A guide to Concordancing in SketchEngine, Part One
This first section explains how to use the concordance function in SketchEngine to search for collocations and see examples of language in use in BAWE. Click on any heading to open that section.
You can also download and/or print the guide in the following formats:
1 Go to the free corpora in Sketch Engine: https://app.sketchengine.eu/#open
2 Select the BAWE corpus
3 From the list of tools, select Concordance
1 Type your search term on the line
2 Click the red Search button
SketchEngine contains far more information in its concordance results than tools like Flax. The information is presented as:
1 In the top-left you can see the frequency of the search term. This is the raw frequency (how many times the word occurs in the corpus), and the relative frequency (how many times it occurs per million words)
2 On the left of each line you can see the type of text the concordance line comes from (eg essay, case study, research report). You can also click on the information symbol to find out more about the source text, e.g. the title of the writing, the number of words, etc.
3 The key word is presented in the classic KWIC (key word in context) view, ie the key word highlighted and in the centre of all the lines. The KWIC is also known as the node. You can click on any KWIC to expand the context a little further, but please note that it is impossible to see the full essay.
The words to the left of the KWIC are referred to as the left context.
The words to the right of the KWIC are referred to as the right context.
In the example search below we can make the following observations:
- The KWIC is claim as both a noun and in various verb forms, e.g claims, claiming
- The left context tells us that when claim is a verb, it is preceded by names e.g. Bulmer, Coleman, Maguire. When claim is a noun, it is preceded by modifiers e.g. scientific, moderate, this, these, their own. The verb form claiming is preceded by a comma
- The right context tells us that claim as a verb can be followed by that
Concordance lines are presented in the order the texts were uploaded to the corpus. This can often skew results towards one particular discipline or text type. To get a potentially more representative sample, you can click on the shuffle sample (two arrows crossed diagonally) in the top right.
This does not reduce the number of concordance lines, but takes them out of chronological order.
It’s worth noting here that SketchEngine is programmed to always produce the same results for shuffling, so if you do a search and shuffle, then repeat the search and shuffle with students, you should get the same results.
Rather than deal with 1000s of concordance lines, you can use the sample option (a question mark on a die) and specify a random sample of fewer lines. Then click Go.
As with the option to shuffle, sample will produce the same results for a particular set of search parameters, meaning you can replicate the steps with students at a later stage and get reliable results.
Although we can sometimes spot patterns immediately from concordance lines, it becomes easier when the data is sorted. To sort the concordance by the left context:
1 Click left context at the top of the concordance. This alphabetizes the concordance by the word to the left of the KWIC.
2 Move between pages of the concordance by using the arrows beneath it.
3 You can increase the number of rows visible per page if desired by adjusting the number in the rows per page field.
4 Click the yellow arrow in the top left next to the search criteria to jump to words beginning with a particular letter. This can be handy because punctuation marks are treated as tokens in the corpus, so you may need to skip several pages before you get to actual words before the node
5 You can also do this by clicking the Jump To option in the bottom left
You can follow similar steps for the KWIC (this sorts it by lower case to upper case) and the right context.
To copy and paste results into a document:
1 Tick the boxes in the lines you want to copy. If you want to copy all the lines on the page, use the box at the very top left
2 Choose copy from the dialogue box
3 Paste the results into a document by using CTRL + V
You can download the concordance results by clicking the download arrow in the top right.
We recommend downloading current view as pdf, rather than the whole concordance as a spreadsheet file, as it’s much more student-friendly
If you want to limit the pdf to certain lines
1 Select the lines you want to share with students
2 Click the show only selected option at the foot of the page. This then removes unwanted lines.
To share concordance results with students electronically simply click in the address bar of your browser and copy the link there.
Please note this is easiest if you are using freely accessible corpora like BAWE without logging in. If you share results you have garnered after logging in with an institutional account, students will also be required to log in to see the results.
A guide to Concordancing in SketchEngine, Part Two
This second section explains how to use the more advanced concordance search functions in SketchEngine, allowing you to look for specific parts of speech or word forms, search within specific parts of the corpus and filter results according to the words the KWIC co-occurs with.
You can also download and/or print the guide in the following formats:
Although the basic search we saw in the previous guide can be useful, SketchEngine’s concordance tool allows us to be much more specific in what we search for. To search for specific parts of speech or word forms:
1 In the concordance tool, click Advanced
2 Select your query type from the list on the left. Lemma will search for all forms of a specific part of speech, eg claim, claims, claiming, claimed. Word will look for all instances of a specific word, eg claims.
3 Choose the part of speech you wish to search for from the list on the right
4 Type your search term on the line and click Enter on your keyboard.
Results for lemma claim (verb)
Results for word claims (verb)
It is possible to refine your search so that SketchEngine only looks in certain areas of the corpus. To do this:
1 Fill out the advanced search fields
2 Click Text types
3 Choose the drop-down menus to specify where you want the results to come from. This can be by grade (distinction or merit), level of studies, discipline area etc.
One handy feature when searching for the kind of writing our students might produce is is the ability to limit your search results to writers whose first language is not English. To activate this, click on the Subcorpus field and choose All_except_English_as_author_first_language from the list
Note on text types
If you narrow down your search with too many filters, you will get a ‘Nothing found’ message. For example, searching for the term claim in critiques written by third-year L1 Arabic students in Anthropology (4 filters).
To see how words co-occur with another lemma:
1 type your search term
2 click Filter context
3 Select Lemma context
4 Type your second search term
5 Use the drop-down menus to determine how many positions to the left or right of the word you want to look at
6 Click Go.
The example search below shows time and at, filtered for up to 5 positions to the right
To see how words co-occur with a particular part of speech:
1 type your search term
2 click Filter context
3 Select Part-of-speech context
4 Choose the part of speech from the list
5 Use the drop-down menus to determine how many positions to the left or right of the word you want to look at
6 Click Go
The example search below shows time filtered by prepositions up to 3 positions to the left
Notes on filter context
- You cannot use the lemma filter and part-of-speech filter at the same time.
- You can use the lemma filter to search for more than one word or exclude a word, but remember to select the correct option (all, any or none).
Sometimes you might want to change the criteria of a search without starting again from the beginning, eg to change the part of speech of a search term from noun to verb, or to adjust the co-occurrence range from 3 to 4 positions. However, using your browser back button to return to the search field tends to wipe your search criteria entirely. To avoid this:
1 Click the change criteria symbol (it looks like a magnifying glass and recycling logo combined ) in the top right menu
2 Adjust the criteria you want to change
3 Click Go.
