Autoregressive Conditional Heteroscedacity
with Estimates of variance of United Kingdom Inflation.
Residual scatterplots showed no evidence of homoscedacity or heteroscedacity
We tested for heteroscedacity
in variables analyzed using a Brown and Forsythe's test (Brown and Forsythe 1974).
the F-statistic, White: the White test statistic for heteroscedacity
, J-B: the Jaraque-Bera test statistic for normal distribution.
To check model adequacy, we inspected residuals for outliers and obvious signs of heteroscedacity
, tested the normality of residuals using the Shapiro-Wilk test, and checked for autocorrelation within the residuals using the Durbin-Watson test.
Assumptions of ANOVA were checked by examining residuals visually against each independent variable, as well as by using Shapiro-Wilk's test of normality and Hartley's f-max test of heteroscedacity
. Residuals were normally distributed for each treatment level within each factor (legacy and pH treatment).
Most data sets did not meet assumptions of homogeneity, so we used a square-root transformation to correct for heteroscedacity
These estimates include the value and sign of the coefficients, the significance of the coefficients, the [R.sup.2] (and adjusted [R.sup.2]), F-value and the result of other diagnostic tests such as multicollinearity and heteroscedacity
. The respective equations are shown in Table 1.
A log-linear functional form was employed because of the econometric problem arising from the occurrence of heteroscedacity
Approaches to problems of significance, multicollinearity, serial correlation, and heteroscedacity
in time-series analysis are in Franke and Kaul's (1978) and Franke's (1980) appraisals of the Hawthorne Experiments at Western Electric.
West (1987): "A simple positive semi-definite, heteroscedacity
and autocorrelation consistent covariance matrix".