Bibliometrics & verification

Bibliometrics is in short a measuring tool for scientific publication. It has developed to a global, strategic tool for science. At Karolinska Institutet it is used to give KI's management the opportunity for analyses that contribute to an overview of research at KI, and to compare KI's research with research from the rest of the world.

Are you a researcher at Karolinska Institutet and/or Region Stockholm? Log in to verify your articles and to analyse your publications.

About bibliometrics

Bibliometrics involves conducting statistical analyses of publications – i.e. a type of "publication measurement" (biblos: book and metron: measure). Bibliometrics is now commonly used to analyse research by performing quantitative studies on its publications.

Current practical bibliometrics is based on the assumption that the majority of research findings are published sooner or later as articles in international academic journals, and are read in that context by other researchers who then go on to cite these articles in their subsequent articles. So the more citations an article receives, the greater the impact this can be taken to have had.

So, in practice, bibliometrics involves measuring the number of published academic articles by a certain group of authors and the number of times these articles are cited, as well as studying the statistical connections between different articles, authors and subjects.

FAQ: General questions

Your h-index is the largest number X such that you have at least X publications with at least X citations.

There are three different ways of finding your h-index:

1. Through KI RIMS you can export your Qualifications Portfolio using the reporting function. All bibliometric parameters are automatically retrieved from KI RIMS to the qualifications portfolio, including h-index.

2. Through Web of Science by searching all your publications and from the resulting list of publications select "Create Citation Report" (to the right above the search result).

3. You can calculate h-index from your list of verified publications in the bibliometric system, (if you have no publications from before 1995 and have verified all your publications).

Login and verify all your publications. Export them to Excel (or anything else that can read csv files) from the publication list in the Analysis Toolkit.

Sort the publications after times cited (with the most cited publications first and so on) and look in the sorted list where the number of citations is at least equal to the order number.

Example:

Order Citations
1 ......... 23
2 ......... 11
3 ......... 5
4 ......... 2
5 ......... 1

In this case, h-index is 3, since publication number 3 has 5 citations (at least 3) but publication number 4 has 2 citations (less than 4).

You can log in to the Bibliometric Verification Toolkit with either a KI-login (Karolinska Institutet) or a HSA-id (Region Stockholm). Choose which option you want to use by clicking the either the KI Login button or the Region Stockholm Login button on the login page.

KI-login: The Karolinska Institutet IT Centrum offers e-mail accounts, KI-login and VPN (Virtual Private Network), see the Staff Portal.

HSA-id: Region Stockholm uses a login name of 4 characters called HSA-id for most Region Stockholm systems. To be able to login to the bibliometric system you also have to use one of two authentication options:

a card reader connected to your computer;
a temporary authentication code via SMS (If you do not have your mobile phone number registered in EK you will have to contact Servicedesk (08-123 777 77) for contact details to your local EK administrator who receives your case. The number is not displayed externally.)

There are presently two ways that we can identify the research fields of individual articles. One is to use the Clarivate Analytics journal subject field as a proxy for article content. This is very often used in bibliometric studies. But in medicine, many of the publications also have MeSH terms, the controlled vocabulary from Medline, which are much more specific, and at Karolinska Institutet we often use these.

For publications in the intersection between the Web of Science and Medline, both citation data and detailed subject information at an article level are available. At Karolinska Institutet we often use this type of intersection to single out publications in Medicine and Life Science from all Web of Science records from other research institutions. This makes it possible to do better comparative analyses between us, as a medical university, and research institutions with more than one faculty.

The primary aim of the Karolinska Institutet/Region Stockholm bibliometric database is to furnish Karolinska Institutet/Region Stockholm managers, departments/clinics and employees with high quality bibliometric analyses and reports (regarding e.g. publication patterns, co-publication partners and/or by-subject breakdowns). The results provide overviews of the research being conducted by Karolinska Institutet/Region Stockholm researchers as well as tools that enable scientific results to be compared internally and with the work being done in other countries. Analyses results can also be used to follow up and plan activities, and sometimes even to assess and reward scientific quality.

The incentives from the board of research for publishing could be condensed to:

Before you publish, do a little background research into the journal to which you intend to submit your article; it should be included in the Web of Science index.
Check the journal’s impact factor and consider whether you can choose one with a higher value.
One high-cited article usually gives better returns on the bibliometric indicators than several low-cited articles.
Verify all your publications. It is essential to the quality of the database, and there is an explicit intention for the choice of bibliometric model to be such that the verification of all your publications has an aggregate positive effect.

