Dr. Phillip Smith
A Referencing Style Guide
This guide was written by Peter Coxhead and hosted on the School of Computer Science's website for many years. It is kept alive here.
Contents
- Introduction
- References in the text (citations)
- List of references
- Referencing online sources
- Footnotes
- See also
1 Introduction
Correct and consistent use of a standard referencing convention is essential in producing a report, thesis or paper. One version of what is often called the "Harvard system" is briefly explained below. Different authors (and different journals) use variants of the basic system; what matters most is consistency.
This is not a guide to avoiding plagiarism, but proper referencing is one of the main ways for students to avoid being accused of cheating through improperly copying other people's work. So learn to reference fully and properly!
Referencing a source involves two separate steps:
- indicating in the body of a piece of work that some material is not entirely original, by providing a short 'identifier' for its source (here called a reference in the text although it's often called a citation)
- listing, in a separate section of the work, the full details of the source (in a list of references).
In the Harvard system:
- A reference in the text or citation consists of a name – e.g. one or more surnames or the name of an organization – and a date, e.g. "Smith (2005)", "Wilson & Patel 2007" or "United Nations (1948)". (Other systems involve numbers, e.g. "[12]" or "12", or invented identifiers, e.g. "[Smi2005]". These are not the Harvard system.)
- The list of references is sorted by name (including initials if appropriate) and date. (Other systems may sort by the numerical or alphabetic order of the references in the text.)
2 References in the text (citations)
There are two ways to include a reference in the text.
-
If the reference name naturally forms part of a sentence, then include it exactly as if no reference is being given. Follow the name by a space and the full year of publication, enclosed in parentheses. (Multiple authorship is dealt with in the same way.) Examples:
Carson
(1970) argued that ...
The system developed by Brown & Smith (1986) is ...
The declaration of human rights published by the United Nations (1948) was ...
AI has been effective as Hamza (1983) claims ...
On the other hand, Jones et al. (1988) have reported that ...
Carson's (1970) paper argues ...The last example is not universally acceptable; the possessive can always be avoided by appropriate re-phrasing, e.g.:
A paper by Carson (1970) argues ...If the parenthesized year is crossed out, the remaining sentence, including punctuation, must still be correct. Thus the following are wrong:
An earlier paper Carson (1970) states that ...
It has been claimed that in this area AI has been effective, Hamza (1983).Multiple references by the same author(s) can be dealt with by placing a list of years in parentheses. Examples:
Jones (1980, 1983, 1987) has repeatedly argued that ...
The system developed by Brown & Smith (1986, 1988) is ... -
If the reference name does not form a natural part of the sentence, include both the name and the year, enclosed in parentheses, at an appropriate point in the sentence. (A comma can be placed between the name and the year, but this must be done consistently.) Examples:
An earlier paper (Carson 1970) argues that ...
The ARGA program (Brown & Smith 1986) is ...
It has been claimed that AI is effective (Hamza 1983) ...
The Universal Declaration of Human Rights (United Nations 1948) specified ...
On the other hand, it has been reported (Jones et al. 1983) that ..If the parenthesized name and year are crossed out, the remaining sentence, including punctuation, must still be correct. Thus the following are wrong:
An earlier paper by (Carson 1970) argues that ...
As claimed in (Hamza 1983), ...In the style of referencing in which formats such as [Hamza83] are used, it would be correct to write something like "As claimed in [Hamza83], ...". In the Harvard style, it is not. Don't mix up the two styles. If you decide to use the Harvard style, use it consistently.
Multiple references can be placed in a single list, using semi-colons to separate different authors:
This point has been made a number of times (Jones 1980, 1983a, 1987; Brown & Smith 1986; Carson 1970), but ...
Note the following points:
I prefer to join two names by "&" rather than "and". An advantage of this is that sentences like:
Both Brown & Smith (1986) and Carson (1970) argue ...are clearer because the "&" and the "and" are distinct.
If there are two authors, both names should be given in the text. If there are three or more authors, it is permissible (but not obligatory) to use only the first name followed by "et al.". Note that "et al." is usually italicized. (In hand-written or typed scripts, underlining may be consistently substituted for italics.) It is an abbreviation for "et alia",[1] meaning "and others"; the full stop must not be omitted.
