25
DRAFT: NOT FOR CITATION OR QUOTATION 1 Coding Texts using NVivo 10 James A. Anderson Callie Spencer Heather Stone Introduction A text can be considered as any assemblage of signs and symbols presented as a text through any medium or combination of media. This definition deliberately eliminates very little from the purview of textual analysis. It does lay some responsibilities on the analyst, however. A text needs to have its boundaries as well as the manner of its constitution as a text defined. And the manner by which it is a text "of something" and by which it connects to other texts needs to be described. The text has to a member of a body of texts because coding is inherently reductionistic depending on the commonalities a given text shares with other texts. Coding attempts to show that a set of characteristics can usefully describe something of the underlying forms, conventions, thematics, cultural values and the like of a body of texts. As a result it both reveals what is hidden in each individual text and reduces the surface of each text to these underlying features. What is presented as new knowledge is not the collection of texts but the codes warranted by the collection of texts. Deductive coding common in metric approaches uses a predetermined set of codes along with the rules of their attachment to content that has been derived from theory independent of the body of texts. The code book is in effect a hypothesis about the underlying features of the set of texts. When the codes meet standards of instrumentality and reliability—they work well—the hypothesis about the characteristics of the body of texts is supported. Emergent coding, common in interpretive and hybrid approaches, is centered in the argument that justifies the recognition of an aggregation of texts as a genre, set, body, collection of texts that does some shared, collective cultural work. This argument defends the claim that are common features within the aggregation that elevates the particular selection of cases to a congregation of meaningfulness and validates the utility of the coding work that will reveal those meaningful features. Our focus here is on emergent coding of which there are two major kinds along with multiple variations between them. The two kinds can be described as the "Discovery Method" of grounded theory and as the Axial Method" of grounded theory. The essential difference between the two is the role that existing theory plays in the coding process and in the final argument. In © JAA, 2012

Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

  • Upload
    others

  • View
    6

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

1

Coding  Texts  using  NVivo  10  James A. Anderson

Callie Spencer Heather Stone

Introduction  A text can be considered as any assemblage of signs and symbols presented as a text through any medium or combination of media. This definition deliberately eliminates very little from the purview of textual analysis. It does lay some responsibilities on the analyst, however. A text needs to have its boundaries as well as the manner of its constitution as a text defined. And the manner by which it is a text "of something" and by which it connects to other texts needs to be described. The text has to a member of a body of texts because coding is inherently reductionistic depending on the commonalities a given text shares with other texts. Coding attempts to show that a set of characteristics can usefully describe something of the underlying forms, conventions, thematics, cultural values and the like of a body of texts. As a result it both reveals what is hidden in each individual text and reduces the surface of each text to these underlying features. What is presented as new knowledge is not the collection of texts but the codes warranted by the collection of texts. Deductive coding common in metric approaches uses a predetermined set of codes along with the rules of their attachment to content that has been derived from theory independent of the body of texts. The code book is in effect a hypothesis about the underlying features of the set of texts. When the codes meet standards of instrumentality and reliability—they work well—the hypothesis about the characteristics of the body of texts is supported. Emergent coding, common in interpretive and hybrid approaches, is centered in the argument that justifies the recognition of an aggregation of texts as a genre, set, body, collection of texts that does some shared, collective cultural work. This argument defends the claim that are common features within the aggregation that elevates the particular selection of cases to a congregation of meaningfulness and validates the utility of the coding work that will reveal those meaningful features. Our focus here is on emergent coding of which there are two major kinds along with multiple variations between them. The two kinds can be described as the "Discovery Method" of grounded theory and as the Axial Method" of grounded theory. The essential difference between the two is the role that existing theory plays in the coding process and in the final argument. In

© JAA, 2012

Page 2: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

2

Defining Text Members of the coding course coalesced (more or less) on the following: A text is considered any assemblage of signs, symbols, or representations that intends to express semiotic value within a formal structure. A text has durability, a cultural name, material boundaries, and a more or less clear beginning, middle, and end. A text represents the confluence of an historicized constituting performance (author), medium, content, audience, and culture and is activated locally and culturally in an interpretation (of which coding is one such interpretation).

the discovery method, theory is formed and informed by the coding that finds its full justification in the cases. In the axial method, theory participates as an interpretive lens to direct one phase of the content coding. In both methods all work starts with the cases and whatever happens next depends on what emerges from the deep engagement of those cases. The work of emergent coding, then, starts with the successful claim for a body of texts that will be fruitful in understanding some cultural work. It proceeds to the process of the selection of actual cases that meet the terms of the body of texts argument.1 The evaluation of a potential study needs to stand on the success of the argument for a congregation of meaningfulness formed in a body of texts and in the adequate access to a set of cases within that congregation. Because the cases and the interpreter of those case form a unique combination of insight into a set of cultural practices, no uncertainty is reduced in the review of what went before. In old West parlance, one stakes a claim and works the rock for what is there.

