BACKGROUND AND OBJECTIVE: To adjust the Impact Factor (IF) provided by the Institute for Scientific Information (ISI) is necessary for improving bibliometric analysis among the various areas of knowledge. Our objective was to evaluate two parameters (the maximum and the median value of each subject area) for IF adjustment using the original (not corrected) IF as the reference method in a tertiary hospital biomedical investigation model. MATERIAL AND METHOD: We retrieved articles from Hospital Universitario "Del Río Hortega" from Valladolid (Spain) for the period 1999-2004 as data source. We have describe the characteristics of the followed IF distributions: IF Corrected by the Maximum value (IFCORMAX), IF Corrected by the Median value (IFCORMED) and IF without adjustments (IF). Besides, we have analyzed both the inter-annual and the inter-subject differences obtained by the three methods. RESULTS: The three analyzed IF series shown not normal distributions that are positively skewed. The IF adjusted by the median showed the highest coefficient of variation (CV = 357.3%). The IF adjusted by the Maximum value increased the "weight" of journals with the highest not corrected IF for each subject category. Inter-annual differences were similarly estimated by the three methods. The IF adjusted by the median increased the inter-subject difference from 7.3% to 12.4%. CONCLUSIONS: Our results suggests that IF adjusted by the maximum value of each discipline is the best method to correct the ISI-IF values, because journals with the high IF are always rewarded, while IF adjusted by the median infra-estimated most of them.
BACKGROUND AND OBJECTIVE: To adjust the Impact Factor (IF) provided by the Institute for Scientific Information (ISI) is necessary for improving bibliometric analysis among the various areas of knowledge. Our objective was to evaluate two parameters (the maximum and the median value of each subject area) for IF adjustment using the original (not corrected) IF as the reference method in a tertiary hospital biomedical investigation model. MATERIAL AND METHOD: We retrieved articles from Hospital Universitario "Del Río Hortega" from Valladolid (Spain) for the period 1999-2004 as data source. We have describe the characteristics of the followed IF distributions: IF Corrected by the Maximum value (IFCORMAX), IF Corrected by the Median value (IFCORMED) and IF without adjustments (IF). Besides, we have analyzed both the inter-annual and the inter-subject differences obtained by the three methods. RESULTS: The three analyzed IF series shown not normal distributions that are positively skewed. The IF adjusted by the median showed the highest coefficient of variation (CV = 357.3%). The IF adjusted by the Maximum value increased the "weight" of journals with the highest not corrected IF for each subject category. Inter-annual differences were similarly estimated by the three methods. The IF adjusted by the median increased the inter-subject difference from 7.3% to 12.4%. CONCLUSIONS: Our results suggests that IF adjusted by the maximum value of each discipline is the best method to correct the ISI-IF values, because journals with the high IF are always rewarded, while IF adjusted by the median infra-estimated most of them.