Read the full recommendations.

Recommendations concerning the use of bibliometric indicators at an individual level

Within Karolinska Institutet and Region Stockholm there are recommendations for if and how bibliometric methods should be used with data for individual researchers.

Bibliometric methods are less suitable for the assessment of individuals or smaller groups. It is unusual for these to achieve a publication quantity sufficient for the results to be reliable and stable. It is also important that analyses methods do not create undesirable incentives for publication and verification behaviour, and one expressed intention with bibliometric analyses within Karolinska Institutet/Region Stockholm is that verification of a publication should never be counterproductive for an individual researcher. With good knowledge of the limitations existing at the level of the individual, certain bibliometric measures can also be used to supplement visual inspection of a publication list.

According to the EU General Data Protection Regulation (GDPR) we need to inform you on how we handle your data.

When you log in to the Karolinska bibliometric system you accept that we will retrieve and store information about you as specified below in the bibliometric database. Please click the box I agree if you understand and agree to the use of your personal data. You will have the right to cancel your consent at any time, and ask us to remove all information about you, by sending us an e-mail at kib@ki.se.

When you click I agree we will retrieve and store the following information about you from the KI/SLL staff databases in the bibliometric database:

User name (kilogin/hsaid)
Unique database identifier
First name
Last name
Civic registration number (Personnummer)
E-mail address
Organisational affiliation to department, hospital, patient/function area or clinic (where applicable)
Affiliation Type and Role (for example Employee, Graduate student, Research Assistant)
Current affiliation status
In addition the system may contain:

Information about your publications and where you were active when you wrote them
Several versions of your name
Your current field of research
Your ORCID
This information enables us to connect you and your organizational affiliations to your publications and include your verified publications in bibliometric analysis within Karolinska Institutet/Region Stockholm.

We will not share information retrieved from the staff databases with anyone outside the Karolinska Institutet/Region Stockholm organization. If you have any additional questions about our use of the information in the Karolinska Institutet/Region Stockholm staff catalogues, please contact us at kib@ki.se

Restricted profile

You can choose not to have the bibliometric system retrieve information about you from the staff catalogues, and still use the system. However, if you wish your verified publications to be included in bibliometric analyses results this may need some additional action on your part.

The system may still contain

Information about your publications and where you were active when you wrote them
Several versions of your name
Your current field of research
Your ORCID
If you wish your verified publications to be included in bibliometric analysis results, we can connect your account to either of the staff databases without your civic registration number, and retrieve necessary information (your name, organisational affiliation, affiliation type and affiliation role). If you wish to do so, please contact us at kib@ki.se and inform us on how you wish your account to be handled. The civic registration number is what enables us to match staff records between the Karolinska Institutet integration platform (KIIP) and the Region Stockholm staff database (EK) and create a single KI/SLL bibliometric verification account for each user. We can still give you a joint KI/SLL account in the Karolinska Institutet bibliometric database if you include both your KI-login and your HSA-id in your e-mail.

About the European General Data Protection Regulation
When you log in to the Karolinska bibliometric system (address: Karolinska Institutet, Fe 200, 17177 Stockholm, Sweden, organization number SE202 100 2973), you accept that we will retrieve and store information about you as specified below in the bibliometric database. According to the EU General Data Protection Regulation (GDPR) we need your consent to administer and process your data. You have the right to cancel your consent at any time, and ask us to remove all information about you, by sending us an e-mail at kib@ki.se.

Additional information about your rights regarding this can be found at GDPR.eu and Swedish Authority for Privacy Protection (in English).

Owner of the Karolinska Institutet bibliometric database is the research board of Karolinska Institutet. The Karolinska Institutet University Library (Berzelius väg 7B, Solna) is responsible for developing and running the bibliometric database. Non-licensed information in the bibliometric database is subject to the principle of public access to official documents and that information is available to the general public on request. You are entitled to a print out of the information about you that is stored in the Karolinska Institutet bibliometric system (free of charge). To receive this, send a signed, written request to Karolinska Institutet University Library, Att: Fredrik Persson, Fe 200, 171 77 Stockholm.

The Karolinska Institutet University Library will, upon request by the registered person or if an error is found by the University Library in the personal data processed take steps to correct, block, restrict or erase as soon as practicable such personal data as has not been processed in accordance with the European General Data Protection Regulation. (For example, this may relate to incorrect data.) You may report any errors to kib@ki.se who will put you in contact with the person/persons responsible for the primary data source of the incorrect data.

