Systematic literature search requires you to organise and tackle the search process in a structured and preplanned manner. It demands careful consideration of your search terms, selection of databases, choice of search methods, and requires you to reflect on the search results obtained during the process.
With a systematic literature search, you have a greater chance of avoiding disparities and bias, and it enables you to identify gaps in the existing research. In this way you also minimise the risk of reproducing already existing research.
It is important that you document your searches during the process so that your searches are, in principle, reproducible.
Within some disciplines there is a tradition for working with systematic reviews, where there are very specific requirements as to how the literature searches should be carried out and the results reported. There may also be other types of reviews, that are relevant for your purposes. It is not always necessary to be so thorough however. If you follow the basic principles for the construction of structured and systematic searches, you can come a long way.
If you have any questions or concerns, feel free to contact the liason librarian associated with your field, who will be happy to assist you with your questions.
Alternatively, you are always welcome to contact your local library in AU Library.
Your search terms are crucial for determining what you find. So it is worth spending some time working on them. It is up to you to find the best and most relevant keywords to cover the different aspects of the topic you are working with in your research.
It's all about being creative and considering all possible synonyms, close synonyms, related terms, narrower/broader terms, antonyms, abbreviations, grammatical or linguistic variations of your keywords. This mapping of potentially relevant keywords will streamline the task of composing an appropriate search string, one that provides a good balance between breadth and relevance in your search results.
You can reformulate and group your search terms to fit the various aspects of your research question:
Within some fields of study terminologies can be inconsistent, and perhaps also changeable over time, so here you must be extra careful to build up a good pool of keywords which can be combined in different ways. This also applies to interdisciplinary topics and new areas of research that have not yet received a fixed terminology.
If your subject area has a relatively fixed and established terminology, it can often be helpful to check the thesauruses of the databases you are searching in, so you can verify that you are using the agreed form of your keywords.
Building block searches can help you to streamline and structure your work with search terms.
In order to streamline and structure your work with keywords, you can organize them into thematic groups, covering the various aspects within your topic:
It can help to create an overview of potential keywords because:
Depending on language and the various subject-related aspects of your current project, it may be necessary to search in a range of different databases. Here, organizing your keywords in groups can also help you identify where to search.
In the individual databases you combine your grouped keywords with an OR. Across the blocks you combine your different groupings with an AND. Example from the database Pubmed:
Truncation refers to the use of wildcard symbols, e.g. * or ?. Truncation symbols replace one or more letters and are most often placed at the end of your keywords. They can however also be placed in the middle of a keyword or, in some databases, at the beginning of a keyword.
Truncation is especially good for finding the various grammatical forms of search terms, as well as alternative spellings, e.g. English and American, and for transcriptions of texts in other alphabets, e.g. Cyrillic or Arabic, etc.
You should use truncation with care. The location of the asterisk (*) has consequences. You could, for example, consider searching for the singular and plural forms independently instead. Then you will get exactly the word-end variations of the word you want. This can, in some databases, and within some subject areas (with precise specialist terminology), give better results.
Which characters you need to use, and how precisely they function will depend on the individual database. See under 'Help', or search tips, in the relevant base.
Masking is a specific form of truncation where you replace one single letter in a keyword with a masking symbol. Masking is usually used in the middle of a keyword:
Phrase search is used when you want to ensure the database searches for precisely the text you have written. You can do this by placing one or more words in quotation marks. Phrase search is particularly good for composite terms.
Here you will find only references which contain the precise term: "solar energy". A search without quotes will, in many databases, be treated as a search for solar AND energy, where both terms must be present but not necessarily stand together and in that order.
Proximity operators are used to indicate the maximum distance (in number of words) that can exist between your search terms. The use of proximity operators can provide you with more flexibility and control in your searches.
This will find references where the word cognitive appears a maximum of 2 words distance away from the word therapy. This means the search will also catch words and sentences such as ”cognitive behavioural therapy” and ”therapy with cognitive behavior treatment”.
Databases use Boolean operators to combine your search terms. They are used to broaden or narrow your searches.
The Boolean operators are: AND, OR, NOT.
AND narrows your search result. AND denotes the intersection - both your keywords must be present. Your search will become more precise than if you were to search for each keyword individually:
OR broadens your search. Or denotes the union - here just one (or both) of your keywords needs to be present. Your search result will become larger and broader:
NOT limits your search. NOT denotes the set difference. Here the keyword that follows this Boolean operator must not be present in the results. NOT filters out unwanted search results (noise), but you must use it with care, it is easy to filter out relevant results:
You can use multiple operators in the same search, for instance when working with building block searches, but if you are using both AND and OR in the same search string, then it is important to place brackets around the words that belong together in an OR block.
We recommend that you make simple searches with only one operator at a time. After which, you can then combine your search result sets into new combined searches.
Your keywords are crucial in determining what you find. It is therefore worth spending time on finding the best keywords. In many databases, you need to consider whether you want to carry out a full-text search, or whether you want to search using controlled subject terms.
