Literature Review and Technical Reading

Research Methodology

Prof. Jebran Khan

• Review of literature is one of the most important steps in the research process.
• New knowledge in research can only be interpreted within the context of what is already known, and cannot exist without the foundation of existing knowledge.
• It is an account of what is already known about a particular phenomenon.
• The main purpose of literature review is to convey to the readers about the work already done & the knowledge & ideas that have been already established on a particular topic of research.
• Literature review is a laborious task, but it is essential if the research process is to be successful.
• The review process must explain how a research item builds on another one because useful research should elucidate how and why certain technical development took place. Thus, it is easy for the reader to comprehend why the present research is being undertaken.

Literature Review

Definition of Literature Review

• According to university of Wisconsin writing center, A literature review is a “critical analysis of a segment of a published body of knowledge through summary, classification, and comparison of prior research studies, reviews of literature, and theoretical articles”.
• A literature review is an evaluative report of information found in the literature related to selected area of study. The review describes, summarizes, evaluates & clarifies this literature. It gives a theoretical base for the research & helps to determine the nature of research. …(Queensland University, 1999)
• A literature review is a body of text that aims to review the critical points of knowledge on a particular topic of research. …(ANM, 2000)
• A literature review is an account of what has been already established or published on a particular research topic by accredited scholars & researchers. …(University of Toronto, 2001)

Importance of Literature Review

• Identification of a research problem & development or refinement of research questions.
• Generation of useful research questions or projects/activities for the discipline.
• Orientation to what is known & not known about an area of inquiry to ascertain what research can best contribute to knowledge. Determination of any gaps or inconsistencies in a body of knowledge.
• Discovery of unanswered questions about subjects, concepts or problems.
• Determination of a need to replicate a prior study in different study settings or different samples or size or different study populations.
• Identification of relevant theoretical or conceptual framework for research problems.
• Identification or development of new or refined clinical interventions to test through empirical research.
• Description of the strengths & weaknesses of design/methods of inquiry & instruments used in earlier research work.
• Development of hypothesis to be tested in a research study.
• Helps in planning the methodology of present research study and in development of research instruments.
• Identification of suitable design & data collection methods for a research study.

Purpose of Literature Review

• The purpose of a literature review is to convey to the reader previous knowledge & facts established on a topic, & their strength & weakness.
• The literature review allows the reader to be updated with the state of research in a field & any contradictions that may exist with challenges findings of other research studies.
• It helps to develop research investigative tools & to improve research methodologies.
• It also provide the knowledge about the problems faced by the previous researchers’ while studying same topic.
• Besides enhancing researchers’ knowledge about the topic, writing a literature review helps to:
• Place each in the context of its contribution to the understanding of subject under review.
• Describe the relationship of each study to other research studies under consideration.
• Identify new ways to interpret & shed light on any gaps in previous research.

Purpose of Literature Review

• Resolve conflicts amongst seemingly contradictory previous studies.
• Identify areas of prior scholarship to prevent duplication of effort.
• Point a way forward for further research. See what has & has not been investigated.
• Develop general explanation for observed variations in a behavior or phenomenon.
• Identify potential relationship between concepts & to identify researchable hypothesis.
• Learn how others have defined & measured key concepts.
• Identify data sources that other researchers have used.
• Develop alternative research projects.
• Discover how a research project is related to the work of others. Place one’s original work (in case of thesis or dissertation) context of the existing literature.

Source of Literature Review

• Literature can be reviewed from two sources: 1. Primary sources and 2. Secondary sources
• Primary Sources: Regular paper / Research Article
• Literature review mostly relies on primary sources, i.e. research reports, which are description of studies written by researchers who conducted them.
• A primary sources is written by a person who developed the theory or conducted the research, or is the description of an investigation written by the person who conducted it.
• Most primary sources are found in published literature.
• For example: Strukov, D., Snider, G., Stewart, D. et al. The missing memristor found. Nature 453, 80–83 (2008).
• Secondary Sources: Review Paper / Review Article
• Secondary source research documents are description of studies prepared by someone other than the original researcher. They are written by people other than the individuals who developed the theory or conducted the research.
• The secondary sources may be used when primary sources are not available or if researchers want external opinions on an issue or problem or even the results of their own research.
• For example: Chua, L. Resistance switching memories are memristors. Appl. Phys. A 102, 765–783 (2011).

