# Statistical Evaluation of Prairie Restoration Results

How hypothesis testing, control and treatment plots, sample size and variance turn prairie restoration data into a clear, defensible answer.

Statistical evaluation of [prairie restoration](/articles/native-prairie-restoration-getting-started/) results is the step that turns field notes into a defensible answer about whether a planting, burn or seeding actually changed the vegetation. It starts with a clear question, gathers restoration data from matched plots, and uses hypothesis testing to judge whether the difference you see is real. The same discipline keeps you from claiming success from a single good season. 

## Frame the question before sampling 

A sound statistical evaluation of prairie restoration results begins with one written question about a measurable response. Name the plant or bird measure, the direction of change you expect, and the comparison that would answer it. Restoration data gathered without a stated question becomes a pile of observations that no test can sort out. 

## Match control and treatment plots 

Control and treatment plots give your comparison a baseline that shares the same soil, slope and weather. Pair plots that are alike in those conditions and assign the restoration treatment to one of each pair. A before-after comparison within the same plot adds a second check, but it cannot separate restoration effects from seasonal change on its own. Careful plot layout is a job for someone trained in field design. 

## Read variance, sample size and significance 

Variance tells you how widely the plots scatter around their averages, and high variance hides real differences. A sample size large enough to see the scatter is the only honest way to build power into a test. Significance then reports how often a difference this large would appear if the treatment did nothing. 

## Prairie Restoration Statistics Questions 

### How do I run a statistical evaluation of prairie restoration results? 

A statistical evaluation of prairie restoration results runs from a written question to a test chosen for your data type. Collect paired control and treatment plots, record the response the same way each season, and estimate variance before you test. State the result as a difference with its uncertainty, not as a pass or fail. 

### What is significance in a prairie restoration test? 

Significance is the probability that a difference as large as yours would appear if the restoration had no effect. A small probability supports the claim that the treatment changed the vegetation or bird use. Significance does not measure how big or how valuable the change is, so report the size of the difference too. 

### How large should my sample size be? 

Sample size depends on the variance in your plots and on the smallest difference you care to detect. More variable prairie ground needs more plots to reach the same clarity. Estimate the variance from a pilot season, then use that figure to set the number of plots before the main sampling begins. 

### Do I need before-after comparison as well as control plots? 

A before-after comparison alone is weak because weather and seasons shift the vegetation whether or not you restore it. Control plots running alongside the treatment carry that shared change, so the treatment effect stands out. A before-after record strengthens the design when paired with controls, and it helps document the starting condition.
