ADF vs. KPSS Test: How to Check Stationarity in Time-Series Data

Stationarity is an important concept in time-series analysis. Many econometric methods rely on assumptions about how a series behaves over time, so checking whether a series contains a unit root can be an important step before fitting a model.

What is the ADF test?

The Augmented Dickey-Fuller (ADF) test is a statistical test used to investigate whether a time series contains a unit root. Its null hypothesis is that the series has a unit root. A small p-value may provide evidence against this null hypothesis, depending on the selected significance level and test specification.

What is the KPSS test?

The Kwiatkowski-Phillips-Schmidt-Shin (KPSS) test approaches the question differently. Its null hypothesis is stationarity around a specified level or trend. Using both tests can provide complementary evidence when examining time-series behaviour.

Why use both tests?

The ADF and KPSS tests have different null hypotheses. Considering their results together can help analysts understand whether the evidence is consistent with stationarity or a unit root. Results should still be interpreted carefully, taking account of lag selection, deterministic terms, sample size, and the characteristics of the data.

Try an online unit root test

Researchers and students who want a convenient way to explore these tests can use the ADF test online and KPSS testing tool.

Before drawing conclusions, check the test settings and interpret the results in the context of your research question. No single statistical test should replace sound model selection and diagnostic analysis.

ADF vs. KPSS Test: How to Check Stationarity in Time-Series Data

Stationarity is an important concept in time-series analysis. Many econometric methods rely on assumptions about how a series behaves over time, so checking whether a series contains a unit root can be an important step before fitting a model.

What is the ADF test?

The Augmented Dickey-Fuller (ADF) test is a statistical test used to investigate whether a time series contains a unit root. Its null hypothesis is that the series has a unit root. A small p-value may provide evidence against this null hypothesis, depending on the selected significance level and test specification.

What is the KPSS test?

The Kwiatkowski-Phillips-Schmidt-Shin (KPSS) test approaches the question differently. Its null hypothesis is stationarity around a specified level or trend. Using both tests can provide complementary evidence when examining time-series behaviour.

Why use both tests?

The ADF and KPSS tests have different null hypotheses. Considering their results together can help analysts understand whether the evidence is consistent with stationarity or a unit root. Results should still be interpreted carefully, taking account of lag selection, deterministic terms, sample size, and the characteristics of the data.

Try an online unit root test

Researchers and students who want a convenient way to explore these tests can use the ADF test online and KPSS testing tool.

Before drawing conclusions, check the test settings and interpret the results in the context of your research question. No single statistical test should replace sound model selection and diagnostic analysis.