Driver Strength
The strength of drivers in driver analyses is determined by calculating Kendall's tau-c statistic. This statistic is appropriate for measuring the ordinal association between two variables, which is applicable in this context since surveys deal with ordinal data (e.g., survey items measured on a 5-point ordinal scale).
Driver strength is determined by classifying the absolute value of Kendall's correlation coefficient as follows:
-
Very High: > .7
-
High: > .4 and <= .7
-
Medium: > .3 and <= .4
-
Low: > .2 and <= .3
-
Negligible: <= .2
Driver strength is based only on the effect size criteria outlined above. Driver strength is not a measure of statistical significance. This is because the likelihood of observing a statistically significant result varies as a function of the sample size (e.g., number of respondents); larger samples will result in a larger number of significant results — even when the size of Kendall's coefficient is negligible.