Introduction to MANOVA, MANOVA vs ANOVA n MANOVA using R

in this tutorial we learnt about manova you will see málaga vs. anova side by side you will see huge cases of manova where do you need to apply manova you will also see null hypothesis and execution steps of manova and you will see a workout demo using our formula 1 let's first start with an introduction of manova in ANOVA you have one dependent variable and one or more independent variable which are called factors like you have two factors and probably one dependent variable the big change that happens in case of maloha that you have two dependent variable if you see the difference here is that there is another dependent variable that has come here let me relate to an example in case of ANOVA probably you have these are three different teaching ads and this is the dependent variable score in case of manova you have again d3 teaching ax but you have two dependent variable score and probably orientation score for continuation of study so does it like only help in getting score or does it also keeps the motivation high that students will like to continue the study that's what is the another dependent variable you have so if you look at the difference here is that you have got another dependent variable so there are many possibilities here like because there are multiple dependent variable multiple independent variable there are many possibilities for example relationship between independent variable relationship between dependent variable and relationship between independent and dependent variables in ANOVA you are mostly bothered about relationship between independent and dependent variable however in this case there are other possibilities that also occur let me explain you some huge cases of manova where do you apply it so let us talk of a case which is very popular known as all data it is all about the skull of human being had change over time or not the skull is defined by four parameters this height the back height which is maximum rate you have naxal height and you have this distance if you look at it from chain to this the base of this head if you want to find out that the skull of human being had changed over time or not if you think of your ear is the independent variable over here the skull safe had change or not and a skull shape is getting defined by four parameters this MB P H and H and L B so if you look at here one independent variable and multiple dependent variable and that is why I teach a one factor ANOVA let me give you another case where you have more than one dependent variable that's about a pottery when pottery comes you actually like to check the content of these chemicals aluminum iron magnesium calcium sodium etc if you look at you are trying to find is there a difference in percentage of beach chemical across different states and how does it help it will help you because then you can probably predict by knowing just this content that this particular pottery would have come from a particular state and that will be of humain self in prediction if you look at in this scenario how many dependent variable that you have probably you can pause video and you can think of you have at least five dependent variable if you are interested in knowing the percentage of these five chemicals you have at least Beach five at the dependent variable now how will you solve it let me explain you the null hypothesis of a no manova and hard-edged it execute make a note in case of ANOVA you have you check the depend variable average because here you have multiple different variable you look at group mean vectors because you have multiple dependent variable so you are looking at the average of all the variables and that saw you are getting two vectors so you're looking at group mean vectors of all dependent of all the classes are same or not so here you know it's scenario is very much like this mu1 mu2 mu3 or not however here mu1 mu2 mu3 artha group mean vectors in one line it is as good as the means of the variables the dependent variables are same across different treatment groups are not so how will it get executed in case of anova you click create sum of squares and you distribute in two parts between sum of square and within sum of square and then you actually calculate ratio between sum of square to within sum of square which is called F statistics and if p-value associated with F statistics is quite small you reject null hypothesis which simply means that all the group means are not same manova works very much similarly the only difference is that here you had one variable that so you're talking of sum of square here you call up total scatter matrix because you have many dependent variables and again you divide that total scatter matrix into two part between group matrix and within group scatter matrix then you calculate W inverse B which is very similar to this if you think of and if the p-value is associated with this matrix is quite small you reject null hypothesis which means that the dependent variables are not same across different groups let me show you a workout example so I am going to use the skull data and I am going to show you that how do I use this syntax so up to this part is anyway very simple to the are here just I am using C by means like I am taking all the for dependent variable of the skull data if you look at the skull data you had four dependent variables BH + b BL and MB so you are taking all these for depend variable using AM bind command and that becomes your dependent variable and add factor you are taking time Apoc which is the independent variable which is the year here let me show you the data and show you the execution so here you have the data and if you just take a look at the pivot you realize this data belongs to 4000 BC 3300 BC 1 8 5 0 we see 200 BC and 150 AD and if you look at the MB it is increasing over time so this particular width is increasing over time if you look at BH it is going down over time so this particular height is going becoming less over time this height you know BL again its reducing and if you look at Annette it is slightly increasing but you have 4 different variables let's look at hard-edged our solve it so I have already saved this data at CSP and I am just going to read it here so here you know these part is very simple because here you are just going to read this data that CSV file so here I have this data in the CSV format and you can take a look that these are the different we have 30 30 data from each of the period so you have 150 records this is how you run it and let us look at the summary the most important portion here is that this p value because this p value is very simple it simply means that all the four dependent variables are not same across time group which means it has changed significantly across time thank you I'll suggest you to subscribe to this channel so that you can keep updated when I upload new huge thank you for watching

Channel: Gopal Prasad Malakar

Views: 23,108

Keywords: MANOVA, Basis Statistics

Id: jUksjmKvwos

Length: 8min 57sec (537 seconds)

Published: Tue Jan 31 2017