Bell-Ringing
/If you lined up a large random sample of people, ranging from the shortest at one end to the tallest at the other, you’d notice that most are clustered around the average height, with fewer and fewer people towards the shorter and taller ends. You could then plot the distribution on a graph, with height in intervals of one centimetre along the horizontal axis and the number of people in each interval on the vertical axis. Smoothing out the dots, the distribution would be more or less bell-shaped – hence the bell-curve.
The concept achieved notoriety with the publication of a controversial book in 1994, ‘The Bell Curve: Intelligence and Class Structure in American Life’, by Murray and Herrnstein. The authors argued that IQ is predominantly genetic and unchangeable, which led them to echo (though without actually using the word) the long-discredited eugenic view that those with low IQs should be discouraged from reproducing and those with high IQs encouraged to do so.
Even more controversial, they generalised the argument to endorse supposed differences between IQs of white and black Americans. In fact, during the early part of the 20th century eugenic views were widespread, and it took the genocidal application of eugenics by the Nazi regime during the second World War to expose the idea as simple bigoted racism. Since then, the principles of eugenics have resurfaced from time to time under the guise of ‘science’, but while they still find adherents they are robustly challenged.
Unfortunately, one consequence of the publication of Murray and Herrnstein’s book has been a kind of halo effect, where the distinction between specious theory and statistical distributions has become blurred. Just as with height, many real-world features are in fact bell-curve distributed. There are of course different distribution patterns, and the curve might not conform to a smooth bell shape, but in traits that are bell-curved there is a range between opposite extremes, with the majority grouped around the average or mean.
More or less symmetrical bell distributions around the mean are called normal distributions, and normal distributions have unique properties that make them particularly advantageous for statistical data-analysis. One of these features is that you can estimate how many people would score within specified distances above and below the mean. Formally, this is expressed in standard deviations: for example, approximately 68% of the sample would fall within one standard deviation above and below the mean, and about 95% within two standard deviations.
The relevance of the bell-curve for the Challenge of Change Resilience Training® lies in the eight discrete scales comprising the psychometric Challenge of Change Profile®. Scores on each scale range from 0 to 10, and all of them display approximately symmetric bell-curve distributions in random samples of people. The magnitude of the score is particularly important for the scales we call ‘one-edged’, where for example either low scores (for Rumination) or high scores (for Detached Coping) are definitely preferable, but there is a preferred direction, high or low, for all of the scales.
Scores on psychometric scales are always subject to various forms of bias. For example, the longer the question is pondered the less reliable the answer will tend to be. The first response that comes to mind is generally the most reliable, hence the encouragement to people to base their responses on that rather than trying to consider every possible variation in the situation described in the question. Response bias is further minimised by the extent and sophistication of the statistics used to construct questionnaires, and an important feature of the Challenge of Change Profile is that the scales have all been subjected to lengthy and ongoing statistical analyses. The validity of the scales has also been informally endorsed by the participants themselves, in their responses to anonymous end-of-course feedback questionnaires.
A degree of caution is nonetheless required when interpreting the scores, and an advantage of the normal distributions is that appropriate cut-offs can be determined. This allows ‘high’ and ‘low’ scores on the scales to be qualified: a high score isn’t just 10, but falls in the range from 8 – 10; a low score is in the range from 0 – 2. By taking response bias into account, statistically-determined ranges help to guard against misinterpreting the results. Again based on distributions, the bell-curve for the Emotional Inhibition scale tends to be skewed towards comparatively higher scores for women than for men, so that more nuanced recommended score-ranges can be offered for this scale.
The Challenge of Change Profile scales provide valid and reliable assessments of the habitual ways that people behave with regard to stress and resilience. They measure a set of relatively stable, habitual behaviours rather than fixed biogenetic predispositions, and underpinning all of the scales is the evidence-based reassurance that, with dedicated practise of the training, scores on all eight scales do shift significantly towards the more favourable range. The training programme shows clearly why these changes contribute to enhanced health and well-being, and offers a unique four-step process for making them happen.