How do I read the graphs
Bar chart of Standard Scores explained:
Standard Scores are like IQ scores. The mean of the distribution has been "normalized" to be 100, and the standard deviation has been set at 15 by rescaling of the data as necessary. For this purpose the norming data have been arbitrarily fitted to a normal distribution. Such a fit is not always a good representation of reality. For example, in the case of omission errors (the Sustained Attention measure) it is not possible to score better than no errors at all. For adults, it is normal to not make such errors. The "no-error" case is therefore close to the population mean. Since it is not possible to score any better, the entire upper part of the normal curve is missing. Since omission errors are rare in adults, every such error ranks high when calculated in terms of standard deviations. That is to say, a change of even a single omission error can mean a very different standard score. The data are very grainy. For these reasons, the normal curve is not a good approximation of reality for omission errors in adults. Hence, the data should not be over-interpreted.

More plentiful, and therefore more fine-grained, are the commission errors (the Impulse Control measure). However, a ceiling still exists at zero commission errors, and that ceiling establishes a maximum possible score that is achievable. This again makes the normal curve only an approximation.

The response time and variability data do not have the ceiling problem, nor do we have a graininess problem. (These are the "Speed of Response" and "Consistency of Response" measures, respectively.) Hence, these data may be the most revealing of nervous system functioning, particularly at the upper end of the performance spectrum where we may have few commission and no omission errors. These data should be evaluated in some detail.

We provide a plot of reaction times over the four quarters of the session. Of particular interest are trends over the session, as well as differences between first and second halves, and between quarters within the halves. If the major problem is in the first half, we may have an issue around the maintenance of vigilance in the test. If the major problem is in the second half, it may be a matter of poorer performance under pressure, which may be anxiety-related. If performance improves within the first or second halves, the initial decrement may also be anxiety-related. If there is a decrement in the second quarters of each half, that may be vigilance- or fatigue-related.

We also provide a histogram of reaction times for each quarter, allowing trends to be observed. Of particular interest is whether the response time distribution looks like a "normal," i.e., Gaussian, curve. A tail in the distribution calls attention to the existence of long-latency reaction times that may indicate nervous system instabilities or in the extreme, even brief "absences." Perhaps the most significant measure of progress in training is the suppression of these outliers, and the normalization of the variability measure.