Selection bias in the serology study subsample is also evident in the ORs for the 2008 09 seasonal influenza vaccine, for which the ORs were different, according to whether a serologic or clinical outcome was used, because the effect would be expected to be in the same direction.
Additionally, we performed several sensitivity analyses in selected subsamples to address specific potential biases (e.g. in the subsample of non-movers to address potential bias due to health-related changes in exposure).
In the subsample setting, the bias towards the null and the bias in the direction of the observational association (which is usually in the same direction as the causal effect when this is present) can balance each other out, as demonstrated in the left-hand panel of Figure 3 when the ratio n X: n Y is 0.5.
This may be due to the slight residual bias in the direction of the observational estimate in the subsample case and in the direction of the null in the 2-sample case.
However, because weak IV bias is related to F rather than to R, bias does not vary with the n X: n Y ratio in a uniform way, since increasing n X both increases the proportion of participants in the subsample (leading to a greater bias towards the confounded association) and increases the F statistic, leading to a reduction in weak IV bias.
The subsample regression resultsk showed that first, only in the subsample in which mother has a lower status did sibling sex composition significantly affect educational attainment as expected.
Then the PC traits in the subsample were KYV.
aOnly children in the subsample were allergy tested.
Motivational Variables: Intercorrelations in the subsample of individuals with Mild Cognitive Impairment (MCI) and in the subsample of individuals with Mild Alzheimer Disease ADD).
The mean age (SD) for those analysed in the subsample was 56.1 (12.0) years and for those excluded in subsample because of missing data was 55.4 (13.3) years.
