What software does and does not do
and Christina Silver²
Affiliation: ¹University of Huddersfield and
Date written: 9th Dec 2005
Computer aided qualitative analysis data software is, as it says, an 'aid' in your analysis. It provides the tools but you have to do the thinking.
- Structure work - Enables access to all parts of your project immediately,
- 'Closeness to data' interactivity - Instant access to source data files (e.g transcripts),
- Explore data - Tools to search text for one word or a phrase,
- Code and Retrieve Functionality - create codes and retrieve the coded sections of text,
- Project Management and Data Organisation - Manage project and organise data,
- Search and interrogating the database - Search for relationships between codes,
- Writing tools - Memos, comments and annotations,
- Output - Reports to view a hard copy or export to another package.
This section is based on the paper: Lewins, A. and Silver, C. (2009) Choosing a CAQDAS Package - A working paper by Ann Lewins & Christina Silver. The full version is available online as a PDF at: http://caqdas.soc.surrey.ac.uk/PDF/2009ChoosingaCAQDASPackage.pdf
- Do the analytical thinking for you (though it can do things that help you do that thinking),
- Do the coding for you. In general, you need to decide what can be coded in what way. Some software supports automatic coding the results of text searches, but it is still important to check what has been automatically coded. One program, Qualrus, after you have coded a passage, makes suggestions about how else you might code it. The same warning applies here. It is up to you to decide if what is suggested by the program makes sense.
- Reduce bias, improve reliability or, on its own, improve the quality of your analysis (though it does have functions that can be used to help improve the quality of analysis),
- Tell you how to analyse your data. (No-one thinks a word-processor can write a report for them.)
- Does not calculate statistics, though some programs will produce simple counts and percentages
It is useful to think of the software as a type of database, filing cabinet or library catalogue where you can store items in meaningful categories and link them so they can be retrieved when required. In the software these categories are the codes and links and models you have created as you analyse the data. Once you have coded the data you can use search tools to explore the data and begin to identify themes and patterns.
- Understand what qualitative data analysis involves, See Intro QDA
- Understand your methodology or analytic approach and the processes involved, As a starting point look at the section on Methodologies for qualitative data analysis
- Know what you want the software to help you to do, See the software tools and How to use the software
- Attend a training workshop in you chosen CAQDAS package. For what is available, see the Training and support page
Organised and controllable data set. Many qualitative projects become very large and unwieldy, Not only is there a lot of primary data to deal with, but notes, memos, comments etc. increase the bulk of text you need to deal with. Add to this images, photos, diagrams and even video and audio recordings and the amount of material can become overwhelming. Keeping most of this in CAQDAS software and using the organising features like variables, sets, families, groups and, of course, coding will help keep it organised. On top of this, the software has tools, such as search, that will enable you to find things rapidly.
Support for coding. Most programs support the coding of text. This is a key tool for keeping a record of your analytic thoughts about the data and a way in which you can develop an analytical understanding and interpretation of your data. Most software not only has tools for coding, but also has ways of supporting the analytical development of codes using e.g. coding hierarchies, commenting, definitions, memos, grouping, modeling and networks and linkages and linkage types.
Searching Text and codes. This is one of the powerful tools that all software includes. There are two kinds of searches, searches for text and searches for codes. The first is like the search facility in a word processor, but often with much more capability. Some programs include facilities for searching for multiple words, Boolean links (and, not, or), phrases, wild cards, and even more complex expressions (using GREP http://www.anybrowser.org/bbedit/grep.shtml). In some programs this is combined with the facility to code the occurrences of the words that are found.
Searching for codes is actually searching using codes, since what is searched is the text that has been coded. This tool is a very useful one for looking for the relationships between coded passages. Thus you could look for text that is coded as, say, ‘illness’ that is also coded as ‘pain’ to find where people talk about pain associated with their illness.
Using software for searching in these ways is almost certainly faster (and more accurate and consistent) than looking through the text, by eye, to find certain terms, or retrieving by hand text coded in two different ways.