Bibliometric indicators are calculated for each publication individually. Indicators for an individual researcher will be shown when this person logs in to the Karolinska Institutet bibliometric database. An analysis based on field normalized indicators for an individual researcher with less than 50 publications cannot be ordered by someone else.

List of indicators available through the bibliometric system. (For a full list of bibliometric indicators and explanations to how they are calculated, see Bibliometric handbook.)

P: Number of publications identified with the current analyzed set of publications.
At Karolinska Institutet/Region Stockholm: Number of verified publications with a recognizable KI address (including university hospitals). Verified means that the publications in the Karolinska Institutet bibliometric database have both been identified by the researcher in question and that an organizational affiliation has been supplied by him/herself.
C: Total number of citations
The Total number of citations is the sum of citations (self citations included) from all publications present in the database to all verified publications for the analyzed set of publications.
c: Number of citations to a single publication
c_f:: Field Normalized Citation Score for a single publication
The Item Oriented Field Normalized Citation Score normalizes the number of citations to a single verified publication by comparing it to the mean number of citations to documents of the same type, published the same year, in the same research area.
Avg C_f: Field Normalized Citation Score Average
The average Field Normalized Citation Score is calculated on The Item Oriented Field Normalized Citation Score for all verified publications for the current analyzed set of publications. The world average is about 1, and an indicator of for example 1.2 means that the analyzed group of articles is cited 20% above the world average. (At present the standard method of using the ISI Journal Classifications to identify the research area is being used.)
Sum C_f: Total Field Normalized Citation Score
The sum of all the item oriented field normalized citation scores for all verified publications for the active selection.
P_f5%: Field Normalized Top Publications
The indicator Field Normalized Top Publications shows the number of verified publications for the active selection that belong to the 5% most cited publications in the world. The normalization is made by comparing the number of citations to each verified publication to the 95th percentile of citation distribution to all publications from the same year, in the same subject and of the same document type. (At present the standard method of using the ISI Journal Classifications to identify the research area is being used).
Share P_f5%: Field Normalized Share of Top Publications
The indicator Field Normalized Share of Top Publications shows the percentage of verified publications for the current analyzed set of publications that belong to the 5% most cited publications in the world. The normalization is made by comparing the number of citations to each verified publication to the 95th percentile of the average citations to all publications from the same year, in the same subject and of the same document type. (At present the standard method of using the ISI Journal Classifications to identify the research area is being used).
JIF: ISI Journal Impact Factor
The ISI Journal Impact Factor is a measure of the frequency with which the "average article" in a journal has been cited in a particular year. The ISI impact factor for a specific journal, one specific year (Y), is calculated by counting the number of citations to articles in that journal the two preceding years (Y-1 and Y-2) from publications in year Y and dividing this with the number of publications.
∑ JIF: Sum of ISI Journal Impact Factors
The sum of the ISI Journal Impact Factors connected to each verified article or review in the active selection.
Avg JIF: Average of ISI Journal Impact Factors
The average of the ISI Journal Impact Factors connected to each verified article or review in the active selection.

Note!

No field normalized indicators are calculated for too recent publications (with a publication year equal to the present year or last year)
publications belonging to fields with fewer than 30 publications per year
publications belonging to fields with an average citation rate lower than 0,2 cites.
for a subset of publications less than 50 publications.

Moreover, the compiled indicators are only based on publications with document type "Article" or "Review". This is to ensure statistical stability.

In the verification toolkit each verified publication with a not too recent publication year (see note above) also displays the value of the "Item Oriented Field Normalized Citation Score" for that individual publication, and will if it is one of the 5% most highly cited publications from that year in its field show a text indicating this.

It is possible, and desirable, for authors to verify all of their publications, also when written at other organizations than Karolinska Institutet/Region Stockholm. This is because two different types of analyses may be of value in different situations.

The internationally more common type of analysis, the “organization-based” analysis, uses knowledge about where the publications were written. It mirrors everything done by an organization during a certain time. This is the method of choice if one wishes to study the development over time of a specific research environment such as a university.

An organization based analysis of one individual researcher at Karolinska Institutet will include the publications that he/she wrote while active at Karolinska Institutet. At KKarolinska Institutet/Region Stockholm data for this type of analysis comes either from the authors' addresses as written on the publications or from the addresses that the authors supplied when they verified the publication in the bibliometric system.

