Once you have run your search on the databases identified in your protocol, you need to screen the results. Screening is a two-part process in which you determine whether each individual article meets your inclusion criteria, and therefore should be included in your review.
To reduce bias, you must have a minimum of two reviewers to screen - yourself and someone else from your review team.
How do I screen?
After you've finished creating your search strategy, save the final version in each database that you identified in your protocol. Once you have them all ready to go, run your search on each database and export all of your results to EndNote.
In EndNote, use a separate group to store the results from each database. Keep your EndNote library safe and backed up, as you will need to refer back to it for full text and your PRISMA flowchart (reporting).
The Two Stages of Screening:
First pass (Title / Abstract)
Second pass (Full text)
|This is where you examine titles and abstracts to remove obviously irrelevant material. At this stage you may not need to provide a justification for your exclusions.||This is where you examine the full text for compliance with your eligibility criteria. At this stage you must provide reasons why you exclude documents.|
Having trouble tracking down some of your references for the full text review stage?
Always check using Library Collection search
If 'Find it' (or similar) links from a database or the 'Find Full Text' function in EndNote do not locate the full text, directly search the UniSA Library Collection.
Journal articles (including meeting abstracts published in journals)
If there is no result in Collection search for the article/meeting abstract title, search on the full name of the journal to find any journal level records. There isn't always a record at the article/meeting abstract level. For example:
If the journal level record/s aren't obvious, refine the Resource Type to Journals.
There may be multiple journal records if the journal has changed name, or if the Library has access in various ways. Look for a record with the correct journal name and then the right 'coverage' - does the record indicate access to the right year, volume and issue?
Journals: print holdings
No online access? The Library may have the relevant journal issue (usually off campus) in print. In this case Request the appropriate volume to be sent to your campus of choice for pickup. For example, you would need to request the print volume to access a 1997 article from the Australian Journal for Science and Medicine in Sport:
Journals: online access
If the View Online access links in the journal record indicate the right year coverage, click through and browse for your reference's year, volume and issue. This is especially useful for finding references to meeting abstracts, which are often in supplements (with an S in the reference) and can sometimes only be found by looking for the right page range.
If the View Online access links indicate the right coverage, but browsing does not find the publication, it may be that the stated access is out of date. If you Request an interlibrary loan add a note to say the Collection record is inaccurate.
In focus: meeting abstracts
Interesting abstract, but you need more detail about the study? You could try the following:
Finding the full text: other options
A search engine or Google Scholar search may find the publication freely available online, often via an institutional repository or academic network such as ResearchGate.
The Library can usually arrange access to publications outside UniSA's collection for reviews - request an interlibrary loan.
Speed up access to full text via UniSA Library with LibKey Nomad and Google Scholar Library Links
Add the LibKey Nomad extension to your browser to see a quick PDF download button where it matches with a UniSA Library holding - this is especially helpful when visiting publisher pages off campus.
You can also configure your Google Scholar Settings > Library links to show Fulltext at UniSA. This can speed up accessing final published versions of works (the 'version of record') when using Google Scholar.