How does Sparta identify meaningful change?

The Short Answer:

Sparta defines a moderate change if the data is greater than or equal to 1 standard deviation away from the individual's baseline (or mean), and a large change if the data is greater than or equal to 2 standard deviations away from the individual's baseline. This simple distinction allows for moderate and large changes to be visually highlighted within the Sparta software to ensure Admin users are aware of any meaningful changes that are taking place for an individual.  It is important to note that changes can either be viewed positively (improvement) or negatively (deterioration) and the interpretation is, as always, context-dependent.

 

The Long Answer:

When it comes to identifying "meaningful" or "significant" change in data, there are a lot of different thoughts and opinions that often confuse practitioners. Things that are statistically significant are often practically irrelevant so it is critical for practitioners to not get too bogged down in academic jargon while still ensuring best practice to remain objective and scientific.

We can do this by simply understanding two concepts: means (average) and standard deviations (variation). Sparta identifies meaningful change for an individual by first comparing data to the historical mean or average for that individual, also termed their baseline (see article). For this, Sparta defines an individual's "baseline" by taking the average of their last 90 days worth of data.  If there is no data within the last 90 days, Sparta will utilize an average of the last 3 scans.

Now that we have identified an individual's baseline, we need to determine what is a meaningful change from that baseline. The greater the variation the more spread out the data is and in turn the greater the standard deviation (a measure of the amount of variation) will be. This is important as if there is a large amount of variation what seems like a large change might actually be within this normal variation.

For example, a 0.1 change might be considered meaningful for a sprint time (4.50 compared to 4.40 seconds) but not so much if we are looking at a bodyweight change (225 compared to 225.1 lbs).

Sparta defines a moderate change if the data is greater than or equal to 1 standard deviation away from the baseline (or mean), and a large change if the data is greater than or equal to 2 standard deviations away from the baseline. This simple distinction allows for moderate and large changes to be visually highlighted within the Sparta software to ensure Admin users are aware of any meaningful changes that are taking place for an individual.  It is important to note that changes can either be viewed positively (improvement) or negatively (deterioration) and the interpretation is, as always, context-dependent.

The above example shows a large increase from baseline in Load, a moderate increase from baseline in Explode, and a large decrease from baseline in Drive.

 

For further information on trends we see related to fatigue or poor recovery check out: How does fatigue affect the assessments?

For further information on how to determine an individual's baseline check out: How do I determine an individual's baseline?

 

* For further clarification: When utilizing T-scores, by definition the mean is 50 and a standard deviation is 10 for the total population. However, when identifying meaningful change, the mean is based on the mean of that individual's data (see baseline above) and the standard deviation is based on the within-subject variation.