An “author-based” analysis includes all publications written by researchers presently affiliated to the analyzed organization, regardless of where they were active when they wrote it. This more accurately mirrors the current research potential of the organization and is therefore the method used for many bibliometric analyses within Karolinska Institutet. At Karolinska Institutet/Region Stockholm data for this type of analysis comes from the authors' current affiliations according to the staff catalogues IDAC and EK.

The Karolinska Institutet bibliometric database is based on data from PubMed/Medline and three citation indexes from Web of Science Core Collection produced by Clarivate Analytics:

Science Citation Index Expanded (SCI-EXPANDED)
Social Science Citation index (SSCI)
Arts & Humanities Citation Index (A&HCI)

The following citation indexes are not included:

Conference Proceedings Citation Index - Science (CPCI-S)
Conference Proceedings Citation Index - Social Science & Humanities (CPCI-SSH)
Emerging Sources Citation Index (ESCI)

The KI bibliometric database is limited to records available in those databases from 1995 and forward.

For a copublication analysis we use information about who wrote the article and where they were active when they wrote it as a proxy for the research institution or country involved.

However, the quality of this data in the Web of Science is limited. Different people may publish under the same name, and some researchers (especially those with a double family name) may be published under many different name forms. Also the quality of addresses may bee of varying level with dozens of different versions of for example Karolinska Institutet.

Using the information in the addresses field, we can for instance look at co-publication at the country level. This kind of list is useful when assessing the level of international collaboration.

We can also look at the different authors in the publications and list the most frequent co-authors. Combining author and address information, we can also see who at Karolinska Institutet publish together with which organizations in one specific country or within any other subset of publications we can define, such as a subject field or a specific set of journals.

Using the information in the addresses field, we can look at co-publication at the organizational level, for example all KI publications together with Chinese organizations. This kind of analysis is useful for finding potential organizations for a formalized cooperation, or to evaluate already active cooperation agreements.

Within Karolinska Institutet we can use the higher data quality ensured by the verification process where we get to know the different namestrings to an individual researcher and the standardized affiliation of every publication in order to do copublication analysis with much more detail than otherwise possible.

There are presently two ways that we use to identify the research fields of individual articles. One is to use the Thomson Reuters journal subject field as a proxy for article content. This method is very often used in bibliometric studies. But in medicine, many of the publications also have MeSH terms, the controlled vocabulary from Medline. These are much more specific and are often used at Karolinska Institutet.

An analysis of the journal categories of a unit's/person's publications gives an indication of which areas the analyzed unit is active in. From this type of analysis, it is also possible to identify the type of audience that can be expected to read, and potentially cite, the publications. It can be used as the basis for a discussion about whether this is where you want to publish and be seen.

An analysis of the MeSH terms is much more specific and can help you see how your publications are shown in an international database like Medline, ie what MeSH terms other researchers need to use to find your publications. If you are not familiar with the analyzed group, a MeSH term analysis may show what major and peripheral aspects of research the group is involved in.

Combined with information about individual authors and/or organizations we can produce maps that show what factions co-publish in which certain areas.

Your author placement is highlighted in your list of verified publications. However, there are some limitations when calculating the author placement:

The publication must be of the type "article" or "review"; only such types of publication are included in the distribution of funds.
If the publication is in PubMed but not in Web of Science, it must have Medline status. It is common for new publications in PubMed not yet to have such status.
Information about shared corresponding author and about shared first / last authorship, we base on the information available in Web of Science and PubMed. We do not make any manual adjustments regarding this.
Manual additions are currently made to Karolinska Institutet / Region Stockholm publications that lack information and are included in the distribution of funds from the year of publication 2013.

FAQ: Verification toolkit

Starting 23 February 2021, Karolinska Bibliometrics database includes open access status on many publications. The data is retrieved from Unpaywall, Open APC Sweden (in Swedish only) and Open APC.

The information on OA status is available to support individual researchers to comply with the open access policy at KI and to fulfil the various open access mandates of research funders. It is also used to follow up the general open access compliance at KI and as a tool in license negotiations with publishers. Currently the data on open access status is not used in resource allocations at KI.

Publications in your publication list in the bibliometrics database are marked with one of the symbols listed below, or no symbol. The symbol is visible to the far right of the publication.

When logged in to the bibliometrics database, you can access your publication list by clicking "XX verified publications" or the button "View/Export to the right of the link. The open access information is also available in the publications list in the "Analysis Toolkit".

If you notice any irregularities or faulty information, please contact Publication Support at the library.

