How to research quickly a state of the art – Practical guide

Every scientific oeuvre starts with a state of the art. A state of the art is a text that summarizes the key concepts on the topic, explains the current status of the subject, and discuss the latest advances and challenges in the field.

This text is the starting point of the arguments that follow. It establishes a common ground between the writer and the reader. It is like an updated map of the subject that serves both the author to know how to organize the ideas and the reader to understand the context.

Create such a map is very useful when you are learning a new topic. You need a map to orient yourself. How the topic looks like. Its structure, variants, extension. Once you grab the main structure of the topic and its possibilities, you can lose yourself in the details wherever part catches more your attention.

I am talking about finding a few key resources that capture the essence of a topic and used them to sketch a mind map to orient yourself. This strategy is helpful in many situations. During learning, your mind map helps you to understand where each piece of information fits in the overall picture. When you are discussing a complex topic, your mind map puts your arguments into perspective with respect to the arguments of the rest of the people in the conversation.

Educate yourself quickly about a specific topic from the source is a useful skill. For example, assess the state of the art of a new technology or summarize the key arguments on a subject.

Scientific subjects can be very dense and difficult to follow. Having clear the basics makes a whole difference when diving into the topic.

So, how can we build this map in practice?

Search

Let assume that you are already familiarized with the basic terms of the topic of interest. You can express the subject in your own words and explain the key ideas to someone else. You aren’t novice anymore. You are ready to read SCIENCE, with capital letters.

At this point, you want to read scientific literature to create your own deep understanding of the topic. You need a list of few keywords on the subject. Don’t worry if your list is not too extensive, with 4 or 5 terms should be enough to start. As soon as you make your first queries, you will fine-tune those terms with related keywords.

If you are unfamiliar with scientific databases and you don’t know which one to use, I suggest starting with Semantic Scholar. It is a general database with many cool features enhancing the literature skim.

For the first query start with my 3-terms-formula

  1. Top 1 keyword of your topic. It can be a single word or a multi-word term
  2. Niche or application. This part is optional. It depends on how specific your initial keyword is, but in most cases, the top keyword is too broad.
  3. Review. Add the word “review” to get review articles. Review articles will give you a broader view of your topic and they contextualize the rest of the publications in the field.

Let’s imagine that you want to learn the state of the art of artificial intelligence applications in the healthcare industry. Your first keyword would be “artificial intelligence” and the second keyword or the niche “healthcare” and finish with “review”.

Within the first 10 results, you can already see two interesting articles. You can continue the search with different angles, playing with related terms like “machine learning” or “medicine”, and see what comes out.

Once you find a good keyword combination, you can move to the searching options. In Semantic Scholar there is an option where you can sort by relevance, citation count, most influential, and recency. I would play with these options to see how the results vary.

If after these tweaks you still haven’t found a perfect match you can narrow it down by filtering the field of study and publication type.

Try also removing the word “review” to see if there is an article that hit the bull’s eye.

Select

The scientific literature is vast and there is more information than we will ever be able to read. So, be very selective especially with your first articles

Select between 5 to 10 articles. Even if you think you are missing something by leaving out certain papers, don’t worry. Each review article will give you tons of references to read.

Chose a minimum of 2 articles from different authors. Although keywords are similar between authors there are always analogous terms for the same concept. Capturing this variability will help you to understand better the topic.

Look for the most relevant publications in the last two to five years. This is a rule of thumb. It depends a lot on the field. For fast-paced topics, 3 years is already old material.

Recent articles are better, not only because the information is updated, but also because it is easier to go back in time than forward. Think that if you read a paper from 2016, all the studies cited in the text will be from 2015 or older. You are missing references to the most recent studies.

Read

Now, you are ready to go. Download the articles. Add them to your library. Read, highlight, tag.

This will take you at least a couple of days. This process is time-consuming: identifying relevant articles, selecting which ones are worth adding to your scientific library, and actually reading the articles in depth. So, the stricter you are during your selection the more meaningful/practical/productive your reading time will be later.