Support for comparative analysis. Of course the search tool is one good way of facilitating comparisons. Use it to look for other uses of terms in the text. But comparisons also means looking at other cases, other text coded the same way and other text coded similarly. Many programs allow you to define cases (if these aren’t the same as your documents) and you can then retrieve coded text for different cases in order to compare the results. In some programs the search tool allows you to search for coded text and cases to facilitate the same kind of comparison. All programs support retrieval of coded text (in a variety of ways). With medium to large data sets, using a computer to support such comparisons is a lot easier than dealing with a large number of pieces of paper.
Models, networks and diagrams. Many programs now have tools that enable you to produce charts, diagrams or networks and link the object in them to things like quotations, codes, documents and memos in your project. Objects in the diagrams can be linked (with lines arrows etc.) and these links can be given different meanings so that you can use the diagrams to construct models of your data. Such modelling can be a powerful way to develop you theories about your data. It is often easier to draw such diagrams on paper (especially if you need to learn how to use the software first), but the big advantage here is that the models and networks are linked into the data set and you can easily move back to the data on which the ideas are based to check that they make sense and are well supported.
Interface with quantitative data. Not all qualitative researchers will want to link their data to quantitative data. But for those who do, most programs contain tools so that you can import data from quantitative data sets and link it directly with relevant qualitative data in your project. Thus you might have a project examining the online conversations between learners in a virtual learning environment. If you also have some quantitative data about these learners such as their age, assessment marks, preferred learning styles, and so on them the software will allow you to import this information in such as way that both the messages sent by a student are linked with the quantitative data that relates to that student.
Some programs also allow the opposite procedure. That is, you can process the qualitative data in such a way that you create novel categorisations of cases or individuals. These can then be exported as quantitative data for integration into a quantitative analysis. For example, you might examine passages from online conversations that have all been coded as about collaboration. You might then identify 5 kinds of collaboration (from reading the text) and then export this as a variable for use alongside other quantitative data about the students.
Coding and the decontextualisation this produces. A common complaint about CAQDAS programs when they was first developed was that coding and retrieval with them tended to decontextualise the passages retrieved. Retrieval brings together all the text coded in the same way so it can be compared no mater where that text come from. It may therefore arise in different parts of people’s discussion and may also come from different people. Retrieval is done so that text on the same theme can be compared easily, but this is at the cost of seeing clearly the context in which that theme arose in the original talk. Recent versions of the programs have ameliorated this in a variety of ways. Generally they either make it very easy to move from the retrieved passage to the spot it came from in the original document or they support retrieval without removing the passage from its context.
Little support for complex mark ups Some approaches to qualitative analysis require the use of a range of symbols and marks to indicate things like pauses of different lengths, rising and falling intonation, overlapping speakers etc. In some cases these symbols and marks are not easy to include in a word processed document and hence in the rtf files required by most CAQDAS programs. In addition, most programs do not have ways of recognising the importance of these marks (though as characters you can search for them).
However, more significantly than this, is the fact that several approaches to analysis, such as conversation analysis, discourse analysis and narrative analysis do not really utilize a thematic approach. There is therefore much less incentive to use CAQDAS with its support for thematic coding. Add to this the fact that such approaches often use just a small number of cases and quite short texts which they examine very intensively, and the need for software in greatly diminished. Some researchers doing such analyses do use CAQDAS programs but only for preliminary, case selection stages before they focus on the more intensive conversation, discourse or narrative work.
Need to learn how to use the software (and need to buy it) For many analysts, who are new to QDA or have not used software before, this is a very significant consideration. For those on a tight budget, the software is not cheap (thought by specialist software standards it is not expensive either). So you have to allow for its purchase in research budgets. Then you have to learn how to use it. There are some books on the software and several general methods books have chapters on CAQDAS. In addition there are training courses you can go on (See Training and Support). And more and more, use of packages forms part of postgraduate training and universities are setting up labs with the software installed. But still, learning takes time and effort. You may also be faced with the decision of which program to purchase and use. The best advice here is to allow time for this decision. Look at the material on this web site to help you choose and ask colleagues and friends at other institutions and units what they have used. Remember, it is very useful if there is a local expert to whom you can go for advice when you are learning. So what they know well might sway your choice.