Skip to main content

Table 5 Factors correlating with the number of non-English studies included in Campbell systematic reviews

From: The prevalence of and factors associated with inclusion of non-English language studies in Campbell systematic reviews: a survey and meta-epidemiological study

Models

B (CI 95%)

Sig.

Standard error

Std. coefficients (beta)

Model 1 (n = 101) survey model

 

0.05

  

 Constant

− 0.93 (− 2.39–0.52)

0.20

0.73

 

 Number of different working countries represented by authors

0.56 (0.07–1.03)

0.02

0.24

0.41

 Education group

0.94 (− 0.10–1.98)

0.08

0.52

0.32

 Number of included studies

0.01 (0.00–0.02)

0.06

0.01

0.34

Model 2 (n = 47) survey with questionnaire model

 

0.07

  

 Constant

− 4.53 (− 9.38–0.32)

0.07

2.34

 

 Number of different working countries represented by authors

0.96 (0.07–1.86)

0.04

0.43

0.42

 Crime and Justice groupa

3.99 (− 0.02–7.99)

0.05

1.93

0.91

 Education groupa

2.85 (− 0.31–6.02)

0.08

1.53

0.89

 International Development groupa

1.92 (0.08–3.75)

0.04

0.89

0.49

 Social Welfare groupa

2.85 (− 0.57–6.23)

0.09

1.65

0.97

Model 3 (n = 77) language explicit model

 

0.09

  

 Constant

− 0.82 (− 2.78–1.14)

0.40

0.94

 

 Number of different working countries represented by authors

0.65 (0.05–1.25)

0.04

0.30

0.48

 Number of screened studies

0.00 (0.00–0.00)

0.09

0.00

− 0.55

 Number of included studies

0.02 (0.00–0.03)

0.06

0.01

0.40

  1. The independent variable is the number of non-English studies included in the systematic reviews published by the Campbell Collaboration. The variable, number of studies meta-analysed, was dropped from the analyses because it had an unacceptably high correlation with the variable accounting for the number of studies included in reviews (Pearson’s R = 0.85; p < 0.01). All other variables were included as planned, but only significant variables are reported. Missing data is excluded pairwise
  2. R2 = model 1, 0.37; model 2, 0.59; model 3, 0.40
  3. aThe substantial implication of belonging to any one subject area covered by C2 is illogical and is thus disregarded