Main Sources of Literature

Electronic Sources of Literature

• Computer-assisted literature search has revolutionized the review of literature.
• These searches, however, for a variety of reasons may not provide the desired references.
• Electronic literature search through web may be very useful, but sometimes it can be time consuming & unpredictable because there are many website & web pages that can lead to information overload & confusion.
• Currently one of the most general literature search can be conducted through search engines like Google (www.google.com), Yahoo (www.yahoosearch.com), Bing, MSN search, Alta Vista, or Excite.
• The most relevant computing databases are: ACM Digital Library, IEEE Xplore Digital Library, dblp computer science bibliography, Microsoft Academic, Springer Lecture Notes in Computer Science along with Elsevier Digital Library for Computer Science.

Bibliographic Database: Web of Science

• The Web of Science is a website that provides subscription-based access to multiple databases that provide comprehensive citation data for many different academic disciplines.
• It allows the researcher to search in a particular topic of interest, which can be made by selection in fields that are available in drop down menu such as title, topic, author, address, etc.
• The tool also allows sorting by number of citations (highest to lowest), publication date.
• Put quotes around phrases, add more keywords, or use the “Refine Results” panel on the left to narrow down the search by keyword, phrases in quotation marks, type of material such as peer-reviewed journal articles, date, language, and more.
• Structured search like this that enables narrowing and refining what one is looking for is effective to ensure that the results throw up relevant sources and time spent in studying those is likely to be well utilized.
• Based on the researcher’s need the search result can be broadened or narrowed down using the built-in fields provided in this website.

Bibliographic Database: Google Scholar

• Google Scholar provides a simple way to broadly search for scholarly literature. You can search across many disciplines and sources: articles, theses, books, abstracts and court opinions, from academic publishers, professional societies, online repositories, universities and other web sites.
• Basic techniques of searching in Google Scholar:
• OR—Broadens search by capturing synonyms or variant spellings of a concept. Example: Synchronous OR asynchronous will find results that have either term present.
• Brackets/Parentheses ( )—Gather OR’d synonyms of a concept together, while combining them with another concept. Example: RAM (synchronous OR asynchronous).
• Quotation marks “ ”—Narrow the search by finding words together as a phrase, instead of separately. Example: RAM (synchronous OR asynchronous) “Texas Instruments”.
• Site—limits the search to results from a specific domain or website. This operator is helpful when searching specific websites such as the BC government, which is Example: RAM(synchronous OR asynchronous) “Texas Instruments.” site: http://ieeexplore.ieee.org.
• Filetype—limits the search to results with a specific file extension one could look for pdf’s, PowerPoint presentations, Excel spreadsheets, and so on. Example: RAM (synchronous OR asynchronous) “Texas Instruments.” site: http:// ieeexplore.ieee.org, filetype: pdf.

Effective Search: The Way Forward

• Scholarly publication is one wherein the published outcome is authored by researchers in a specific field of skill.
• Such work cites all source contents used and is generally peer reviewed for accuracy and validity before publication. Essentially, the audience for such works is fellow experts and students in the field.
• A researcher must consider what type of information is needed, and where it could be found.
• Research can utilize the keywords their respective field to search for the required articles.
• Searching is an iterative process:
• Experiment with different keywords and operators;
• Evaluate and assess results, use filters;
• Modify the search as needed; and
• When relevant articles are found, look at their citations and references.
• After the search is complete, the researcher needs to engage in critical and thorough reading, making observation of the salient points in those sources, and summarize the findings.

Steps of Literature Reviews

Steps of Literature Reviews

• Stage I – Annotated Bibliography:
• At this stage, researchers read articles, books & other types of literature related to the topic of research & write a brief critical synopsis of each review.
• After going through the reading list, researchers will have an annotation of each source of related literature.
• Later, annotations are likely to include more references of other work since previous readings will be available to compare, but at this point the important goal is to get accurate.
•  Stage II – Thematic Organization:
• At this stage, researchers try to find common themes of research topic & organize the literature under these themes, subthemes, or categories.
• Here, researchers try to organize literature under themes, which relate to each other & are arranged in a chronological manner.
• Researchers try to establish coherence between themes & literature discussed under theses themes.

