Fig 27. Implementation of findings in developing countries according to author residence. There was no significant difference in proportion of findings implemented (a) with a corresponding author resident to the country of study (X 2 = 0.02, df = 1, p= 0.89) (b) with any author resident (X 2 =0.22,df = 1,p=0.63) Further analysis, however, revealed a three way interaction between: status of the country,
whether or not the authors were affiliated to NGO/government, and whether authors were
resident in the country of study.
Table 9. The interaction between country status, author affiliations, and author residence. Summarised output of logistic regression showing the significance of the three way interaction in the model Factor
P(>|Chi|)
Country status
Author affiliations
Author residence
status:affiliations
status:residence
affiliations: residence
status:affiliations:residence
0.03
0.0003
0.34
0.22
0.15
0.22
0.03
This was due to the fact that author affiliations and residence have more of an impact in
developed countries than in developing. In developing countries, a high number of co-authors
resident to the country had NGO or government affiliations (fig. 28 (a)), but this made no
difference to implementation of findings. However, there were lower levels of implementation
when only the corresponding author was resident to the country (n=10), even than when there
was no author resident to the country (fig. 28 (b)).
55
Fig 28. (a) Author affiliations and (b) proportion of findings implemented, as a function of whether both (Both), co-authors only (Co), corresponding author only (Cor), and no authors (None) were resident in