The good news is that you only have to do it once. After that, you will set up a system to keep you updated automatically. Because the purpose of building your own scientific library is to have a resourceful tool and not a burden.

Let’s recap for a moment. Out of our initial search of the state of the art, we should have identified a maximum of 10 articles. Most of them reviews. Preferably published in the last 2 years.

Again, do not be afraid of left interesting articles out of your initial top ten selection. If they are worth it, they will come back to you via a citation in context or in future searches with different search criteria.

Keep up to date

After reading all this new knowledge, you will know where you want to go next. From this point, you have two roads to follow. The active and the passive way.

The active way consists in look for all the references that have caught your attention during reading, plus additional searches of new relevant terms you have found out. This is the case when you want to know more about the subject immediately. In this way, you can go as far and as deep as you want.

The passive way is intended to keep you updated in long term. It consists in feed yourself automatically with relevant publications on a weekly basis. We will explore both methods in the next posts.

Conclusion

Those are just the first steps in outlining the main lines of a new knowledge map. Educating yourself and comprehending the current status of a complex subject can’t be done overnight, but with these steps, you can quickly reach a good point of understanding.

A state of the art consists in read a lot. What is important is not only to read but also to capture the information to use it later. In this blog, I discuss how can we do that in a practical and efficient way.

For now, start your scientific library following these steps in your topic of interest. In the next posts, we will see how to keep it updated with little effort and massive results.

Your turn

What topic you would like to research the state of the art?

Leave us your answer in the comments.

Scientific databases – which one should you use

We have already explained in this blog what types of scientific articles exist. Now, you may wonder – where I can find scientific articles? How do I know what has been published on my subject? Because googling doesn’t lead me to scientific studies.

Well, I have good news. Even on Google, there is a specific section for searching scientific studies. It is called google scholar. But there are many more resources other than google.

Scientific databases

Usually, the first time you arrive at a scientific study is by following the references of other texts. As we saw in where to start the research of a new scientific topic,  you follow a path from general sources to more specific sources.

Once you have pass the basic level, you have read a few sources and you want to gain perspective on the subject.

For that, you need to review the studies published on that specific topic. You need a tool to search scientific literature. This tool is a scientific database.

Scientific databases, or academic databases, not only encompass journal articles, but also conference proceedings, books, doctoral theses, and other scientific-related documents.

As mentioned above, we can use google scholar. Technically, it’s not a database, it’s a search engine that looks for information everywhere, including scientific databases. However, since it looks at the whole web, which is too vast, makes it difficult to rank on the first page the references you are looking for.

So, what are the alternatives?

The two biggest scientific databases

In the world of scientific databases, there are two big commercial actors: Scopus and Web of Science, previously known as Web of Knowledge.

These two are the biggest databases with over 70 million references each. The problem is that these databases are intended for institutions, and they are quite expensive. These are the databases you will find in the library of a university. In many cases, if you go in-person to the library of a university you can use one of these databases for free in the computers of the library even if you are not a student. So, if by chance you have the opportunity to access a university maybe it is worth passing over the library and search on these databases.

The problem with these databases is that you can’t use them unless you belong to an institution that pays for access. But they are so relevant, that you need to know about their existence.  The rest of the resources I am going to show you below are accessible for everyone, at least the bibliographical data. The documents themselves, in some cases, will be under a subscription fee or a paywall. In a future post, I will explain how to deal with that.

Free access general databases

An alternative to these multidisciplinary databases is JSTOR. Although access to most of the documents is not free, it is a great tool to find research papers.

JSTOR has a cool feature called text analyzer, currently in beta, which could be interesting in particular cases. You can paste a text or drag a document and the tool identifies the keywords of the text giving you back related documents.

A tool like that is interesting to trigger a serendipity search to find documents related to your specific topic of interest which initially don’t seem naturally connected to you. Serendipity accidents are considered to have a potential for innovation because connect ideas from different fields that are not related a priori. Let us a comment below with your experience using this tool.

