V1.4 - Harvtest

The test is used to determine if the relationship between variables in a linear regression model is truly linear or if it is instead convex or concave.

It performs a t-test on recursive residuals. If the relationship is non-linear, the mean of these residuals will differ significantly from zero. Hypothesis: Null Hypothesis ( H0cap H sub 0 ): The true relationship is linear. Alternative Hypothesis ( H1cap H sub 1 ): The relationship is non-linear (convex or concave).

In R, the function harvtest(formula, order.by = NULL, data = list()) is used to execute the test. Key Components and Usage HARVTEST V1.4

The test returns an htest object containing the t-statistic , the p-value , and the degrees of freedom .

If "HARVTEST V1.4" refers to a different utility, it may be confused with: Harvest: Time Tracking Software With Invoicing The test is used to determine if the

It is frequently used alongside other diagnostic tools like the Breusch-Pagan test (for heteroskedasticity) and the Durbin-Watson test (for autocorrelation). Related "Harvest" Software

When utilizing this test within a regression diagnostic workflow, the following parameters and outputs are typical: formula : A symbolic description of the model to be tested. Hypothesis: Null Hypothesis ( H0cap H sub 0

order.by : An optional vector or formula to order observations (crucial for time-series data).