables you also have another type of variables they are called confounded variables confounded variable is one whose effect cannot be separated from the supposed independent variable. Now we will we can illustrate this with example suppose you are trying to find out if men and women have any specific preference for colors you want to now this is a hypothesis that you want to test whether there is a specific preference. Now you think you take a group of men and group of women and then you do some experiments with them okay and you find some difference and you try to attribute that difference in preferences to the gender but what you are not taking care of is the fact that this women men and women are they have a particular age. So over the years they might have been exposed to particular type of colors this is called past experience you do not know supposing because people are exposed to a particular type of color they like it they may develop a liking for it now this past experience is something it is a confounding confounded variable because it is effect you cannot separate out very easily. Therefore if you can devise an experiment in which you can separate out the effect of the confounding variables or confounded variables then that is something significant. So one must always be aware that there can be dependent variables there can be independent variables and there can be variables whose effect cannot be separated from that of other independent variables and confounded variables. So your experiment should not have confounded variables or you should have some method of eliminating the effects or at least if you cannot eliminate you can at least state in your experiment that this is a confounded variable whose effect we have not been able to detect. Some other attributes of variables quantitative variable it is something that varies in amount you can assign a number for it categorical variable it varies in kind gender is a categorical variable right male or female continuous and discrete variables. So you can have variations of quantities in discrete manner or continuous manner now we have discussed the various types of variable in experiment let us look at another very important issue any experiment you are making a measurement without a measurement there is no experiment I have emphasized this fact in my introduction that measurement is very very important just qualitative understanding of how when one variable is increased the other variable increases or decreases is not sufficient that is that does not complete the experiment you must measure. So that is where errors come in so two types of errors can be there in measurement random errors. So random error varies and equally likely to be positive or negative it is an error which is equally likely to be positive or negative as against this a systematic error is something which is constant within course throughout the experiment the word constant is to be not to be taken literally in fact if you want you can say non random errors so random error is one which varies equally likely which is equally likely to be positive or negative that is nature of random error but systematic error it has a certain trend predictable trend traditionally the word systematic error has been used therefore we will retain the error retain this word but you can understand this as a non random error and now let us illustrate with example supposing you are measuring the terminal velocity of a ball in viscous medium like glycerin suppose this is the experiment you are doing now let me draw a diagram to show that so this experiment all of us who are in engineering or physics would have done in our first year now this is a jar or a tube and it is filled with glycerin which is a viscous medium and you drop a ball in this medium now the ball initially accelerates downwards because of gravity but ultimately it acquires a constant velocity because as its velocity increases the viscosity produces a frictional force which increases with velocity so for some velocity the downward force because of gravity is exactly balanced by the frictional force or opposing force because of the viscosity of glycerin if you were to drop the ball in vacuum then it would accelerate continuously it will not reach a constant velocity experiment is to measure the terminal velocity of this ball terminal velocity is a constant velocity that it acquires after some time now how will you measure it the common method of measuring is you note for example when it crosses this particular mark with a stop clock so you start a clock then when the ball crosses a different mark you stop so the clock would have come somewhere here some time has elapsed now you measure this distance and measure the time and then the terminal velocity is given by distance by time so this experiment so in this experiment what are the sources of error and how do you classify the errors as random and systematic as shown in this slide random errors can occur in starting and stopping the clock or in estimating the ball location on a scale okay so let us look at the diagram again now you say you are noting the time when the ball crosses this mark now are you noting when the bottom end of the ball is crossing the mark or when the top end of the ball is crossing or is it that you are noting the middle this is one point and it may not be possible for you to precisely note any one of these so there can be some error because you are repeatedly when you do the experiment repeatedly right this something that you are not able to determine precisely and it can be random then starting and stopping of the clock you are looking at the ball and at the same time you have to start the clock so there can be a delay between the two events so you may start the clock before the ball has crossed the particular mark or you may do it little bit later and this thing cannot be controlled so these are the two sources of random error in this experiment these are examples of two sources of error which are random now what is the systematic error here that can happen supposing your clock after you start progressively become slow this is an example of a systematic error or when you repeat the experiment what happens is the there is a friction between the ball and the glycerin now this friction causes the glycerin to heat up after all how is the energy dissipated how is the friction the energy dissipated because of friction it is dissipated as heat so the temperature of glycerin will go on rising if you take a very long time to complete your experiment so this is another example of systematic error so like this