-
If a reference would be ambiguous because it has the same author(s) and year as another reference, lower-case letters ("a", "b", etc.) are placed immediately after the year, e.g. "(Jones 1983a, 1983b)".
-
It is often desirable to pin the reference down more tightly – particularly when the source is a book. Page numbers can be placed after the date, preceded by "p." for "page" or "pp." for "pages". Examples:
This algorithm is taken
from Jones (1980, pp. 32-3).
One possible definition (Brown 1980, p. 12) is ...An alternative to "p." or "pp." which is sometimes found is the use of a colon. Thus instead of "Jones (1980, p. 12)", you may see "Jones (1980:12)" (with or without a space between the colon and the number). A slight disadvantage of this style is that you can't always be sure that the number after the colon refers to a page. If you include the "p.", then "Jones (1980, p. 3)" is clearly distinguished from, for example, "Jones (1980, section 3)". With just the colon, you can't be absolutely sure what is meant. If you do use a colon, use it consistently.
For some further notes on referencing online material, see Section 4 below.
3 List of references
There is a distinction between 'References' and 'Bibliography'. References are those sources actually referred to in the text. If the reference "Jones (1980)" occurs in the text, there must be a full description of it in the list of references. Short of blatant plagiarism, there are few more serious academic sins than the floating reference! If there is also a bibliography, then it lists those sources which were consulted and found relevant, but are not actually referred to by name in the text. If you use a combined bibliography and list of references (not ideal), make sure this is clear to the reader.
A single list of references should be given at the end of the thesis or paper – never one per chapter. The list must be ordered by name, then year. For this reason the year is usually placed immediately after the name, as it makes ordering simpler (although putting the year last is also logical and conforms more closely to library cataloguing systems). Multiple authors, however many there are, must never be reduced to "et al." in the list of references, even though they were in the body of the text.
It is not easy to give an exhaustive list of the possible formats for a full reference; variation in style is common, but the twin goals must always be completeness and consistency. 'Completeness' means that from the reference alone (i.e. with no other knowledge of the subject), the reader can locate the original source or obtain a copy. Some examples are given below. (Online material is discussed separately in Section 4 below.)
-
A complete book.
Jones, P.J. (1980),
Introduction to Algorithms, London: Methuen.
Jones, P.J., Smith, R. & Watson,
E.P. (eds) (1988), Artificial Intelligence Reconsidered (2nd edition), New York: Wiley.The title of a published book is italicized (or underlined, but only if italics are not available). Convention says that the place of publication as well as the publisher must be given, although with a large multinational publishing company, the former is less relevant. If more than one place of publication is listed, use the first. Only the publisher's name is needed, not the full legal title of the company, e.g. "Wiley" rather than "John Wiley & Sons". The ISBN is sometimes included; this is a unique identifier for a book.
-
A chapter in an edited book.
Hamza, K.A. (1988), "Vision Systems", in Jones, P.J., Smith, R. & Watson, E.P. (eds), Artificial Intelligence
Reconsidered (2nd edition), New York: Wiley, pp. 12-34.The chapter title is not italicized. The book title is italicized (or underlined where italics are not available). Page numbers should be given (not merely the number of the chapter). The "pp." is redundant, and hence sometimes omitted, but avoids any possible ambiguity. If several chapters from a book are referenced, it is permissible to include the book in the references list in its own right, and then use a format such as:
Hamza, K.A. (1988), "Vision Systems", in Jones et al. 1988, pp. 12-34. -
An article in a journal.
Carson, P.R. (1970),
"An Approach to Intelligent Planning", Journal of Applied Artificial Intelligence 38(3), 4-11.The title of the journal, which is the title of the published entity in which the article can be found, is italicized. The journal title should not be abbreviated, unless this is insisted on by an editor. If in doubt, don't abbreviate. The volume number is in bold (or underlined, but only if bold is not available). The part number, if relevant, is placed in parentheses, to make absolutely sure that it is distinguished from the volume and page numbers. Note that the first and last pages should be given. The main thing to avoid is something like "Journal of Applied Artificial Intelligence 3, 4, 10"; is this Volume 3, Part 4, starting at page 10, or Volume 3, pages 4 to 10, or what?