Padlock symbols

Padlock	Open Access status	Explanatory note	Colour
	Gold	Article published with immediate open access in an open access journal.	Orange
	Hybrid	Article published with immediate open access in a subscription journal.	Orange
	Green	Manuscript published with immediate or delayed open access in an online open repository.	Green
No padlock	Unknown/ Bronze	Unknown open access status, incl. articles published with delayed open access (bronze).	No padlock
	Not OA/ Closed	Not open access.	Grey

This could be due to two things:

One of the duplicates might be a meeting abstract with an identical title to an original article. If this is the case, please verify both of these publications.
If it’s not a meeting abstract, it could be a real duplicate. This might happen if there is a duplicate in the database sources Web of Science or PubMed/Medline, or if the same article from Web of Science and PubMed/Medline has not been correctly matched together in the bibliometric database. To have this corrected, please contact the University Library by email, kib@ki.se.

ORCID (Open Researcher & Contributor ID) is a unique international researcher ID. The point of such an ID is for example to make it possible to distinguish two researchers with the same name from each other, and to find all the publications from one person even if the name is not written exactly the same on all articles.

ORCID is founded by initiative of universities, individual researchers, research grant instances and journals. It may be of use when you submit a paper or apply for a grant and need to identify yourself or possibly populate electronic forms with data from your ORCID profile. Publishers such as Elsevier, Wiley och Springer Nature, and databases such as Web of Science and PubMed are using ORCID in various procedures.

You may add a biography or add a publication list to your ORCID profile. The primary goal of ORCID is however not to gather information about researchers but to provide a registry of unique researcher IDs that other services can use in order to minimize research administration.

An ORCID of your own may in the future give you a simpler work flow when applying for grants or sending manuscripts to publishers.

To return to your start page, click on your name which should be on the left-hand side of the main navigation links at the top of the page.
You can check and update your profile on your start page under the heading My profile. Among other things, you can add or remove namestrings there which will the be used to search the bibliometry database.
Our Help link is at the bottom of the page. The help page will open in a new window.
To search for your articles by PMID-number, click on the Find and Verify link on your start page, then click on the tab Find by PMID(s). NOTE: Your search is limited to the publications in the KI bibliometry database and therefore you may not retrieve results for all PMIDs.
To search for a publication by title, author, address, year or journal name, click on Find and Verify on your start page and then click on the tab Search.
After you click on Find and verify, select the publications you want to verify and click on the green Confirm button. The articles should now be visible in the next step: Supply addresses to your publications.
To remove publications from your List of suggestions that appears after you click on Find and verify, select the publications you want to remove and click on the red Unconfirm button.
In the step Supply addresses to your publications, choose the address where you wrote the respective article by marking the correct checkbox. Then select the relevant publication(s) and verify them by clicking on the green Submit all pending addresse(s) button. NOTE: verified publications cannot be removed. Address information cannot be removed either, however you can add additional addresses to a publication.
If you have confirmed an article by mistake, you can remove it in the step Supply addresses to your publications by selecting the article in the list Publication without address(es) and clicking on the red X icon to the right of the publication.
In the step Check if you're done a red exclamation mark will indicate if you have publications that still need to be connected to an address. Articles are not completely verified until they are connect to an address. In this stage, you can also see a summary of your status in the system.

MEDLINE is the greater part of PubMed and is limited to a more controlled subset. For example, newly published articles are available as temporary citations in PubMed before they are indexed and included in the MEDLINE subset. This process can take several months.

More information on the difference between PubMed and MEDLINE can be found on the NLM web page.

My profile shows the information about you that is known by the bibliometric system. Some is imported from the Karolinska Institutet and Region Stockholm staff catalogues IDAC and EK, but some can be entered or changed by you directly in the bibliometric system.

My namestrings

The My namestrings section shows the namestrings that we use to locate you as an author on the publications in the database. The primary use is to simplify for you to find your publications via the auto suggestion list on the first page, but sometimes they are also used in analyses to identify your place in the author list.

Your primary namestring is automatically generated from IDAC or EK. This cannot be deleted or changed by you and is for technical reasons not updated automatically even if changed in IDAC/EK, so if you do not publish under the name listed here you need to contact the university library to have it changed (kib@ki.se).

Add alternative namestrings

Under Add alternative namestrings you can add any other names that you've published under. This is for example advisable if you have a middle initial on some of your publications, if you have a double family name (with or without hyphen) or if you have changed your family name during your career as a researcher.

