Can we conduct bibliometric analysis using 3 combined sources like PubMed, Science Direct and Google Scholar just similar to the PRISMA Guidelines of systematic review?
"wolf in sheep's clothing" or "a snake in the grass"
A simpler way to say is "A hidden enemy". Enemy of science is, Bibliometrics.
But let me start with this:
Three answers for you
1- NO, you can not combine or even compare these three "you mentioned". Science Direct is a biblio/fultext gateway database to its mother publisher, which is Elsevier, only, and they own everything. But PubMed is a borderless repository that covers publications from all of the world "but with their own criteria" they own nothing. Google Scholar (GS) is a semi citation search engines, with very open algorithms, collecting every thing the GS robots consider them scientific, again they own nothing. So you can not combine a database with a repository and a search engine, simply they don't have coherence, they are different entitles with different functions.
Before I go for the second answer it may be useful you have a look at these discussions
In this second part I say Yes you can do it, when they are compareable in terms of variables such as, coverage (for example they cover social science/s and/or humanities and/or Index, medical sciences... and many more) Like comparing CAB, AGRIS & AGRICOLA. Also you can compare them from the viewpoint of the usability, such as interface, inbuilt search capabilities...(below you find published examples)
Besides what I told you yesterday that bibliometrics now has become enemy of science and it gives us false picture of what is happening in the world of science/ research and academia. These are not up-to-date topics,
Before I open the third part and go through details about my relatively new findings regarding Metric studies (biblio, sciento, webo, infor, alt, METRICS), I suggest you read these from my publications regarding this part (2/3).
Article Evaluation and comparison of free and commercial user interf...
Article The Ability of The Web Search Engines in Answering General R...
is a very well-known researcher in "information science" all over the world. According to his Google Scholar page he has 851 publications *.
At least I have found two articles partially or completely devoted to review his publications. One is this:
Scientometric portrait of Mike Thelwall **
In this 2016 article, he has been seen with a traditional method of bibliometircs. that is finding and counting his published Items and the like.
But when we use "Inferential Scientometrics" we see unseen and hidden truth.
For example some 70 articles from him have been published in journal of scientometrics, where he was an editorial board member (at least up to my first waves of scientific criticisms to him in 2021 ***) and now marked in "Distinguished Reviewers Board" ****
Now, you compare Mike Thelwall (as an insider) published Items in Scientometrics journal, and published Items of a researcher that has No role or responsibility in that journal (as an outsider)!! then you understand better why you can not count on bibliometrics. You see the dark side of science.
These articles from all over the world tell you more about those people who work and have official roles in scientific journals, and why this phenomenon is unhealthy. (There are too many others search and read if you wish)
I Publish in I Edit? - Do Editorial Board Members of Urologic Journals Preferentially Publish Their Own Scientific Work?
10.1371/journal.pone.0083709
Editors publishing in their own journals: A systematic review of prevalence and a discussion of normative aspects
10.1002/leap.1449
Where and what do they publish? Editors’ and editorial board members’ affiliated institutions and the citation counts of their endogenous publications in the field of communication
10.1007/s11192-019-03169-x
Do scientific publications by editorial board members have shorter publication delays and then higher influence?