Another great tool is Semantic Scholar. In this case, they use the power of artificial intelligence to extract meaning from scientific literature making the search more efficient than the traditional keyword search.

Most of the databases use the metadata defined by the authors to sort the document and make suggestions. However, Semantic Scholar extracts the topics discussed in the article directly from the text and suggest related studies base on it. It also differentiates the type of citation (highly influential, background, methods, results) which is very valuable information to get the ball rolling in the direction of your interest to find relevant references. Other features that have enormous potential are: filter by papers with pdf available and create feeds of latest papers by topic of interest (currently in beta).

Another alternative worth exploring is ScienceOpen with several features to enhance research discovery.

This is the nice part! Every tool has small tweaks that make them special. You have to find the one that works better for you.

Thematic databases

The thematic databases are the most practical resources to find relevant scientific articles because you are already looking within a specific area of knowledge reducing the ambiguity of queries that belongs to several fields.

For example, if you query “arm biomechanics” in a life science database you will retrieve studies of the mechanical aspects of the human arm. However, with the same query in an engineering database, you will retrieve articles on robotics devices. In a computer sciences database, you will get research on human-machine interaction. And, in social sciences databases, you will obtain papers on anthropology. So, by choosing the database you will already reduce the scope of your search, which is good.

For the sake of simplicity in the following list, I highlight the main scientific databases sorted by area of knowledge. All databases provide similar tools for search, filter, and create alerts. I invite you to explore all the possibilities of the database most relevant for your scientific library.

  • PubMed – medicine and life science
  • IEEE Xplore – engineering, technology, and computer science
  • arXiv – physics, mathematics, and computer science
  • dblp – computer science
  • SSRN – social sciences and humanity
  • EconBiz – economics
  • ERIC – education science

Open Access Repositories

All the resources mentioned above are great, and lead you to a ton of references worth reading. However, there is a small problem. In many cases, they lead you to a scientific article that is under payroll where you only have access to the abstract and the list of references but not the full text. This could be a bit frustrating especially at the beginning when you don’t have resources at your disposal like a university library.

One way to avoid this problem is by searching only on open access repositories. Open access is a type of publication where the reader has free access to the full text because the publisher covers the cost or the authors have paid a fee to publish open access. Fortunately for the world, the scientific publication business is moving in that direction and it is becoming more and more common to publish in open access format.

Again, I list below the main Open Access Repositories for you to explore the one that fits better your interests.

  • DOAJ – all disciplines
  • Paperity – all disciplines
  • CORE – all disciplines
  • OpenDOAR – all disciplines
  • Zenodo – all disciplines
  • BioMedCentral – medicine and life science
  • PLOS – medicine and life science
  • arXiv – physics, mathematics, and computer science
  • JURN – mainly arts and humanities but also business and law

Other resources – Gateways

Gateways are centralized access points to a network. In our case, this means, they are search tools that retrieve documents from specific databases.

For example, ScienceDirect is a well-known gateway to access all the documents published by Elsevier. Elsevier is indeed the largest scientific publisher, but it will not show you results from any other publishers which is quite limiting. You can find similar tools from other publishers like Springer or Wiley.

Another scientific gateway is science.gov, the search tool of the U.S. government on science information that offers scientific and technical information of the main U.S. federal agencies (NASA, EPA, NIH, US Department of Transportation, Department of Homeland Security among others). In turn, science.gov belongs to the WorldWideScience.org, which provides a similar service but for all the participant nations (more than 70 countries).

These national portals offer different possibilities than the pure scientific databases because they give access to additional science-related documents beyond peer-reviewed publications, like technical reports, internal research, surveys, patents, guidelines, etc, by country instead of by discipline.

Bonus line, if you are interested in this last point you may want to check The World Factbook.

