It is surprising that disability as measured by the ODI was
(in contrast with pain intensity) negatively associated with
the emotionally supportive communication scale, although
disability and pain intensity correlated highly positively
(r=0.658, see Table 4). We interpret this phenomenon as a
suppression effect.50 According to Conger,51 a suppressor
variable is a variable “which increases the predictive validity
of another variable (or set of variables) by its inclusion in a
regression equation”. We tested this and were able to show
that this condition for the ODI variable regarding pain intensity
is fulfilled in prediction of the emotionally supportive
communication KOPRA scale. The regression weight of pain
intensity is clearly lower and loses its significance when the
ODI variable is taken out of the model. Further, the unexpected
significance of the ODI variables disappears when
pain intensity is taken out of the model (see Table 3, which
shows no bivariate correlation in the unexpected direction for
the ODI variable with respect to the emotionally supportive
communication scale). Therefore, it can be assumed that
greater disability, mediated by pain, contributes to greater
communication needs in the emotional area. Simultaneously,
disability adjusted for pain intensity is associated with a lower
preference for emotionally supportive communication.