SPSS 8 Normal Distribution Test in 3 Approaches

in this video I'm going to show you how to check normality of data so actually we can do it in three ways so first we can just plot it and see in the histogram if the distribution looks like normal distribution and then we can look into the we can do descriptives look at the skewness and kurtosis values and then finally we can run normality test which is the robust way of doing it okay so let's first start with just plotting it okay so let's say if go to analyze if go to your frequencies will just I'm going to check that normality of readings will go here I'll go to charts histogram show normal cover carve on histogram and I'll click continue okay and okay so this is how it looks like it looks actually rather good mean it's not perfect one but it looks pretty good mean we have most of the Fallas centered around the middle point here in the mean and then the distribution kind of okay in both sides but not perfect okay but yeah then to me it looks rather normal and safe but then we want to be more confirmed right so we will go for analyze descriptives and descriptives and I'll pick reading score here okay go to options will just pick Kunis and kurtosis and I'll click continue I'll click okay so for normal distribution data normally we would like to have this unit value of zero and corrosives cutlasses value of three so here we see the skewness value but the skinners value between plus one and minus one is also fine so which which is quite fine here it is close yeah it's it's it's not large its close to zero so it's fine but Carlos's value is far away from from three so that's problematic so we cannot really conclude normal normal distribution of date okay so then we will go for the most robust approach which is the normality test spaces offers two of them and we'll go to explore descriptive and then explore we'll pick the data here we will go for plots and here resorption normality test with plots will just an take this okay normally test with plots and click OK and ok so here are some basic statistics and here is the test of normality so we normally spaces offers two of them called mogh rubbish minima and shep real-world both the clay both both for both the test we have the significance value less than 0.05 which means our data is not normally distributed in in normality test the null hypothesis is that data are normally distributed okay so when our significance values are less than 0.05 we reject the null hypothesis so we reject the null hypothesis that data is normally distributed that means our data is not normally distributed okay so that's it but actually if we look further here on the QQ plot so normally when we loop in the QQ plot this is kind of advanced things and I'm just discussing it because it's already here so if the observed values and expected values if they fall on the line if the dots fall on the line if we plot observed and observed and expected values on two dimensional figure and if it fits on the line then actually data is normally distributed so in this case it is not exactly fitting on the line but if you see it's more or less on the line so actually kind of say is that data is actually to some we are pretty much normally distributed with some fat tails okay so as we see that here we have few are bit far from the line so we can say with fat tails which which actually our Carter's is also kind of indicated that the value of kurtosis also indicated that right so yeah and also we have more than 200 observations so we can actually say although all the previous things kind of all that all these tests and cars value actually goes against normality but still we can kind of say that they dies rather normally distributed with fat tails so these are the three approaches to look for normality test hope this helps you thank you very much if you like the video please like comment share and subscribe