WEBVTT Kind: captions; language: en-us

NOTE
Treffsikkerhet: 85% (HØY)

00:00:00.300 --> 00:00:08.500
In this video we will see how to evaluate variables in jimovie. Let us first load our data set. 

NOTE
Treffsikkerhet: 91% (HØY)

00:00:15.899 --> 00:00:23.900
In order to evaluate the normality of individual variables we have to do three things. One was the 

00:00:23.900 --> 00:00:30.850
visual evaluation of graphs, the other was the evaluation of indices, and the third was the 

00:00:30.850 --> 00:00:34.450
statistical testing of the normality assumption. 

NOTE
Treffsikkerhet: 89% (HØY)

00:00:34.450 --> 00:00:39.800
These are all done through the exploration descriptives panel, 

NOTE
Treffsikkerhet: 88% (HØY)

00:00:39.800 --> 00:00:44.300
so we can choose which variables to do these for 

NOTE
Treffsikkerhet: 84% (HØY)

00:00:44.500 --> 00:00:54.800
for example the matrices in the word fluency in kindergarten and word fluency in first grade, 

NOTE
Treffsikkerhet: 90% (HØY)

00:00:55.100 --> 00:01:02.700
and the first thing we have to do whenever we have a variable you want to look at is the histogram. 

NOTE
Treffsikkerhet: 91% (HØY)

00:01:05.600 --> 00:01:09.300
So here is the histogram for the matrices 

NOTE
Treffsikkerhet: 88% (HØY)

00:01:09.300 --> 00:01:19.500
for word reading fluency in kindergarten where most kids can't read and the same in first grade. And 

00:01:19.500 --> 00:01:26.500
to compare these distributions to the normal distribution we use the quantile quantile plot, so I 

00:01:26.500 --> 00:01:34.700
check on QQ here and I get the quantile quantile plot for each of the variables. 

NOTE
Treffsikkerhet: 91% (HØY)

00:01:39.800 --> 00:01:47.800
To get the indices that are associated with the asymmetry or symmetry of the distribution and the 

00:01:47.800 --> 00:01:56.900
proportion of values away from the mean I have to go and check on skewness and kurtosis in the 

00:01:56.900 --> 00:02:00.300
distribution checks of descriptives. 

NOTE
Treffsikkerhet: 79% (HØY)

00:02:00.800 --> 00:02:07.950
And the statistical evaluation of normality is done using the Shapiro Wilk test. 

NOTE
Treffsikkerhet: 91% (HØY)

00:02:07.950 --> 00:02:14.800
So the skewness of each of the variables is shown on this line 

NOTE
Treffsikkerhet: 87% (HØY)

00:02:14.800 --> 00:02:22.700
and you can see that the first and third variable have low, very low positive skewness is negligible, 

00:02:22.700 --> 00:02:33.400
and the second one has high positive skewness, and the kurtosis is shown on this row here so the first 

00:02:33.400 --> 00:02:39.900
and third variable have very low negative kurtosis that we can safely ignore but the second one has

00:02:39.900 --> 00:02:43.000
high positive kurtosis that we cannot ignore. 

NOTE
Treffsikkerhet: 84% (HØY)

00:02:43.000 --> 00:02:51.400
And consistent with that the Shapiro Wilk test produces a low p-value for the second variable which 

00:02:51.400 --> 00:02:57.100
indicates that it is very unlikely we would have gotten this kind of distribution by sampling from a 

00:02:57.100 --> 00:03:04.350
normally distributed population, whereas the p-values for the other two are quite high and no cause 

00:03:04.350 --> 00:03:06.200
of concern 

NOTE
Treffsikkerhet: 88% (HØY)

00:03:08.300 --> 00:03:15.300
and that is all you have to do in order to produce the necessary information for evaluating the 

00:03:15.300 --> 00:03:18.400
normality assumption in Jimovie.