Steps of Literature Reviews

• Stage III – More Reading:
• Based on the knowledge gained through primary reading, researchers have a better understanding about the research topic & the literature related to it.
• At this stage, researchers try to discover specific literature materials relevant to the field of study or research methodologies which are more relevant for their research.
• They look for more literature by those authors, on those methodologies, etc.
• Also, the researchers may be able to set aside some less relevant areas or articles which they pursued initially.
• They integrate the new readings into their literature review draft, reorganize themes, & read more.
• Stage IV – Write Individual Sections:
• At this stage, researchers start writing the literature under each thematic section by using previously collected draft of annotations.
• Here they organize the related articles under each theme by ensuring that every article is related to each other.

Steps of Literature Reviews

•  Stage IV – Write Individual Sections (cntd):
• Furthermore, related articles may be grouped together by ensuring the coherence between different segments of the literature abstracts.
• For each thematic sections, draft annotations are used (it is a good idea to reread the articles & revise annotations, especially the ones read initially) to write a section which discusses the articles relevant.
• While writing reviews, the researchers focus on the theme of that section, showing how the articles relate to each other & to the theme, rather than focusing on writing each individual article.
• The articles are used as evidence to support the critique of the theme rather than using the theme as an angle to discuss each article individually. 
• Stage V – Integrate Sections:
• In this section, researchers have a list of the thematic sections & they tie them together with an introduction, conclusion, & some additions & revisions in the sections to show how they relate to each other & to the overall theme. 

Technical Reading

Techniques of Technical Reading

• Finding the right work to read can be difficult because literature where knowledge is archived is very fragmented and there are bits and pieces all over the place. Very rarely will one find everything that one wants close together in one place.
• It is also important to know where to read from; relying on refereed journals and books published by reputed publishers is always better than relying on easily available random articles off the web.
• Given the abundance of journal articles, it is useful to adopt a quick, purposeful, and useful way of reading these manuscripts as its require multiple rereading to understand the concepts and one might expect to spend many hours reading these paper.
• Initial skimming through the manuscript is the best way to decide whether a manuscript is worth of careful reading or not.
• For initial skimming, read the title and keywords (these are anyways, probably what caught the initial attention in the first place). If it does not sufficiently seem to be interesting; it is better to stop reading and look for other articles.

Techniques of Technical Reading

• If the title and keywords seem interesting, one should then read the abstract to get an overview of the paper in minimum time.
• Again, if it does not seem sufficiently important to the field of study, one should go straight to the conclusions and if conclusion is not interesting then immediately stop reading further.
• If the abstract is of interest, one should skip most of the paper and go straight to the conclusions to find if the paper is relevant to the intended purpose, and if so, then one should read the figures, tables, and the captions therein, because these would not take much time but would provide a broad enough idea as to what was done in the paper.
• If the paper has continued to be of interest so far, then one is now ready to delve into the read of Introduction section to know the background information about the work and also to ascertain why the authors did that particular study and in what ways the paper furthers the state of the art.
• The next sections to read are the Results and Discussion sections which is really the heart of the paper.
• One should really read further sections like the Experimental Setup/Modeling (brain of the paper), etc., only if one is really interested and wishes to understand exactly what was done to better understand the meaning of the data and its interpretation.

Conceptualizing Research

• The characteristics of a research objective are that it must have new knowledge at the center, and that it must be accepted by the community of other researchers and recognized as significant.
• Besides being original and significant, a good research problem should also be solvable or achievable and reproducible.
• Therefore, the method and tools that could be used to obtain that new knowledge are quite significant and normally come from the existing recorded literature and knowledge in the field.
• Coming up with a good research objective, conceptualizing the research that meets all of these requirements is a tough thing to do.
• It means that one must already be aware of what is in the prior literatures. One needs to be continually reading the literature so as to bring together the three parts
• Significant problem,
• The knowledge that will address it, and
• A possible way to make that new knowledge or enhance the current knowledge.
• Incorporating these parts will be different for every person and for every field, but the only way to be that expert is by immersing oneself in the literature and knowing about what already exists in the field.

