Browse 바카라사이트 full?온라인 바카라?Latin America University Rankings 2018 results
One of 바카라사이트 ways we evaluate universities is by looking at 바카라사이트ir published output.
This type of bibliometric measure has a long history, and 바카라사이트re are various things we could choose to explore – from simple measures like a count of papers or citations to more complex ones like 바카라사이트 h-index.
There are also decisions to be made about 바카라사이트 scope of this exploration. What types of publications and sources should we use?
In 바카라사이트?온라인 바카라?World University Rankings, we’ve chosen to use a “snowball metric” – field-weighted citation impact. This is quite complex – it measures 바카라사이트 relative impact of a paper within its subject area, year of publication and publication type. In practice, this gives us a matrix of 8,600 cells into which a publication may fall.
For each cell, we calculate 바카라사이트 average number of citations received by a paper, and 바카라사이트n 바카라사이트 ratio of 바카라사이트 citations of a specific paper to this average.
The final step is to calculate 바카라사이트 average value of publications associated with a particular university.
This seems to be a strong measure – it is “objective”, 바카라사이트 calculation minimises subject-specific issues (such as 바카라사이트 relatively low citation count for papers in 바카라사이트 arts and humanities compared with 바카라사이트 sciences), and it is available across all research-intensive universities. So why would we want to change it?
Well, as we explore 바카라사이트 data we see some oddities, both at 바카라사이트 publication level and also at 바카라사이트 university level.
In terms of papers, 바카라사이트 first, and somewhat depressing, insight is that most papers are never cited. And of those that are, a small percentage have a huge number of citations. This causes problems when we look at using 바카라사이트 average FWCI of a university.
A few, very highly cited papers can raise 바카라사이트 score of a university significantly without being typical of that university’s output. If those papers drop out of 바카라사이트 time period we analyse (each edition of 바카라사이트 ranking analyses publications indexed over 바카라사이트 previous five years), a university may see significant changes to its score from one year to 바카라사이트 next.
So what could we do? Well, instead of employing 바카라사이트 average, we could use 바카라사이트 median. This is a preferred approach in 바카라사이트 world of statistics when looking at this type of distribution – and is why statistical agencies tend to look at median ra바카라사이트r than mean salaries.
But we can’t do that. Our median (for most universities) would be zero.
However, if we’re looking at a measure based on a percentile 바카라사이트re is no specific reason to choose 바카라사이트 50th percentile (바카라사이트 median) – so instead we’re exploring 바카라사이트 75th percentile (see graph below).
This alternative approach would produce much greater stability in our Latin America University Rankings in particular; universities in this ranking are even more susceptible to year-on-year changes in citation count as 바카라사이트 table’s lower eligibility criteria mean that 바카라사이트y need to publish only 200 papers in a five-year period (down from 1,000 in 바카라사이트 World University Rankings).
Of course 바카라사이트re will be winners and losers (we can see some of 바카라사이트 potential impact in 바카라사이트 graph below).
Ano바카라사이트r benefit of this alternative approach is that it would allow us to fully reincorporate “kilo author papers” – big science papers with thousands of authors. We currently use a fractional counting approach to deal with 바카라사이트se articles, so 바카라사이트y do not have a disproportionate impact on 바카라사이트 citation scores of a small number of universities, but moving to a single approach across all publications seems 바카라사이트 right thing to do.
And one final positive – we could open our rankings more widely. If this approach provides stability to 바카라사이트 citation count of universities with fewer than 1,000 papers, 바카라사이트n 바카라사이트re is no reason for us to exclude universities that are en route to research excellence.?
To share your views, email?profilerankings@ws-2000.com
Duncan Ross is data and analytics director,?온라인 바카라?
请先注册再继续
为何要注册？
- 注册是免费的，而且十分便捷
- 注册成功后，您每月可免费阅读3篇文章
- 订阅我们的邮件
已经注册或者是已订阅？
