to them, it’s time to move on to the next phase in understanding NVivo’s system for adding demographic, categorical or scaled values to the cases. Usually you will know all the values you need ahead of time, but some of them might become evident to you only after you start to code the data. Sometimes, too, you will create a new attribute as you are working through your data, to categorize your cases on a variable that has become important to your analysis.
Understanding attributes and values
Whether we like it or not, our position in society and our membership in groups influence the way we think and act, and the kinds of experiences we have. An interviewee’s sex, class, nationality and religion may singly, or in combination, colour what he or she says in response to an interviewer’s questions. Within an organization, it matters what role or position one has, and perhaps how much education or training, or how many years of experience. And at a personal level, attitudes, behaviours or experiences may relate to one’s work history, education, family responsibilities or health. To record these kinds of demographic variables in a project becomes important, therefore, so that their impact can be assessed. In NVivo, the particular values one has on each of these (e.g., Education = secondary) are referred to as the attribute values of a case (see Figure 6.2).
If you are accustomed to working with statistics, you would think of Sex (or Gender) as a variable, and male and female as the possible values for this variable.
ATTRIBUTE CASES ALUES Case Gender V Barbara female Charles male Dorothy female Helen female Ken male
In NVivo, we refer to attributes and values instead of variables and values.
An attribute is the same thing as a variable.
Attributes record information known about a case, whether it is specifically mentioned in the course of conversation; collected purposefully with a check list or through a survey; ‘given’ by virtue of their location; or perhaps contained within archival records. While attributes are routinely used to record demo- graphic data, you can use them also to record any of the following types of data:
categorized responses to fixed-alternative questions such as those found in surveys – for example, ‘often’, ‘sometimes’ or ‘never’, in response to a structured question about experience of harassment;
categorical data you might generate in the course of analysing your data – for example, whether the interviewee who is caring for her mother did or did not mention getting help from other family;
scaled responses on instruments designed to measure attitudes or experience – for example, a visual analogue scale measuring level of pain experienced, level of attach- ment to community, or a score from a standardized inventory;
characteristics of a site or organization, where sites or organizations (rather than people) are cases.
Having attributes attached to cases makes them especially useful in an NVivo project as a tool for comparing subgroups and for filtering data. For example, you might want to compare the opinions or the experience of males and females, of leaders and followers, of locals and new residents, and so on. If you recorded these kinds of attributes with the cases, then the comparison becomes a straightforward task using NVivo’s query processes or visualization tools. Similarly, you can use the values of an attribute to filter cases. This would allow you to (a) run a query only on data from females; or (b) to find out whether opinions on real estate devel- opment are associated with attitudes to the natural environment, filtered for locals; and (c) then compare the result with a similar query, filtered for new residents.
When you attach an attribute to a case, as we indicated earlier in Chapter 3, you flag the entire swath of data within the case with the
characteristic of being female or male (or the values of the attribute
you’re attaching). This is what allows you to compare what females
say about tourism, or trust, or loyalty with what males say about
those things.
Creating case nodes and adding attribute values to the cases has several consequences:
Attributes apply to all of the data in a case – they cannot be added to part of a
case.1
Any further data added to the case node will automatically acquire the attributes of the case.
1 This means if you want to add values to only one of three documents
within a case – such as whether it was the first, second or third interview – this will need to be done by creating a set of documents with that
characteristic (explained later in this chapter). Another alternative (which we are less inclined to recommend) is to add source attri- butes to your system and assign these values directly to your sources.
Policy Ratitication date Organization Funding source Age male female 30s 40s 50s Person Gender
Any coding you apply to passages within a source (e.g., trust, love, honour) intersects with any case-level coding (e.g., Barbara, Charles, Dorothy) and thereby with the attribute values on that case.
For now, the point is that once you create cases, you’ll be able to add attribute values to these cases with the long-term goal of sorting the cases, and the data stored at them, by attribute values. In the discussion above, we were using the example of people as cases (and comparing attribute values such as male and female), but keep in mind that your cases might be policies, groups, sites, or critical incidents; the types of cases you study and the kinds of attributes you associate with them are entirely up to you.
Understanding classifications
We suggested, above, that you might have more than one type of case within a particular project (e.g., schools, classes and pupils as three case types in one project; sites and individuals as two in another; or people and artefacts in yet another). Each of these case types will usually require different attributes. NVivo’s classification system is what makes it possible to attach unique attrib- utes to different types of cases.
Thus, a Person might have attributes of gender and age;
Organization might have an attribute of funding source and Policy
might have an attribute of ratifica- tion date. Gender will never apply to Organization or Policy, nor ratification date to Persons. Person,
Organization and Policy are examples of classifications that you
might have in NVivo. Figure 6.3 presents a structural summary of how this would work for these example classifications.
Figure 6.3 The structure of classifications, attributes and values
Attributes Classifications
Values
cases, classifications, and comparisons 131
Values are always associated with attributes. Attributes are associated with clas- sifications. So, to be able to create a location in the database where you can store the possible value of female, you create a classification first, and then an attribute, with values. If you have more than one case type in your project, you might find it helpful to plan out what your classification and attribute system might look like by creating a diagram like the one in Figure 6.3 on paper, or in the NVivo modeller.