in any experiment you have either systematic errors or random errors now you should have the ability to distinguish between random and systematic errors and you should also be aware of the fact that these errors are present and what are the sources of these errors because only then you can reduce the errors it is very very important to reduce the errors we will concern example to illustrate why reducing errors is very very important so before we go to that particular example let us see some further facts about errors we normally use the terms accuracy and precision so what is the meaning of accuracy now on the left hand side here you see a point or a circle which can be regarded as a true value of the particular variable being measured and the crosses around it indicate the values obtained in repeated experiments of the same to give you an analogy in fact you can regard I imagine that this is this represents a shooting event where the solid circle is a target and the crosses indicate the places where the bullet strikes okay when a person is trying to shoot at the target in repeated trials now you will say that the marksman is accurate if all these places where they bullet hits that somewhere near the target are very close to the target all of them should be very close so what it means is both systematic and random errors are small so random errors are represented by the differences between the repeated measurements and systematic error is represented by the difference between location of these crosses and the target so here for example when all these crosses are very close to each other but they are located away from the target then you say that the measurement is precise but inaccurate so precision has to do with ability to repeat the value in successive measurements several trials so you may be getting the same value in repeated measurements that doesn't mean it is accurate it means it is precise so accuracy means these values should be very close to the true value so systematic error should not be there precision means random errors are minimum okay but that doesn't mean systematic errors are absent so imprecise and inaccurate both systematic and random errors are present so this is an important point that we should note the difference between precision and accuracy so precision means small random errors accuracy means both systematic and random error should be small now when you write the reading let us see what are we pointing towards the systematic error or the random error now here for example on this slide you show the measurement of a resistance in some experiment and you write the measurement as the measured value as one particular number plus or minus another number so in this case does plus or minus 0.001 ohm represent a systematic or random error what does it represent evidently it is random error because it is plus or minus so it is equally likely to be positive or negative so it is not systematic error so note that in when you report your measurements in this form you are specifying the random error you are not specifying the systematic error now how do you detect and minimize errors this is the next issue random errors can be detected by repeating the experiment so when you repeat the experiment if you are getting different values then you become aware that there are random errors in the experiment however repeated measurements may or may not reveal systematic error this is very important you cannot reduce systematic error merely by repeating the experiment systematic errors are somewhat more difficult to detect and reduce random errors can always be eliminated by repeating the experiment so many times the research scholars are not aware that when they are minimizing error by repeating the experiment they are only removing the random error or let me not use the word remove they are reducing the random error you cannot remove systematic errors by this approach so first of all many times in fact the research scholars are not aware of the errors at all now how do you know whether a research college is aware of errors or not it you can easily see the way the readings are presented in graphical form measurement okay for example if you find that some measurements are reported like this so I show points these are measured points the measurement of y as a function of x and then you say the behavior is something like this now clearly the person who has drawn the diagram is not aware of errors because you assume that the measured value is precisely this point right whereas just now we have shown on the slide that supposing you are reporting measurement of resistance you will report it as a value that you have read on the meter plus or minus a quantity plus or minus a quantity that is a random error now I will not discuss in detail how do you decide the random error in a situation right you can look at it I am just initiating the whole discussion about the errors so how do you know the research scholar is aware of the errors well if it shows an error bar if it shows the reading as a range and then this value is the midpoint of this range then you know that yes the research scholars aware of the errors now why is this bar very important because now you see if I take lowest point here the highest point here because this is also possible and then the lowest point here then I do not seem to discern any specific trend okay but then you may say this will happen all the time the point is what is this what is the length of this bar this is what is important if the length of this particular bar is small then you can definitely discern some trend but if the length of this bar is long which is what we have seen in our lectures yesterday when we discussed about graphical approach and so on okay how to represent data in graphical form then you may not be able to discern trend any trend in such a case if you are reporting that you have discerned some trend then your conclusion can be in doubt now how to minimize random error you do it by averaging now how does averaging reduce random error let us look at this point also now before we do that let us look at the aspect of control in experiment when do you say an experiment is well controlled so control implies awareness of all the variables in an experiment and the ability to vary them at will so you should be aware of all the variables in the experiment and you should be able to vary all the variables the way you want this enables one to restrain sources of variability in research control also means a standard against which the effect of a particular variable can be compared so you are comparing