What is ANOVA? A beginners guide

When conducting an experiment, you encounter experimental error and random fluctuations that are unavoidable. So, how do you determine if the observed differences in your result variable are due to the actual factor you are investigating or just by chance? The answer is to perform an ANOVA.

What are significant parameters?

In any experiment, it’s crucial to identify which factors significantly impact your results. These significant parameters are the factors that cause real changes in your response variable, beyond what might happen due to random chance. ANOVA helps us pinpoint these parameters by analyzing the variation within our data.

Why do I need to determine significant parameters?

Using ANOVA allows you to save time and resources. To do this effectively, you can start with a screening experiment to identify all significant parameters. If a parameter, such as temperature or pressure, does not have a significant effect on your response variable, you can exclude it from further tests.

What is ANOVA doing?

ANOVA breaks down the total variation in your data into two parts:

  1. Variation due to the factors you are testing (between-group variation)
  2. Random variation or experimental error (within-group variation)

By comparing these variations, ANOVA determines if the factors you’re testing have a statistically significant effect on the response variable or if the effect that you are observing is just random variation. The significance is typically determined using the p-value.

The p-value

The p-value in ANOVA shows how likely it is that the differences we see in the data are due to chance. If the p-value is low (usually less than 0.05), it means the differences are likely real with only a 5 % probability of being due to chance. A p-value of 0.05 is commonly chosen as a threshold. If the p-value is smaller than this threshold, we assume that we are looking at a significant parameter.

Take a look at the next blog post if you want to know more about ANOVA and how to perform an ANOVA in python.

Previous
Previous

ANOVA with Python for intermediates

Next
Next

What is a model in DoE and why do I need one?