(EViews10):Interpret Descriptive Statistics #descriptivestats #interpret #eviews #output

from crunch econometrics are we explaining summary or descriptive statistics in eviews before engage in any regression analysis it is essential that you have a feel of your datasets that is what information is contained in your sample summary statistics is relevant for parametric or nonparametric tests it is essential for both qualitative or quantitative research it can easily tell you whether your sample is normally distributed and what are the outliers in the data or that information that you can obtain will be information on measures of central tendency knowing the values for the mean the median and the mode information on measures of dispersion what will be the range of the series the variance the standard deviation percentile scores house or this house as the case may be what about measures of normality you need to know more about the kurtosis or the skewness of the variables it is important for you to use the raw data of the variable and not the transform data so don't use a lock transformation or the first difference of a series if you want to run summary statistics now on measures of central tendency the mean is simply the average value of the particular variable the median in this case will be the middle value after you have sorted from the highest to the lowest or vice versa while the mode will be the most appeared value for the particular variable on the measures of this fashion you need to know how spread out your data is looking at the range which is a difference between the highest and lowest value the variance tells you how widely dispersed the observations are for this particular variable while the standard deviation tells you how far the observations are from the sample average again use the raw data of the variable and not the transformed data or measures of normality we are only considering two there are the kurtosis skewness the kurtosis tells us about the thickness of flatness of the distribution of the series so if we say a distribution is mesokurtic it simply means it embodies a normal distribution we take kurtosis value of 3 if it's leptokurtic it means it has it positive kurtosis it is a bit curve indicating that there are more higher values than the sample mean for this variable am being played psychotic implies that has a negative kurtosis it is a flat head core with small lower values than the sample mean now on the skewness this measures the degree of asymmetry of the series normal schooners implies that the distribution is symmetric around the mean and the skewness value is 0 for positive skewness it implies that this distribution will have a long right tail meaning there are higher values than the sample mean one negative skewness implies that this distribution will have a long left tail with more lower values than the sample mean so having given you all this preamble let us now manova to eviews to take on an example eviews is off i be using this grouped data can see in the arrow forms I have MVA GDP growth rates and the real exchange rates from 1981 to 2014 so I have 34 observations for each of the variables so these are the raw forms of this data they are not in the a log transformation so to run summary statistics all I need to do is to click on view descriptive statistics then click on individual samples so this is the results for the summary statistics I have here the mean the median the maximum minimum values I have the standard deviation skewness kurtosis I also have the jacoub era and a probability value for the Jacobi artistics I have moved the table here to the screen because I have other information that I need to talk about the main funder here simply tells us the average value for each of the variables so for MVA he seeks for GDP growth rates is 3.7 well for the real exchange rate is 150 5.17 the median simply tells us the middle values for each of these three variables while the maximum and minimum values tells us the highest and the lowest figures in each of these variables the standard deviation tells us the deviation from the sample mean with respect to each of the variables the what I want to talk about here will be the skewness for no mass kunis the value is zero so we can say that um this variable mva mirrors a normal distribution let's look at the kurtosis value of 1.88 ketosis measures the thickness of flatness of the distribution of a series and we'll take associates bundle of 1.88 which is clearly lower than the value of 3 because a value of 3 implies that the distribution is normal is mesokurtic but with 1.88 we can say that mva or door mirrors a normal distribution is clearly played psychotic and petty Kotik implies that this series we have lower values below is sample mean is going to have a lots of values that are lower than 6.0 that is what this pretty Kotik implies so it's going to have a flat surface to GDP growth rates the GDP growth rate the skewness value is 1.15 and like we can see being 1.15 GDP growth rates distribution we have a long right tail it's embodies positive skewness and it's clearly leptokurtic looking at the kurtosis value of 0.33 this is a peaked distribution for the real exchange rate it takes a similar interpretation without of GDP growth rates this Cunha's value is one point six seven is we'll have a long right tail indication is positively skewed and the kurtosis value of 4.03 also indicates diced left so Kotik now let's take a look at the Jacobi roster tasty the Jacobin a statistic measures the difference between the skewness and kurtosis of each of these variables with those from a normally distributed variable so we can see here that the jocular statistic for mg is two point zero five that of GDP WordPress's for several points is real exchange rate is twenty point five and below them and their respective probability values the null hypothesis for the Jacobi era test is that the distribution is normal so we can see that for MVA decreasing the probability value is roe point 35 which is above the significance level points zero five so with respect to MVA we cannot reject the null hypothesis so we can say that MVA is a normally distributed curve it has a normal distribution but we cannot say the same for GDP growth rates and real exchange rates in both situation we clearly reject the null hypothesis of a normal distribution because the probability values are highly statistically significant so again for GDP growth rates and the real exchange rates the distributions are clearly not normal and this answer can be clearly seen evil from the result of the kurtosis and skewness of these two variables so given these we can say that before you write your analysis it is essential that you have a good idea of the summary statistics of the data you'll be working with you can easily see what other is an outlier in the data for instance for GDP growth rates the maximum value of thirty three point seven three may likely indicates an outlier because it's different from every other observation in the data and that could be the reason why the curve is so peaked at eight point three three so you might want to take a look at this our generation process to see whether there's an imputation arrow or something so again having descriptive statistics is important before you start one in any regression analysis in case I am too fast please play back this video so that you can get more understanding on words we have discussed thank you for watching subscribe for more videos from crunch econometrics visit our website and our blog and stay tuned for more tutorials

Channel: CrunchEconometrix

Views: 74,185

Rating: 4.8658681 out of 5

Keywords: summary statistics, central tendency, dispersion, normality, variance, standard deviation, quartiles, kurtosis, skewness, long-right tail, long-left tail, mesokurtic, platykurtic, leptokurtic, how to

Id: S7ZO0cgv2IU

Length: 9min 11sec (551 seconds)

Published: Fri Mar 16 2018