Performs the White test for heteroskedasticity by regressing the squared residuals of a linear model on the original regressors and their squared terms. The null hypothesis is homoskedasticity.
eda_white_test(data, years, alpha = 0.05)A list containing the results of the White test:
data: The data argument.
years: The years argument.
alpha: The significance level as specified in the alpha argument.
null_hypothesis: A string describing the null hypothesis.
alternative_hypothesis: A string describing the alternative hypothesis.
statistic: White test statistic based on sample size and r_squared.
p_value: The p-value derived from a Chi-squared distribution with df = 2.
reject: If TRUE, the null hypothesis was rejected at significance alpha.
Numeric vector of observed annual maximum series values. Must be strictly positive, finite, and not missing.
Numeric vector of observation years corresponding to data.
Must be the same length as data and strictly increasing.
Numeric scalar in \([0.01, 0.1]\). The significance level for confidence intervals or hypothesis tests. Default is 0.05.
The White test regresses the squared residuals from a primary linear model
lm(data ~ years) against both the original regressor and its square.
The test statistic is calculated as \(nR^2\), where \(R^2\) is the
coefficient of determination from the auxiliary regression and \(n\) is
the number of elements in the time series. Under the null hypothesis, the
test statistic has the \(\chi^2\) distribution with 2 degrees of freedom.
White, H. (1980). A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica, 48(4), 817–838.
data <- rnorm(n = 100, mean = 100, sd = 10)
years <- seq(from = 1901, to = 2000)
eda_white_test(data, years)
Run the code above in your browser using DataLab