Some  Distinctions  Let's begin by separating some terminology: text, source, case, unit of analysis, congregations, aggregations, collections, families, and constellations.

Units    and  unities  Text is the general term for the objects of our analysis (see definition). Texts form congregations across some lines of commonality; terms like genre, types, categories, body of works, and so forth are also used, but each has some technical limitations for our purposes. I use the term congregation of texts to distinguish a set of interconnected texts with communalities across properties, practices, consequences, and values from an aggregation of texts: An aggregation is what one sweeps off a driveway (metaphorically, of course). The only coagulating connections among the elements of an aggregate are external to the texts per se—time, place, and reference to the broom. And I use congregation to emphasis the meaningful membership of the text as a function of the text itself. Coding reveals the terms of this membership.

1 In actual publications these steps are often presented as "of course it's true" claims with little development. I believe that to be a fundamental weakness of scholarship in this area.

Page 3: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

3

A case is a specific expression of: • Facts

– Material properties of the text: – Expresses the voice of the process of text

production – Does material work

• Language/symbolic material in use – Identifies resources in play – Supplies rhetorical elements – Demonstrates emotional valences – Indicates meaningfulness and import – Does the material semiotic work

• Framing – Establishes cultural location – Sets requirements of the narrative and

discursive performance – Intentions of the form – Does structural work

• Narrative structure/argument form – Connects concepts to action – Establishes motive and justification – Creates tensions and resolutions – Uses paradigms and syntagms – Does moral work

• Discursive performance – Intentional performance of a cultural

location – Displays competence/incompetence – Entails the audience through rhetorical

force – Has modality (formal, ironic, comedic,

elegiac, etc.) – Does cultural work

Consider, for example, a ethnographic reconnaissance study of one of the Occupy protests. The researcher conducts interviews of the press, the protesters and the police. The interviews are transcribed and are scheduled for coding. The three sets of texts although ostensibly all interviews on the same topic belong to three different congregations of texts that have been aggregated by the researcher. They are three different congregations of text because their historicized moment of production were under the governance of different social meanings and cultural values. Yes, they all fall into the genre of interviews (which is why that term does not work), but interview process is just the aggregating action, because the language, logics, and narratives in use would be expected to be different. The code "anarchy," I would hold, would have very different referents across press, police and protesters A source is a nominated member of a congregation. Sources are rounded up in collections. Collections develop through some process of selection. The rigor of this process is a property of the collection that points to its trustworthiness. Cases are derived from sources. A case is a source that has gone through the classification (and possible extraction) process in use and been retained for coding. A case is an authenticated member of a congregation. Cases can also be created out of (extracted from) sources. In their study of introductions and conclusions Anderson and Colvin (2008) entered entire journal articles as their sources. They then coded the

Page 4: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

4

appropriate content as an introduction or conclusion leaving the rest of the source uncoded. An NVivo report was then run that extracted just the introductions and just the conclusions. Those textual fragments (called a collection) became the cases of their subsequent study. Cases form up in families. I am deliberately using membership and kinship terms for the exemplars that enter into our theorizing and coding to underline that we do not code randomness (in fact we routinely discard on-off codes). We code connections. Finally, in coding studies, but not usually rhetorical, critical, or hybrid studies, the case is further unitized into coding segments across some conceptual property (thought unit, argument, topic, frame) or formal property (paragraph, sentence, grammatical segment). These coding segments are the units of analysis. Units of analysis form up in code-specific constellations that reach across individual cases (that connection thing again).

Analytical  Benefits  This formulation of terms leads to some potentially profitable lines of inquiry: Coding studies generally start with an argument about some set of texts having scholarly significance because of the common or integrated cultural work the set produces. This argument establishes the boundaries of the focal congregation of texts. Those boundaries have to be sensible as boundaries that gather a fruitful set of texts that correspond to the purpose of the study. Those boundaries also describe the requirements of the text collection process. For example, let's say that I want to justify my use of "congregation" in talking about a coalescence of texts by nominating all texts that use any referent to the concept of congregation both as a noun and as a verb. The nomination is silly, of course. I would be overwhelmed by hundreds of texts that have no relevance to the issue and the demands of the collection would overwhelm any process I could put in place. I am clearly interested in journal publications of scholarly texts that address the coding process and use the concept of congregation as the term references a coalescence across properties, practices, consequences, and/or values.2 Why? Because this writing intends to be scholarly and only other scholarly texts can justify its arguments. The move from texts to sources is an equally profitable area for review. Texts reference any assemblage of intentionally semiotic content. Sources describe what we are actually going to look at. In the example above, the move was from all texts to journal articles—a set which would be further reduced by the specific collection process. As we have learned, the demands of coding are high; it is imperative to get to a collection of sources that can actually be managed in the duty cycle of publication and dissertations. Five years of coding is not a viable alternative. Consequently, a scholar like Shugart (2008) in Sumptuous Texts goes from talking about "many," "most," and "all" food films to coding three very fruitful ones (for her argument) in extended

2 I doubt that I would find any that would use the actual word "congregation." Too close to the scholarly taboo against terms with a religious background, but maybe Kenneth Burke in Dramatism and Development would provide an entry.