Conceptualizing Research

• If one is working on a research project that is of a smaller scope and didn’t provide much contributions to the scientific society, then conceptualizing the research is possibly too tough to do, and one does not have the time that it takes to become an expert at the edge of knowledge.
• In this case, the researcher needs the help of someone else, typically the supervisor who may already be an expert and an active researcher in that field. He may advise on what a good research objective might be.
• An established researcher in any field should be able to immediately point to the landmark literature that one should read first.
• Otherwise one would need to spend a lot of time reading the literature to discover.
• As engineers, we like to build things, and that’s good, but the objective of research is to make knowledge as well. So, it is important for the researcher to ask if new knowledge is being formulated from their research or not!
• Even if one is building is new things that has never been built before, but it is something that any experienced and competent engineer could have come up with, then their work runs the risk of being labeled obvious and rejected as research.

Critical and Creative Reading

• Critical Reading:
• Reading a research paper is a critical process. Readers should not assume that the reported results or arguments are correct. Rather, being suspicious and asking appropriate questions is in fact a good thing.
• Have the authors attempted to solve the right problem?
• Are there simpler solutions that have not been considered?
• What are the limitations (both stated and ignored) of the solution and are there any missing links?
• Are the assumptions that were made reasonable?
• Is there a logical flow to the paper or is there a flaw (both practical and theoretical) in the reasoning?
• Use of judgmental approach and boldness to make judgments is needed while reading. Flexibility to discard previous erroneous judgments is also critical.
• It is also important to ascertain whether the data presented in the paper is right data to substantiate the argument of the paper and whether the data was gathered and interpreted in a correct manner.

Critical and Creative Reading

• Creative reading:
• Creatively is harder, and requires a positive approach in search.
• In creative reading, the idea is to actively look for other applications, interesting generalizations, or extended work which the authors might have missed.
• Are there plausible modifications that may throw up important practical challenges?
• One might be able to decipher properly if one would like to start researching an extended part of this work, and what should be the immediate next aspect to focus upon.

Taking Notes While Reading

• This proverb applies to researchers who need to read and build their research based on the prior knowledge. They need to write building on the notes taken.
• Researchers take notes of the papers/articles manually or digitally.
• Each research paper, there are a lot of things that one might like to highlight for later use such as definitions, explanations, concepts, comparison, mathematical equation or explanation, and so on.
• If there are questions of criticisms, these need to be written down so as to avoid being forgotten later on.
• Such efforts pay significantly when one has to go back and reread the same content after a long time.
• On completing a thorough reading, a good technical reading should end with a summary of the paper in a few sentences describing the contributions.
• But to elucidate the technical merit, the paper needs to be looked at from comparative perspective with respect to existing works in that specific area.

The faintest writing is better than the best memory

Reading Mathematics and Algorithms

• Mathematics is often the foundation of new advances, for evolution and development of engineering research and practice.
• An engineering researcher generally cannot avoid mathematical derivations or proofs as part of research work.
• In fact, these are the heart of any technical paper. Therefore, one should avoid skimming them.
• By meticulous reading of the proofs or algorithms, after having identified the relevance of the paper, one can develop sound understanding about the problem that the authors have attempted to solve. Moreover, one can try to modify the proofs or algorithms to achieve better results for their own works.
• Nonetheless, one might skim a technical section if it seems like an explanation of something already known, or if it is too advanced for the research at the present moment and needs additional reading to be understandable, or if it seems to specialized and unlikely to be needed in the course of the research program in which case one can get back to it later on.
• Implementation of an intricate algorithm in programming languages such as C, C++ or Java is prone to errors. And even if the researcher is confident about the paper in hand, and thinks that the algorithm will work, there is a fair chance that it will not work at all.
• So one may wish to code it quickly to check if it actually works.

Reading a Database and Datasheet

• Databases are the collection of structured information which provide the knowledge about the stored information and also an instruction manuals of that corresponding information. Ex. http://yann.lecun.com/exdb/mnist/, https://cocodataset.org/#home, http://surfing.ai/speech-products.html.
• One needs to find out every details of the dataset of their corresponding research from the database.
• Datasheets are instruction manuals of the specific model, which provides detail information about the corresponding models. Datasheets enable a researcher to reconstruct the models or debug any given simulation with that specific models.
• Usually the first page of the datasheet summarizes the functionality of the model and its features, basic specifications, and usually provides a functional block diagram with the internal functions of the part.
• The objective of the authors herein has been to use databases and datasheets as an example to state the need to pay attention to the art of reading such documents.