To avoid possible ambiguity, I prefer not to abbreviate page ranges, i.e. I prefer to write "234-239" rather than "234-9". This has become more relevant with the rise of electronic journals which sometimes have unusual ways of indicating 'page' numbers.
-
An article from a bound volume of conference proceedings.
Jones,
P.J. (1983), "An Attempt to Construct a Knowledge-based Route Planner",
Proceedings of the 8th International Conference on Transport Planning,
Budapest: Hungarian Transport Association, pp. 212-58.The volume is treated as a published book – hence its title is italicized. A problem with more obscure conferences is that sometimes the proceedings are just 'put together', and may not have a clear 'publisher'.
-
An article you have obtained which was read at conference whose proceedings seem not to have been published.
Jones,
P.J., Richards, M., Zhao, C.H. & Reynolds, P.E. (1988), "The Use of
BASIC in AI", paper read to the annual meeting of the Norwegian Association
for Computer Education, August, Oslo, Norway.This kind of 'reference' is irritating (although occasionally necessary), since it is not clear how to get hold of a copy of the article. The assumption is presumably that the reader could approach the organization responsible for the conference, but if the proceedings haven't been published (as is made clear by the absence of a title in italics), this is not likely to be very successful.
-
A report produced in 'duplicated' form by an academic or research institution. Unless the institution's address is readily available, details should be given; the principle is that a reader should be in a position to send for a copy of the report.
Jones,
P.J. (1987), "An Algorithm for Distributed Intelligent Route Planning with
a BASIC Implementation", Report #32, Intelligent Transport Laboratories,
38 West Avenue, Forked Springs, California, USA.
Referencing online material is discussed in the next section.
4 Referencing online sources
Traditional referencing systems were developed for paper-based sources. Adapting them to online sources can sometimes be a problem.
Some documents appear both in printed form and on the web, in which case the printed form should treated as primary, although the URL can usefully be given as additional information in the list of references, for example by adding a note of the form "[online at URL, accessed FULL_DATE]".
Where a work is only published on the web, if the author and date of 'publication' can be found (e.g. from the document itself or by using commands like 'Page Info' in Firefox), then the author's name and the date can be used as a reference in the text in the normal way. For example, using this document as the source:
Coxhead (2009) states that ...
The entry in the list of references should give a title for the document (e.g. the title of the web page) together with the URL and the full date on which the URL was accessed. This latter is particularly important since online material is transient; the date of last access gives some indication of the likelihood of the information still being at the quoted URL. For example, this document could be referenced as:
Coxhead,
P. (2009), "A Referencing Style Guide", http://www.cs.bham.ac.uk/~pxc/refs/refs.html [accessed 17 Oct 2009].
Since a University department is a clearly defined entity, it could also be treated as the 'publisher' and a more conventional style format used:
Coxhead,
P. (2009), "A Referencing Style Guide", School of Computer Science, University of Birmingham, Birmingham B15 2TT, UK [online at http://www.cs.bham.ac.uk/~pxc/refs/refs.html, accessed 17 Oct 2009].
With web pages, it is often necessary to use the name of an organization instead of the name of the author (e.g. "Sun Microsystems (2007c)").
Wikipedia is not a primary source (although often a good provider of references to original sources). Wikipedia articles are mainly useful as overviews; see, for example:
Wikipedia (2013), "Parenthetical referencing", http://en.wikipedia.org/wiki/Parenthetical_referencing [accessed 5 Jul 2013].
Finding a date of 'publication' can be a problem, since dates given in the document are sometimes not up-to-date, and 'Page Info' commands will simply return today's date for web pages generated dynamically by the server. One possibility is to use the date on which the URL was last visited; this at least gives some information to the reader. In the last resort "n.d." (for "no date") can be used.
Where information needed to construct a proper reference is simply not available, it may be better to give up, and just include the URL in the text or as a footnote.