By default, searches based on namestrings are limited to publications with Sweden somewhere in the address field. You can remove this limit on namestrings if you wish (for example if you've published while at another university).

Research active at/My affiliation

This information is retrieved from the Karolinska Institutet and Region Stockholm staff catalogues IDAC and EK. If the information is incorrect you have to contact your local IDAC/EK administrator to get it changed. For Region Stockholm affiliations you can choose to uncheck your active affiliation(s) as research affiliations.

The Karolinska Institutet bibliometric database is based on data from PubMed/Medline and three citation indexes from Web of Science Core Collection produced by Clarivate Analytics:

Science Citation Index Expanded (SCI-EXPANDED)
Social Science Citation index (SSCI)
Arts & Humanities Citation Index (A&HCI)

The following citation indexes are not included:

Conference Proceedings Citation Index - Science (CPCI-S)
Conference Proceedings Citation Index - Social Science & Humanities (CPCI-SSH)
Emerging Sources Citation Index (ESCI)

The KI bibliometric database is limited to records available in those databases from 1995 and forward.

To make sure that the publication is not in the bibliometric database, try and search the publication with each and one of the three tools. If you still can't find it – is it a recent article? It may take anything between a couple of weeks to several months after publication until an article is indexed in the Web of Science, a fact that will be mirrored by this database. Records in PubMed are generally indexed much faster.

If you still cannot find your publication, contact us at kib@ki.se and let us help you!

The list of suggestions is autogenerated and based on the name strings in your user profile: primary namestring and alternative namestring(s). Since this may for some researchers with common names result in extremely many hits, the system filters out publications that don't have Sweden in the address information.

Click on Edit my names to add namestrings and/or uncheck the limit to Sweden.

Note: The list of names should only include varieties of your own name, not any of your coauthors, since this may affect analyses results.

This information is retrieved from the Karolinska Institutet and Region Stockholm staff catalogues KIIP/IDAC and EK.

Information in the bibliometric database about organizational affiliation is updated once a day which means it can take 24 hours for any change in KIIP/IDAC and EK to take effect. If you wish to make the change immediately, use the refresh button on the My affiliations page. Please note that for Region Stockholm, only affiliations defined by Region Stockholm FoU as research units are included.

If the error persists after update, please contact your local KIIP/IDAC or EK administrator to check that the information about you in those systems is correct.

If the information is incorrect in KIIP/IDAC and EK, please send an email to kib@ki.se and we'll help you find out what's wrong.

Yes, we recommend that you verify your publications of all document types. You do not have to worry that abstracts, corrections and other document types will affect your citation indicators, such as the field normalized citation score. Such indicators are only calculated on the document types Articles and Reviews, and in some cases also Letters and Notes.

In all bibliometric analyses at Karolinska Institutet we make certain that verifying a publication should not negatively affect the individual researchers results. One way of achieving this is for example to only use the five or fifteen publications with the highest indicator values when calculating averages.

The best way to know what a person's research is about is to ask the researcher her-/himself. Karolinska Institutet therefore asks you to supply information about your current research field according to the 2011 SCB/HSV standard. You can choose up to three fields, and rank them individually with "1" as your main field.

The information you supply can be used to improve analyses and reports of KI research, and may be used to create lists of researchers active in specific areas.

Information about the standard (in Swedish only) is available at the Statistics Sweden.

FAQ: Analysis toolkit

There are variations in citation counts, and other bibliometric indicators, between different databases. The main reason for this is coverage.

Google Scholar for example has one of the largest coverages, although their bibliometric quality is not the best and there may be not just peer reviewed journal articles behind those citations. There may also be duplicates.

The Karolinska Bibliometric System also has a different set of citation indices than the one currently available to KI in the Web of Science interface, (see What is included in the bibliometric system?). The major difference is that the bibliometric system does not include the index with sources that have not been fully evaluated yet, i.e. the Emerging Sources Citation Index.

As a rule, bibliometric analyses at KI are not used to analyse individual researchers or smaller groups (see How does my publication behavior affect my bibliometric analysis results?), so citation counts for individual articles as a rule have only small, if any, effect on bibliometric analyses results.

The term patient group is an organizational expression within the Karolinska Universitetssjukhuset theme organization.

A patient group is a group of patients with the same or similar diagnosis or symptom, for which a follow up on a group level is meaningful.

Each patient group is defined by a list of ICD codes. To be able to compile the research that has been done within a patient group area, a search profile is created based on the translation of these ICD codes to the MeSH terms (Medical Subject Headings) used to index material in PubMed.