Conclusion

All these resources are great, but you should not forget your principal goal. The purpose is not to know all the databases that exist but to find the one that works the best for you and create a system that keeps you updated on autopilot efficiently and effortlessly, as we will see in the next post.

Your turn

Which of the databases works the best for you? In which field are you interested? Do you use other resources to find literature?

Leave us your answer in the comments.

Types of scientific articles – structure and purpose

Not all scientific articles, aka papers, are the same. Each kind of article has a different purpose. For example, some papers present new data while others discuss data already published.

Understanding the differences and the purpose of each type of article makes dive into a topic more efficient. It also helps you to skim the articles much faster. An important skill to keep your scientific library as a practical tool avoiding been submerged in an ever-growing folder of articles pending to read.

There are two main types of scientific articles that you need to master, original research articles and review articles.

Original research articles

The original research article is the standard and most important type of article. It is the basic block of scientific communications and it is also the most common article you will find.

These articles are primary sources. That means it communicates original data generated by the authors. It could be data obtained by experimentation, observation, or theorization.

This type of article follows always the same structure:

  1. Introducction of the scientific question or hypothesis.
  2. Methodology employed to perform the study.
  3. Results of the study.
  4. Discussion of the results.

Review articles

The review articles, on the contrary, are secondary sources. Review articles do not present new data, but collect, reorganize, and summarize the existing literature with a new perspective.

This type of article is intended to give a comprehensive analysis of a specific topic by synthesizing the key points in the field. Some review articles go a step further and formulate new hypotheses based on the open questions of previous studies.

Review articles are generally written by authoritative researchers that understand deeply the topic. They string together all the ramifications of the field in a compressive way, proportionating a wide perspective of the topic.

For this reason, review articles are the most enlightening text to read when you are new in a field. I recommend reading review articles before reading original research articles because the review articles give you the context necessary to frame the relevance of a particular study.

Other types of articles

Without doubts, these two types of articles are the most important and they constitute the gross of the scientific literature. Nevertheless, there are other types of articles worth knowing.

Short scientific communications are quick synopses of preliminary results. Usually, they do not have enough results for a full paper, but a few data points worth sharing. Therefore, they are intended to stimulate discussion and proportionate the seeds for deeper studies.

Conference proceedings are abstract or short scientific communications that have been presented in a conference. The relevance of these articles depends on the field. In most fields, conference proceedings are not included in the formal categories of scientific literature. In part because the content of the proceedings ends up published in a formal paper within a more complete study.

However, in fast-changing fields, like computer vision, artificial intelligence, or cybersecurity among others, conference proceedings are more relevant because the advances happen so fast that the peer-reviewed process is too slow to cope with the new developments. These fields are so competitive and fast-paced that researchers do not bother to spend time writing a formal paper. They rather prefer to invest that time in continuing the research.

If you are interested in these fields, it is worth it to follow those conferences closely, since relying only on formal articles may keep you a little behind the latest results.

Editorials are articles expressing the author’s view about a particular issue relevant to the audience of the journal. Usually, criticize or discuss a publication accepted in the journal presenting briefly the subject with no aim of a full review.

Other resources

PhD thesis or dissertations are the results of a PhD student’s research. This text is very useful to dig into the details on how the research was carried on. Generally, in a thesis dissertation, you can find greater details of the methods than what you can read in the publications derived from that research. Usually, they also include lots of appendices with raw data.

Books or tertiary literature synthesize primary literature that has been around for a long period and is widely accepted in the research community. These are also great resources to get familiar with a topic and understand the basis of the field.

Other resources could be interesting to include in your scientific library depends on your topic of interest:

  • patents
  • best-practices guidelines
  • technical reports from governmental agencies, ONG’s or other institutions
  • official reports from public authorities, or private firms

All those documents are gold mines for you to build your knowledge database, to back up your arguments, and to create a solid base for your work. You have to explore what type of documents are more relevant for you.

Tell us

What type of article do you find more useful in your field?

Leave us your answer in the comments.