Figure 1
The flow begins with a text box labeled “Records identified from Web of Science (n equals 480, 282)” on the top. From “Records identified from Web of Science (n equals 480, 282)”, a downward arrow arises and points to a text box labeled “Records retained (n equals 13, 304)”. From this downward arrow, a right-pointing arrow arises and points to a text box labeled “Records in non-management field removed (n equals 466, 978)”. From “Records retained (n equals 13, 304)”, a downward arrow arises and points to a text box labeled “Articles screened (n equals 8, 573)”. From this downward arrow, a right-pointing arrow arises and points to a text box labeled “Non-articles superscript a excluded (n equals 4, 731)”. From “Articles screened (n equals 8, 573)”, a downward arrow arises and points to a text box labeled “Articles screened (n equals 6, 347)”. From this downward arrow, a right-pointing arrow arises and points to a text box labeled “Non-S S C I articles excluded (n equals 2, 226)”. From “Articles screened (n equals 6, 347)”, a downward arrow arises and points to a text box labeled “Articles assessed for eligibility (n equals 6, 306)”. From this downward arrow, a right-pointing arrow arises and points to a text box labeled “Articles in non-English excluded (n equals 41)”. From “Articles assessed for eligibility (n equals 6, 306)”, a downward arrow arises and points to a text box labeled “Articles included in review (n equals 725)”. From this downward arrow, a right-pointing arrow arises and points to a text box labeled “Irrelevant articles superscript b excluded (n equals 5, 581)”.The process of article selection