In a next step, a new binary variable is generated which equals 1 in case that both partners work, and 0 in case of complete specialization.
Binary variables were generated using these tertile cut-points.
Per position, separate binary variables were generated for each alternate amino acid, indicating the presence or absence of that allele in a given sample.
Binary variables were generated for working days per week (i.e. ≤5 days or >5 days), shift work and job satisfaction (i.e. yes or no).
12, 13 Binary variables were generated to describe, in summary, seropositivity by species: for example, P. falciparum seropositive samples were defined as seropositive against either PfMSP-1 and/or PfGLURP.
Using conflict termination data, four binary variables are generated in the analysis: terminated in a peace agreement, terminated in a government victory, terminated in a rebel victory, and low activity.
For the proxy measures of child tooth cleaning frequency and effectiveness Debris Index and the Modified Gingival Index, binary variables were generated for the presence or absence of debris on the teeth and gingival inflammation.
New binary variables were generated, in both Lima and Durban datasets, as follows: (i) for each allelic variant on each of the three HLA loci (HLA-A, HLA-B, and HLA-C) a binary variable was created with value 1 if the subject had that respective variant and 0 otherwise; (ii) for each OLP a binary variable was set with value 1 if the subject has responded to this OLP and 0 otherwise.
A current depressive symptoms binary (Yes/No) variable was generated using the cut-off score of ≥10 [ 30].
Due to the highly skewed distribution of the continuous ADDQoL score, a binary ADDQoL variable was generated, which categorized patients into those reporting a negative ('1'), or no/positive impact of diabetes ('0'); the range of the continuous ADDQoL score values at 1 year was -9 to -0.07 ('1') and 0 to 1.15 ('0'), and -7.36 to -0.07 ('1') and 0 to 0.23 ('0') at 5 years.