against a standard the measurement against a standard the effects of all but the independent variable can be eliminated okay effects of all variables except for the independent variable now as we have mentioned in one of our earlier lectures the several independent variables can be there and then you will have one effect now how do you find out what is the relation between the different independent variables and the dependent variable so when you want to determine this you must have one independent variable being changed and monitor the effect and at that time the effects of other independent variables should not be there this is one important thing that you should ensure in your experiment this is what is the meaning of control another meaning of control now how do you can how can you ensure that so there are several ways of ensuring depends on the situation so the effects of all but the independent variable can be eliminated by either removing them if you cannot remove them by maintaining them constant or by screening them there is a method called screening you can screen the effect or by counterbalancing counterbalancing means you cancel the effect of one variable with the other variable and there is another approach that is called systematic randomization now all these various methods are very very important I do not have a time to go through them you can note down these words such as screening counterbalancing and systematic randomization and you can try to find out yourself this is an assignment because removing them one can understand maintaining a variable constant this is also clear screening what is the meaning of screening it is clear but how do you do screening right please find out how do you do counterbalancing some practical examples and how do you do systematic randomization we will see some example of how you can remove the effect when you cannot how you can remove the effect of the variable other than independent variable when you cannot remove the variable or you cannot maintain it constant one example we will see now how can there be difficulty in control let us take some examples to illustrate how difficulties arise in control of experiments concern experiment related to magnetostriction so what is magnetostriction the experiment for measuring magnetostriction is shown on this paper slide so magnetostriction means change in the length of the material because of magnetism change in the length of the material because of magnetism that is magnetostriction you know that when you heat some material it will elongate so that is change in length because of heat like that you can have change in length because of magnetism now supposing you are doing an experiment to detect whether a particular material has this property of magnetostriction so a schematic diagram of the experiment would be like this so you have a solenoid to create the magnetic field and inside this solenoid inside this coil sorry you have a coil to create a magnetic field and then inside that you put the material for which you want to measure the magnetostriction now when you switch on the current magnetic field is created and then you make the measurement of the length from this end to this end and you indeed find for example suppose the length has increased supposing in a particular situation you found the increase in length was say 0.005% increase in length because of magnetism now you are happy that you know yes you have detected magnetostriction in the material but unfortunately what you have failed to notice is that when you pass current through the coil the coil gets heated up and because the coil is heated up the material is heated up and because the material is heated up it will expand now are you sure that the expansion that has happened that you have detected or measured is solely because of the magnetism and not because of the temperature so what you have to do you have to measure the coefficient of expansion because of thermal reasons of the material separately and then you should try to see whether the expansion that has occurred here is more than that or you will have to find some method of measuring keeping the temperature constant but this is an example where you can see that when you want to increase the magnetic field the temperature increases okay so these are the kind of difficulties which are there in control another example terminal velocity of a body in viscous liquid like glycerin we have already discussed the apparatus by which you measure the terminal velocity now as we have said the temperature of the glycerin can go on increasing as you increase the number of trials of the experiment so now if you are measuring the terminal velocity as a function of for example diameter of the ball you are doing an experiment to measure terminal velocity as a function of diameter suppose this experiment then you have balls of different diameter and you are going on doing experiment with the different balls and you think that you have seen some trend as the diameter increases terminal velocity increases or decreases whatever you have seen some trend and you think that is because of the viscosity of the medium but the temperature of the medium has also changed so how do you know that the terminal velocity behavior that you have measured is not because of temperature right and because of viscosity alone so here again there is a problem in control then the third example is something that I already mentioned gender differences in color preferences gender is confounded with past experience so past experience confounding variable now in the previous examples the effect of temperature is the confounding variable okay in magnitude restriction experiment and terminal velocity experiment now why should one be bothered about precision why should one try to be as precise as possible so when experiments give results at variance with theoretical ideas new discoveries are made right this is the main reason why you must be bothered about precision let us consider two examples where precision played a role in great discoveries discovery of argon by Rayleigh and Ramsey this is first example now how was argon discovered the people knew that air contains nitrogen oxygen and traces of water vapor carbon dioxide and so on this is what people had known before argon was discovered now someone did an experiment to see the density of nitrogen that you get in two different to separate methods of getting nitrogen so one approach was you remove the oxygen from air and you try to remove the moisture and so on also since you are not aware of any presence of any other gas so you assume