Controlled subject terms are the "pre-selected" keywords the database owners have attached to the individual references to denote what a document "is about".
The controlled subject terms are selected from a controlled vocabulary. They ensure that you can create consistent subject searches, as references on the same subject can then be found with the precise same keyword, irrespective of the wording in titles and abstracts.
The controlled subject terms vary from database to database, and you cannot expect a search term used as a controlled keyword in one database to be found as a controlled subject term in another database.
The list of controlled subject terms may be hierarchically organised into a thesaurus. Here you can choose between more generalized broader terms (BT) or more specific narrower terms (NT). Thesauri are mainly found in the fields of medicine, psychology and the natural sciences, and education.
When you do a full-text search, you search for words in titles, abstracts, notes and keywords, etc. It is therefore important to find all possible and relevant keywords, as well as variations thereof. In other words you need to be creative and consider all possible synonyms, close synonyms, related terms, narrower/broader terms, antonyms, abbreviations, grammatical or linguistic variations of your keywords.
It is important, especially when searching across disciplines that you do not lock yourself in to the terminology of your own particular field, but also "translate" your terms to those found in other subject terminologies.
When you make a free text search it can vary as to which fields the individual database actually incorporates into the search. Some databases include the reference list, or the full text of the publication. In order to specify a search, you can use field searching. When doing a keyword search you could, for example, narrow your search to only search through titles or abstracts. Example from the database Pubmed:
This gives a search result where the phrase "patient education" appears either in the title or the abstract.
Field searches can also be used to scan through some very specific fields such as ”Geographic Terms”, ”CompanyEntity” or ”Conference location”, thus minimising non-relevant results in your search.
If you need to find "everything" on a particular subject, for a systematic review for instance, then you will need to use both controlled subject terms and full-text search. But in other cases there may be a good argument for choosing one or the other approach, depending on what you are working with. Generally speaking, you can say:
It is a good idea to think about your inclusion and exclusion criteria while you are developing your searches. Which parts of the literature would you like to include, and which would you like to avoid?
Consider, among other things:
The criteria associated with your topic must be reflected in your search terms.
Depending on how clearly defined your inclusion and exclusion criteria are many of the databases offer tools that can help you filter your search based on various formal criteria, such as year of publication, language, document type and peer-review status.
In some cases it may be more appropriate to make your own filtering after your first broad, rough set of search results, so that relevant references are not missed. Rough searches tend to go no further than assessing the relevance of the results simply on the basis of the titles and abstracts returned. In this way, a large number of references can be reviewed relatively quickly.
Almost all databases provide the opportunity to select relevant results from a search set and save them and/or export them to a reference management tool.
When you have created an overview of your keywords, you then need to select which databases, you want to search in. There are differences in the various types of search tool and database. As a starting point it is always best to search in several databases that can supplement each other:
The strength of the bibliographic databases is their focus on the subject area, rather than where it is physically located and what form it may have. You probably already know important databases within your subject area, but it can be a good idea to check the coverage provided by the individual databases, so that you are aware if the database has a bias towards, for example, American literature over European, or journal literature over monographs? Therefore check, for example:
Also consider whether you want to search in one database at a time or whether you can make a combined search in several databases at once.
See which databases, you have access to from Aarhus University.
See selected databases and resources for your sucject area.
A combined search, where you search across multiple databases with multidisciplinary coverage can have its own advantages, especially if your work is itself multidisciplinary in nature.
Combined searches can be performed in the large database hosts such as Ebsco and ProQuest, where you can easily select the individual databases that you want to include in your search.
Combined searches can also be very good in the initial phases of your searching in order to find the best databases for your subject area.
Google, and particularly Google Scholar, is used by many researchers as a first stop when making literature searches. In the initial phases this can be fine. However, we cannot recommend Google Scholar for systematic literature searches for the following reasons:
The search mechanism behind Google Scholar can, however, be very useful when it comes to finding similar articles and finding several versions of the same document (in those cases where this is relevant). You can also use Google Scholar to find full-text versions of articles, either via AU Library's subscriptions or as Open Access.
Google Scholar is a good tool for verifying references, and you could potentially use it as supplement to citation searches in Web of Science and Scopus. Often you get different citations in Google Scholar than in Web of Science and Scopus (which database provides most hits will vary according to subject area and the specific keywords chosen).
Read more about Google Scholar:
How much detail you should use when documenting your searches can vary. Generally speaking, we recommend that the documentation, as a minimum, includes:
It can be helpful to save your searches in the different databases, so that you can constantly keep track of, and retrieve, your previous searches. The databases also allow you to create alerts, so that you can be notified when new material that matches your search criteria is published.
Read more about pre-defined protocols, reporting standards, checklists for quality assessments, etc. in relation with systematic reviews. Read about other types of review.
Read more about how to get alerts on articles, new editions of journals, as well as new results for previously performed searches.