Divide the data into two time periods or types of data, based on when or how activity is to be found. For example, if you took measurements on your subjects before administering a drug and then again after administering a drug, divide your dataset into two smaller datasets; one for pre-administration and one for post-administration.
Calculate the means for the datasets. Starting with the first dataset, add all of the values together. Follow by dividing the sum by the number of data points in your dataset (call the number of data points “n” for convenience). Call this value “m2.” Repeat the process for the second dataset, calling the final value “m1.”
Calculate the variance for the datasets. Subtract m2 from every point of data in your first dataset. Square the resulting numbers and add them together. Divide the resulting sum by the number of data points, n. This is the variance for the first dataset; call it “v1”. Repeat the process for the second dataset, using m1 in place of m2. Call the result “v2.”
Compute the Z-statistic. Use the formula Z=(m2-m1)/(v1/n+v2/n)^(1/2).
Determine whether activity is present. Compare the Z-statistic to the value 1.96. If you find that Z > 1.96 is true, then there is activity present in your sample (with confidence 95 percent). Otherwise, you should state that there is no evidence of activity present in your sample and that there is no need to estimate activity for this reason.
Estimate the effect size, or magnitude, of the activity. Take the mean of the original dataset by summing all of the data points and dividing by n. Subtract this value from every original data point. Square the new data points. Sum these squares. Divide this sum by n. Take the square root of this value. Call the result “s.” Estimate the effect size by the formula (m2-m1)/s. This is a unitless measure of the activity in the data. Values above .3 show moderately high activity.