 Dear student, today we are going to learn the factor loading. So factor loading is the correlation between the factors and the items. If it is greater correlation, so you can say that if it is 0.7 correlation, so you can say that there is a high correlation between the factor and the item. And if it is the 0.3 correlation, so you can say that we have the modified, moderate correlation between the factor and the item. This is the concept of the factor loading. So factor loading, basically, we have to check the correlation between the factors and the items. The factor analysis model assumes that there are K underline factors. So K underline factor means that the total variable we have, we will make K factors. Where K less than equals to P and P stands for the dimension or you can say that this is the variable which we denote by F1, F2 up to so on Fk. So we have made K factors total. In the variable step, we have P variables and the K factor is less than the variables. And that each observed variable is a linear function of these factors together with the residue variable. Basically, what we have done is we have developed a model, we have generated a model. What we have in the model, these are the Xi, this is the model Xi, this is the random variable which associated with the lambda I1, F1 means factor 1, lambda I2F2, factor 2 up to so on lambda IKFK, factor K, total 5th K factor is what we have plus EI. So EI denoted by the residual variable or you have residual means error. This is called the equation number one. In the above equation, in this equation, we have the weights which lambda Ij are usually called the factor loadings. This is basically you have factor loadings, with every factor, we have its loadings attached. Means with that loadings, if we say one factor solution, then what is one factor? X1, lambda, F1 plus E1, we have one factor solution, that is we have just one factor. And this lambda which is equals to the factor loadings or factor loadings, there is the correlation between the variable and the factor, this is called the factor loading. So that the lambda IK is the loading of ith variable on the kth factor and FI here is the FI is the common factor and the variation between the I which is equals to the 1, 2 up to so on P. So we have the P variables, the variants EI describe the residual variation specific factors. So basically, we have factor analysis, we have factor loadings, factors and EI. So lambda stands for factor loadings, F stands for common factor and EI stands for specific factors. This is the model. Now the factor loadings is a statistical measure that indicates the strength and direction of the relationship between an observed variable. Basically I have explained to you what are the factor loadings. Basically factor loading is the correlation between the factors and the variables. We call them variables, we call them items, basically what are the correlations. If it is positive correlation, so direction or strength, if it is high correlation, so there is a strong correlation between the factors and the items or strong correlation or high correlation point 7 above it, so we have strong correlation or more than point 3 to point 7 as that is called the moderate correlation and less than point 3 it means the low correlation or low correlation or you can say that the mis strength or strength of that is low. So the relationship between the observed variable, observed variable, basically what are the variables that we are studying, such as a survey question. Basically what are we doing in survey question, what are the observed variables that we have, we are doing psychological test, we are seeing sadness, mood disorder and fatigue. Now what are the observed variables that we have, we are calling them observed variables. And the test score and the latent factor identify through factor analysis. Now what are the variables that we have, we have to check these variables, what are the factors that we have to check, basically we have 10 factors, we have 10 variables and how many factors we can make, so for that we have all the procedures, for that we basically need strong factor loading, what are the factor loadings, we will find out its correlations. It may be how much of the variation in the observed variable is explained by the factor, how much of the observed variable is explained by the factor, if it is 68% variation of the observed variable, so we can say that 68% variation is explaining these factors. And 32% variation is explaining other factors, so basically what we have is the measure how much of the variation in the observed variable is explained by the factor. In factor analysis each observed variable is assigned as a loading score for each factor. Now loading score, what do you have? As I told you previously, lambda, we have a loading, this is called the factor loading, factor loading score has a numerical value, maybe its numerical value is 0.6. So there is a strong or moderate correlation between the factor and the item. So the factor score we have is the numerical value. The loading score represents the correlation between the observed variable and the factor, I have explained that the loading score value will tell us how much of the correlation between the factor and the variable. The higher the loading score, the stronger the relationship between the variable and the factor. So basically what we have is the range, the correlation range you know is minus 1 to 1. The higher value we have is the factor score or the loading, so we can say that the stronger relationship between the factor and the variable. So here is the factor loading can range from minus 1 to 1 because q minus 1 to 1 because we are measuring the correlation, so the correlation range is minus 1 to 1. A loading score of 1, 1 indicates a perfect relation between the variable and the factor. 1 is very rare that we have 1, but if we have the maximum value then we have 1. If we obtain 1, then that is called the perfect relation between the factor and the variable. While a score of 0 value, if it is 0 then we can say that no relation between the factor and the variable. Negative loading score indicates an inverse relation between the factor and the variable. Negative we have, it means that we have a negative relation. It means that we have used the word not in the question. So if you see it, I will explain it to you later. When we are asking a question directly and if we have not existed, not good, modus not good or fatigue is always we are talking in this way. When we talk in this way, it means that you have used such words in the question which are not equal to not. So there you have this score of negative. And finally the researcher use factor loading to interpret the meaning of the factors identified through factor analysis. So basically what we have is factor loadings because factor analysis is a very important part. So factor loadings basically we will tell you what you have to interpret or its correlation between the factor and the variables. High loading score suggests the observed variable is a good indicator. Observe that you have selected or selected the variable, its value is existing that means its higher score. Observe the high loading score suggests that the observed variable is a good indicator of the underlying factor. That means that the factors you have made in this variable are involved in this factor and in this factor we have converted. So while low loading score suggests that the variable may not be a good indicator of the factor. That means that low loading is coming. It means that there is no chance of it or if we ignore it or neglect it, then it will not have any effect on our further analysis. If we are doing K factor solution. One factor, two factor solution is different. If we have K factor solution then we have some variables that we have to ignore or remove from it. Those factor scores are minimal so we have no effect on them as such. But we have to see its interpretation. What is the interpretation? Low loading score suggests that the variable may not be a good indicator of the factor. This is the basic concept of the factor loading.