The impact factor is based solely on citation data and only looks at the citation frequency of articles from a journal in their first couple years of publication. Journals with articles that are steadily cited for a long period of time (say, 10 years) rather than only immediately lose out with this calculation.
Large and small journals are compared equally. Large journals tend to have higher impact factors--nothing to do with their quality.
It’s important to remember that the impact factor only looks at an average citation and that a journal may have a few highly cited papers that greatly increase its impact factor, while other papers in that same journal may not be cited at all. Therefore, there is no direct correlation between an individual article’s citation frequency or quality and the journal impact factor.
Impact factors are calculated using citations not only from research articles but also review articles (which tend to receive more citations), editorials, letters, meeting abstracts, and notes. The inclusion of these publications provides the opportunity for editors and publishers to manipulate the ratio used to calculate impact factor and falsely try to increase their number.
Rapidly changing and growing fields (e.g. biochemistry and molecular biology) have much higher immediate citation rates, so those journals will always have higher impact factors than nursing, for instance.
There is unequal depth of coverage in different disciplines. In the health sciences, the Institute for Scientific Information (ISI), the company which publishes impact factors, has focused much of their attention on indexing and citation data from journals in clinical medicine and biomedical research and has not focused on nursing as much. Very few nursing journals are included in their calculations (around 45). This does not mean that nursing journals they do not include are of lesser quality, and, in fact, they do not give any explanation for why some journals are included and others are not. In general, ISI focuses more heavily on journal dependent disciplines in the sciences and provides less coverage for areas of the social sciences and humanities, where books and other publishing formats are still common.
In some disciplines such as some areas of clinical medicine where there is not a distinct separation between clinical/practitioner versus research journals, research journals tend to have higher citation rates. This may also apply to nursing.
The information above comes from Michigan State University's Research Guides