[<] [>]
[Cymraeg]

Physics projects for Y3 and Y4 students


Literature search

Literature search

At the beginning of week 4, you should submit a two-page document showing the strategy you have employed to locate refereed literature on your topic. This can be in tabular or graphic format, or a combination of the two. The purpose of this exercise is to ensure your literature research is as exhaustive as it needs to be without sidetracking you from the topic of your project. Use the document to discuss your literature effort with your supervisor. Only students on ph375 submit this as assessed work, but it's a worthwhile exercise anyway.

A tabular representation of your searches on Web of Knowledge could look as shown in the table:

no. query hits evaluation
1 "renewable energy" 1733 too broad - limit to one type at a time
2 "renewable energy" AND wind 322 still too broad - use likely keyword to narrow down
3 "renewable energy" AND "turbine*" 105 acceptable set
4 "renewable energy" AND solar ...
Phrases (2 words)Occurrences
of the
(rejected: not salient)
329
renewable energy
(rejected: search keyword)
167
wind farms44
wind speed29
life cycle27
capacity factor16
There is no need to list in your document
any phrases that you reject.
Word cloud generated from search 3.

You also need to analyse the abstracts for common keywords - these can be used to further refine your searches, but also to identify important concepts in your field. You can do this by highlighter pen on paper or by marking up keywords using text processing software. There are also online text analysis tools (such as this) or utilities producing a graphical representation of keyword frequencies. Any of these approaches are fine, as long as the results are interpreted carefully: The aim is to generate a list of relevant keywords relating to your field beyond those you've used already to define the search in the first place. You should look for individual words as well as short phrases of two or three consecutive words. The insert table shows a heavily truncated example of this. In any case, you need to read the majority of your abstracts - a compelling reason to take some time to devise a clever search strategy!

Venn diagram of topics covered by abstracts.

Try to group your abstracts into classes with a common sub-topic. You could represent this analysis as a Venn diagram with intersecting bubbles of different size, depending on how common a particular aspect is. This doesn't need to be pretty, but it must show the topical relationships between the abstracts you've found. Based on this, you should select a number of papers from each bubble, aiming at a broad overview of all aspects of the topic while going in depth in the most important areas closest to the topic of your own project. With a selection of perhaps 15-20 papers, which you may want to discuss with your supervisor, you can proceed to obtain the actual papers and start reading them for your literature review. You may have to come back to the search if the content of these papers suggests a new aspect that you hadn't considered previously.

Finally, provide a list of references of the papers whose abstracts you've chosen (see referencing for guidance on how to style the list properly) and indicate the location of these abstracts in your Venn diagram by marking up the diagram with citations.

Marking scheme

30%Sensible choice of initial keywords
30%Evidence of interactive refinement of keywords following initial search
40%Engagement with abstracts; sensible classification and choice for detailed study

In summary, your literature search document should comprise four sections:

  1. an overview of the searches with their keywords and hit numbers that you've used,
  2. an analysis of common keywords or topics in the abstracts, other than the keywords you've used yourself in the search,
  3. an analysis of correlations between topics in the abstracts, which enables you to select the most relevant papers to read.
  4. a list of abstracts you've chosen for detailed study in the literature review

All of this is based exclusively on the abstracts you have found. You will need to read the full papers corresponding to your chosen abstracts for the next step, the literature review.


Content updated: ruw/190613