I have always wondered whether the impact factor was a reasonable metric to describe the journal performance of Optical Engineering. This month, I have had some pretty good discussions with SPIE staff and some of the other SPIE journal editors about the topic.

For those of you who are not familiar with the impact factor, it is described on the Thomson Reuters site ( http://thomsonreuters.com/products_services/science/free/essays/impact_factor/) as follows:

In addition to a few of the issues described above, I have wondered whether impact factor is a good metric for Optical Engineering because in general engineers do not publish their work as often as basic scientists or medical researchers, and when they do, the number of cited references tends to be lower than citations in papers written by those in science and medicine. Thus, those who do publish their work are generally not cited as often as authors in other occupations. Maybe the number of downloads would be a better metric. I know many optical engineers who download papers and use the results, but they do not publish themselves. Also, impact factor does not take into account the number of readers and the amount of circulation, so that larger journals that make a larger difference with their constituents are not given credit for a large following. Finally, how long a journal has been around is not counted in impact factor, so papers that are older but are still generating citations do not contribute to the metric.

I think things are changing. Many organizations use the $h$-index as a metric to describe the productivity and impact of their researchers. Wikipedia describes the $h$-index as follows:

My $h$-index can be determined from the graph in Fig. 1. My highest cited publication was cited almost 120 times. My next publication was cited 75 times and so on. If publications are sorted in order of number of citations, the $h$-index value occurs where the publication number equals the number of citations. In my case, my $h$-index is around 18. This is not a bad number for an engineer with many years of experience. Scientists with comparable longevity often have higher $h$-indices. So, what gives a researcher a high $h$-index? Someone who publishes a good number of papers where the papers are highly cited will have a high $h$-index. The $h$-index goes up with researcher age and full professors are expected to have higher $h$-indices than younger researchers.

Fig. 1

Driggers citations per publication from highest cited publication to lower cited publications. The $h$-index is determined at the point of intersection of the two curves.

Why did I go through the $h$-index with you? Google Scholar has developed a new journal metric called the h5-index that is very similar to a researcher’s $h$-index but is applied to a journal as a whole. It is the $h$-index for all articles published in the journal in the last five years. Google Scholar defines the h5-index as follows:

Thus, for 2012 a journal’s h5-index is the largest number $h$ such that $h$ articles published in 2007–2011 have at least $h$ citations each. It is very similar to the $h$-index for researchers, but the citations are only counted for the past five years. As you can imagine, journals that publish a good number of papers are likely to have a higher h5-index, as are journals that have high citation numbers. Finally, the h5-median for a publication is the median number of citations for the articles that make up its h5-core list of papers.

Optical Engineering has a current impact factor (for 2011) of 0.959 and the h5-index is 28; that is, 28 OE papers published from 2007–2011 have been cited at least 28 times by publications covered in Google Scholar. The h5-median for these 28 papers is 36. Given the number of papers published relative to other journals in the field and the journal’s emphasis on engineering, I am reasonably pleased with this number. OE’s impact factor has improved over the past few years, but I am a little concerned about next year since in 2011 we had a 45% increase in papers published over 2010. This increase in paper count could increase the h5-index assuming that the average paper quality is maintained. The 2013 impact factor should also increase, since our acceptance rate decreased significantly in 2012, and our paper count increased by 16% over 2011. In both cases, the h5-index should improve with increased paper quality and a larger number of published papers.

At Optical Engineering, we will continue to monitor these journal metrics, and we are committed to improving the quality of the journal. We will continue to improve the journal such that it serves our optical engineering and optical sciences constituencies in the best possible manner. Finally, I want to make sure our junior scientists and engineers pay attention to their $h$-index. The short pep talk is to continue to publish papers each year and, if possible, publish papers that are likely to be downloaded, read, and cited.