Although these errors are computed over the entire time section, there are no time intervals over which the self prediction results are worse than those when the OTC are used to predict the clinical data.
High self prediction included the answering alternatives 'rather probable','probable' and 'very probable', and Low self prediction the alternatives 'rather improbable'improbableble' and 'very improbable'.
Also High self prediction is a well known predictor [ 14, 15, 28, 29].
Of the total of 50 variables, 2 socioeconomic variables, Low total prior sick-listing and Young age, and 1 subjective variable, High self prediction, were finally considered.
What is perhaps more surprising, is the fact that using office visit data for self prediction apparently has the lowest error.
As cognitive-behavioural therapy, among other items, addresses dysfunctional beliefs [ 52], Cognitive-behavioural rehabilitation given to half of the patients might have a greater impact on the self prediction and result in an underestimation in the association between High self prediction and return-to-work.
We should point out that when we use the clinical data alone (i.e. for self prediction) then the prediction equation is of the form, where the subscript j could be left out since only one filter is used.
28 This is also in line with our newly published prospective study, in which High self prediction qualified as a predictor of stable return to work, while work-related variables did not.
In primary-care patients with non-acute, non-specific spinal pain, including back and/or neck pain, the strong predictors of stable return-to-work were 2 socioeconomic variables, Low total previous sick-listing (including all diagnoses) and Young age (max 44 years), and 1 subjective variable, High self prediction (the patients' own belief in return-to-work).
We also find that manipulations designed to strengthen intentions to carry out a behavior have a larger impact on self-predictions than on the behavior being predicted (Studies 1 and 2), whereas a manipulation designed to influence the ease with which intentions are translated into behavior has a larger impact on actual behavior than on self-predictions (Study 2).
The matrices were constructed using predicted sites in the targets, but as observed above, "self-prediction" was avoided.