Each patient group focuses on either adults or children (or rarely both). Since ICD codes and MeSH Profiles are usually not limited to specific age groups, you can in the bibliometric system separately limit you search results to, for example, either adults or children.

Options:

Children or adults
- No limit to age, this option includes all publications on the topic no matter if an age group is mentioned or not.
Children, exclude publications only about adults
- This option includes all publications except the ones that only talk about adults. This means that articles not discussing an age group will be included here.
Adults, exclude publications only about children
- This option includes all publications except the ones that only talk about children. This means that articles not discussing an age group will be included here.
Children - explicitly
- Publications that explicitly mention children.
Adults - explicitly
- Publications that explicitly mention adults.

Network graphs can be a powerful tool to visualize how for example researchers, institutions and research areas are connected through publications.

A network graph consists of 'nodes', representing different types of data, and 'edges', representing the connections between the data sets. Usually these are shown as circles and lines.

In our network visualizations, the nodes usually represent publications by for example a particular author or an organization. The size of the circle indicates the number of publications, so that the larger it is, the more publications by this author/organization has been found in the visualized dataset.

The line width indicates the number of connections between nodes, for example the number of copublications that a KI author has with a particular university.

Nodes in different data sets have different colours, so that authors are for example purple but the organizations are grey.

We mainly use three different tools for network visualization. Within the bibliometric analysis team we use Pajek and Gephi. For web visualizations we use the the JavaScript InfoVis Toolkit (c) 2012 Sencha Inc. - Author: Nicolas Garcia Belmonte (http://philogb.github.com/)

To be statistically valid, field normalized indicators need to be based on a minimum of 50 publications. If your selection has less than 50 verified publications for the analyzed period, no value for these indicators will be shown. If you choose to group the results by year you can analyze the trends of different indicators at the cost of precision.

The smaller the sample the more inaccurate is the calculated indicator, especially for field normalized indicators. For a sample size of 100 publications the yearly variation of the field normalized citation rate has been estimated to about 0.2 units for a stable research environment.

When you log in to the bibliometric system we collect information about the different affiliations you have in the two staff catalogues, IDAC (KI) and EK (Region Stockholm). The collected information is displayed on your start page at Verification Toolkit. You can only see analysis results for yourself as a person and you are restricted to viewing bibliometric aggregated data for your affiliated department/clinic/patient area or the entire organization KI or Region Stockholm.

In the Patient group research module you can select themes and patient groups defined by the Karolinska University Hospital as a part of the work with value based health care.

If you are a head/administrative head of a department at KI or a head of research and development at SLL (including heads of patient areas or function areas) you will also be presented

A boxplot graphically shows groups of numerical data through their five-number summaries: the smallest observation, lower quartile, median, upper quartile, and largest observation. The spacing between the different parts of the box helps to indicate the spread and skewness in the data. Here you can see a statistical summary of the indicators of your publications compared to those of your organisation.

The decentile bars show average indicator values divided into decentiles within your organisation. The bars consist of the same number of authors grouped according to their indicator value. The marked bar indicates the decentile where your own value is present. If the bar with the highest value is filled that means that your indicator value is among the 10% highest ranked researcher in KI according to the selected indicator and year span.

To reduce the effect of career age and sheer size of publication sample on bibliometric indicators for a researcher/constellation, we sometimes calculate some indicators based on the 5, 10 or 15 publications with the highest indicator values for that specific indicator. For example, the indicator Citation average (C/P) can be calculated on the most highly cited publications for each person.

In the bibliometric analysis toolkit the result show how highly cited your 5, 10 or 15 top publications are compared to top publications by other researchers at Karolinska Institutet / Region Stockholm. The purple bars show in which decentile you can be found based on the average number citations to your top publications. You can also see the average value for each decentile.

In the table of co-authors all unique name strings are considered as unique authors. If you add alternative name strings however, such as historical surnames, in your profile, they disappear in the list of co-publication authors. If you find a name string in the list that belongs to you, you should probably add that to your profile to get better automated guesses on publications to verify.

Bibliometrics at Karolinska Institutet and Region Stockholm

In 2005, Karolinska Institutet's management decided to use bibliometrics as a tool to become Europe's premier medical university by 2010. This decision was one aspect of the KI-05 initiative. KI's bibliometrics activities began in project form at the beginning of 2006. In 2008, the project became an operational unit, funded by the Board of Research.

