You can make it look even nicer by drawing a horizontal line at zero. Then select Type/Scatter, click Apply, and click OK.Ī nice looking scatter plot will appear. With these factors in mind, open the graph and select Options.
Also, given that income is not measured in equally spaced intervals, dots are preferred to a line graph. You need to tell EViews to change the graph so that INCOME is on the x-axis and EHA T is on the y-axis. It may not look like you expected! Unless told otherwise, EViews will assume you want both INCOME and EHAT graphed against the observation number. The one that is to go on the x-axis comes first.Īfter clicking OK one more time, a graph object will appear in your workfile.ĭouble clicking on this object will open it.
#White test eviews series
As a name for the graph, we chose EHAT_ON_INCOME.Īfter clicking OK, you will be asked for the series that you want to graph. Returning to our task of graphing the residuals, we create a graph object by going to Object/ New Object and selecting Graph. Examples of these two alternatives for the first command follow. Recall that these commands can be executed by typing them in the upper EViews window or by clicking on Genr and writing the equation to generate the series in the resulting box. To graph the residuals against income we begin by naming the residuals and the fitted values. To help you assess which observations could be viewed as outliers, dotted lines are drawn at points one standard deviation (a = 89.517 ) either side of zero.ġ.2. Nevertheless, residual graphs like the one below are important for examining which observations are not well captured by the estimated model (outliers), and, in the case of time series data, for discerning patterns in the residuals. Given it is this latter relationship that we are really interested in, it is preferable to graph the residuals against income. The reason such is the case is that the observations are ordered according to increasing values of INCOME, and the absolute magnitude of the residuals increases as INCOME increases. In the residual graph that follows it is clear that the absolute magnitude of the residuals has a tendency to be larger as the observation number gets larger. As an example, consider the Residual Graph selected in the following way. In each case the series are graphed against the observation number.
#White test eviews free
The Standardized Residual Graph is a graph of e/a the residuals have been standardized (made free of units of measurement) by dividing by the estimated standard deviation of the error term. In terms of the names of the series in your workfile As you might expect from the names of the options, each alternative presents information on one or more of the series actual, fitted and residual. At that point you will see a menu with the following options.Ĭheck each of these options to get a feel for the different ways in which they convey information. After estimating the equation and naming it ls_eqn, go to View/Actual, Fitted, Residual. Let us begin by checking the obvious ones. There are a variety of ways in which EViews can be used to examine least squares residuals.
If they increase with increasing income, that suggests the error variance increases with income. To carry out a preliminary investigation of this question, we examine the least squares residuals. We are now concerned with whether the error variance for this equation is likely to vary over observations, a characteristic called heteroskedasticity. Data in the file food.wfl were used to find the following least squares estimates.
In this chapter we return to the example considered in Chapters 2 to 4 where weekly expenditure on food was related to income.