that when you remove oxygen and traces of moisture and so on what is left with nitrogen so you measure the density of the remaining gas let us put it like this now you think it is nitrogen that is your hypothesis another experiment was done where nitrogen was produced from ammonia that approach produces pure nitrogen you are very sure it produces nitrogen it was found that the density of the air that is left when you remove oxygen and moisture and so on was very marginally higher than the density of nitrogen that was produced from ammonia now marginally means how much it was just half percent higher now many times you may say this half percent kind of error can occur in your measurement anyway why bother about it but then when they did the experiment very precisely where this half percent also become significant then and they investigated the cause of this discrepancy they found that you can explain this discrepancy by assuming that there is one percent argon another gas of a particular weight and so on which they called it as argon so one percent of argon was you know detected and that is how the argon was discovered so you can see how precision is very important another example of discovery because of very high precision in experiment discovery of deuterium by Birge and Menzel 1931 and Ure Brickwede and Murphy 1932 how did this discovery occur so people produced hydrogen in two different ways and tried to measure the mass you call it as atomic mass one method was chemical determination as shown on this slide the chemical determination method it gave the ratio of the mass of hydrogen that was obtained by this method to mass of oxygen with atomic mass of 16 particular isotope now this measurement was reported as shown here that is 1.00799 plus or minus 0.00002 in a different experiment the mass of hydrogen ratio of similar ratio mass of hydrogen to mass of O16 was found to be 1.00778 plus or minus 0.00005 now 1.00799 plus or minus 0.00002 what this means is your numerator can be anywhere between 1.00797 to 1.007801 numerator here however the numerator can be 1.00773 to 1.00783 now you see that this range is definitely lower than this range even when you include the errors so differently here the mass of hydrogen was different in this particular experiment than in the chemical determination method now once this difference was detected and then it was hypothesized that you can explain this difference if you assume a very small percentage of hydrogen with atomic mass of 2 instead of 1 that is deuterium so if you assume one part of hydrogen atom with atomic mass of 2 you can explain this discrepancy and that is how with further experiments it was shown that yes this is indeed correct spectrum of hydrogen was taken in a very sensitive experiment and then it was found that you did have lines corresponding to a hydrogen atom of having a mass of 2 units so it is another example where precision is very important so if your this plus or minus value was much larger you would not have detected any significant difference in the two cases so this where it was very very important to reduce this these numbers which means you want your measurements to be made very precise so this is how precision is very very important in our experiments so we can reemphasize what we have said just now with another example significance of results depend is on precision if for example the measured resistance at two different temperatures is 2.0025 and 2.0034 we can say that temperature has an effect on the resistance can we say this yes if the error is plus or minus 0.001 ohm no if the error is plus or minus 0.01 ohm why because you can see that if the error is plus or minus 0.001 ohm then your values of resistance at 10 degree centigrade can be 2.00.024 2200.026 and at 20 degree centigrade it would be 200.033 to 200.035 now clearly the higher value here the lower value here is significantly more than the upper bound here so clearly at 20 degree centigrade the resistance is different than at 10 degree centigrade so you can say yes temperature has an effect but supposing your error was 0.01 ohm then your values would be 10 degree centigrade 200.015 to 200.035 this is one range that is possible at 10 degree centigrade at 20 degree centigrade it is 200.024 to 200.044 your measurement could have been anywhere in this range now you can see that 200.024 the lower end lies in between 200.015 and 200.035 so supposing these were the actual values then you find that as a function of temperature the resistance is decreasing on the other hand if these were the values it is increasing so you cannot discern any particular trend from here so this is how precision is very very important so if you plot the same thing on a graph I have now shown it in terms of numbers I will leave it as an assignment that the same illustration can be much clearer if you were to plot the things on a graph your graph would look something like this when the error is small your two readings at 10 degree centigrade and 20 degree centigrade this resistance this is temperature they would be like this so the maximum value here is less than the minimum value there so you can clearly see yes there is an increase this is when the error is 0.001 if the error is 0.01 what would happen is on a graph it would appear like this 10 degree centigrade 20 degree centigrade this is how it would be so now you cannot say you may it may also be that the resistance is not changing at all or it may be that it is decreasing or it may be increasing but in this case it is very clear so this is how this actually represent this length represents the precision or the extent of random errors how much precision now we have seen that precision is important now how much precision we should aim at it depends on the purpose of the experiment if a resistance is to be used as a standard in the range 10 to 20 degree centigrade and the precision required is 0.01% an error of 0.01 plus or minus ohm is adequate and it is unnecessary to reduce the error to plus or minus 0.001 okay so in your experiment you should have an idea of how much precision is required so this is another skill that you should have now each of these points can be discussed in great detail with further examples so you can have several hours of experimental skill right courses on experimental skill so here I am by just pointing out the important issues you can read up on your own you know what kind of research you are doing the goal of this discussion is to make you aware of the various issues you can you know read up the required material and then understand these issues in depth