The primary aim of the bibliometrics unit is to give Karolinska Institutet's management the opportunity to analyse the institute's research as a basis for organisational development. The analyses are to provide an overview of the research conducted at Karolinska Institutet and Region Stockholm, and an instrument to compare it with research from the rest of the world.

The secondary aim of the project is to provide other stakeholders within Karolinska Institutet and Region Stockholm with the opportunity to order bibliometric analyses and disseminate information to the outside world about the quality of Karolinska Institutet's academic production and collaboration.

Since 2009, Karolinska Institutet and Region Stockholm have had a common model for allocating direct research funding to departments and clinics on the basis of performance. Bibliometrics is one of the four components that are used to calculate the activity rate for each department, hospital and clinic.

If you want to learn more about the use of bibliometrics when allocating research funding at Karolinska Institutet and Region Stockholm, you can read this and in the following documents:

On resource allocation (Swedish only). The bibliometric model is described in the document KF budget och resursfördelning.
Detailed description of current calculations for the bibliometric component in the Karolinska Institutet/Region Stockholm resources allocation model (Swedish only).
Original description of the bibliometric component Karolinska Institutet/Region Stockholm resource allocation model (Swedish only).

Bibliometric database

The bibliometric database contains information about international academic publications from 1995 onwards. It is based on information from the database Web of Science that is owned by Clarivate Analytics.

In addition to data from Web of Science, the bibliometric database also contains data from a local copy of PubMed/Medline, also from 1995 onwards. The material from Medline (the main part of PubMed) includes the individual article's medical subject headings, known as MeSH terms.

Currently (2019), the bibliometric database contains complete bibliographic records for c. 39 million publications. In addition, it includes all citation connections between the articles' reference lists and corresponding articles in the database.

New records are obtained from Clarivate Analytics each week and the bibliometric database is judged to be, in principle, just as up to date as their own database. Via an internal website, Bibliometric Verification Toolkit, Karolinska Institutet's own researchers can go in and verify their own publication in the bibliometric database, increasing the quality of the analyses performed.

To search in the bibliometric database for articles written at Karolinska Institutet/Region Stockholm, there is a public interface available: KI Publications.

Publication behaviour with respect to analytical results

The primary aim of the Karolinska Institutet/Region Stockholm bibliometric database is to provide managers, wards/clinics and employees with high-quality bibliometric analyses and reports (on, for example, publication patterns, joint publication partners and/or analyses divided up by subject area).

The results provide an overview of the research conducted by researchers at Karolinska Institutet/Region Stockholm and is a tool that makes it possible to compare academic results internally and with work taking place in other countries. Analytical results can also be used to monitor and plan activities, and sometimes to assess and reward academic quality.

Read more about KI's recommendations under Strategic publishing.

Within Karolinska Institutet/Region Stockholm there are specific recommendations for the use of bibliometric indicators on an individual level. Bibliometric methods are less appropriate for assessing individual people or small groups. It is rare for these to achieve a publication volume that is sufficiently large for the results to be reliable and stable.

It is also important that the use of bibliometrics does not provide unwanted incentive for publication and verification. An explicit guideline for the bibliometric analyses conducted within Karolinska Institutet/Region Stockholm is that verification of publications should not be to the individuals disadvantage.

However, with good awareness of the limitations there are at the individual level, certain bibliometric measurements can be used as a complement to a visual inspection of one person's publication history.

Read recommendations concerning bibliometrics for individual researchers in their entirety.

Education and collaboration

Members of Karolinska Institutet's bibliometric team participate in meetings and courses in order to learn more about bibliometrics. The team wants to encourage discussion concerning the use of bibliometrics and of the advantages and disadvantages of different bibliometric methods.

Discussion and collaboration in the academic sector can lead to improvements in bibliometric methods. In turn, this can improve and refine the use of bibliometrics, both in the organisations participating and across the whole country.

Creative Commons BY SA

Bibliometric Handbook

Practical bibliometrics as an instrument for monitoring activity and as a basis for decision-making is a relatively new phenomenon. Previously, bibliometrics has primarily been used for research purposes within its own discipline. This means that the various indicators are not yet completely standardised.

Therefore, the bibliometric team at the University Library have written Bibliometric Handbook for Karolinska Institutet and the associated Bibliometric Indicators – definitions and usage for KI.

This document shows how the bibliometric team calculate different indicators, as well as their advantages and disadvantages. The handbook also contains a discussion of how bibliometric indicators and results from bibliometric investigations should be interpreted.