Page 5: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

5

interpretive descriptions.3 She states: "I argue that these films are unified not only insofar as they feature food but also, and more importantly, with respect to how they use food to engage and assuage anxieties attendant on contemporary cultural ambiguities and permeabilities, especially around race/ethnicity and gender" (p. 69) The move from sources to cases can also be a fruitful process for examination. Not all sources become cases and not all cases are equal. It is unusual to see an extended discussion of sources that fail to become cases, but sources do fail and for a number reasons. Some of those reasons are technical: the source can be properly digitized or unitized for analysis; others are formal: the text does not cohere or is a incompetent member of the congregation. An examination of what gets turned away gives us more information about what remains. Finally, that little old unit of analysis provides the strongest testimony to the rigor of the study. Coding is difficult, fatiguing, time-consuming work; we often know the central features of the case well before the coding is completed. Everything about coding seems to conspire to feed destination fever and a rush to judgment. It is only the demands of the unit of analysis that stand in the way and only then if the unit of analysis is properly honored. Conceptual units of analysis are especially subject to shifting definitions and boundaries: segments get larger, ideas conjoined, difficult writing ignored. But formal units are also susceptible to sloppy work: codes are extended across multiple paragraphs in a mindless way, multiple ideas in a paragraph get a single code, oral sources are poorly paragraphed. In unit analysis, however, the author can show the reader the work: proportion of content coded and uncoded, ratios of codes to units, transitions to units, and range of a code across contiguous units, among many others. This sort of unit analysis is a prime safeguard against the self-fulfilling argument—the argument that cannot fail—that sometimes appears in the critical/rhetorical/hybrid study. That is the problem of inspiration undisciplined by evidence. On the other hand, the very detail provided by a fine sieve unit of analysis can itself be a forest-for-the-trees problem. Systematicity is not a substitute for insight. In every case there, will be a continuing tension in finding the balance between too close and too far. That balance is probably best struck in some combination of case coding that addresses the whole of the case and both conceptual and formal units of analysis. In my critical reading of interpretive work, I do not trust—and the methods give me no reason for trust—the declaration of close reading or even less the mere extraction of claim. These approaches lead to the pristine argument where everything coheres in clearly delineated patterns. No text is without contradictions, glosses, lapses and other conflicts. Yet all too often the critical/interpretive argument fails to recognizes them, deliberately ignores them, or refuses to mine them for their value.

3 Yes, the pun is intended. Extended interpretations of scenes and narrative intent are indeed coding with the interpretive concepts serving as codes that are subsequently presented in narrative.

Page 6: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

6

I have argued too far down this line to not draw the obvious conclusion: The effort required of interpretive excellence is high. All too often that effort is not justified by the problem addressed.

Primacy  of  the  Case  Whatever the argument and whatever the method, the quality of the work cannot be greater than the foundation of the case. A case, as we have seen, is an authenticated instance of a text that is a member of the body of texts that is the focus of the study. The case may be a complete textual form or a fragment of a text. It is the content over which coding will proceed. Whatever the argument for the body of texts, the case is one example of that body. With the cases in hand, the development of a literature review, the application of theory, the description of structure and content of the cases, and similar activities should await the close analysis and coding of the cases. All else is anchored in and emerges from the cases. Cases form the foundation of emergent coding. Selection of those cases, therefore, is not trivial. Selection of those cases depends on the procedures used to develop the collection of sources from which the cases will be drawn. What should those procedures entail: A random sample of all members of the congregation? Not possible. There is no way to develop a sampling frame that would capture all potential exemplars. What of a premier selection? This is the procedure used by Shugart: She picks three films that have all the components that will provide for her argument: mainstream films that feature food, ethnicity and gender. A perfect match to her argument. Legitimate? Absolutely, fish where there are fish. But do note that her collection process establishes the reach of her argument.

Starting  a  Project  and  Entering  Sources  in  NVivo  104  Work in NVivo starts with the creation of a project. Once the project has been created, Sources can be entered. NVivo 10 distinguishes between two kinds of sources: internal and external. Internal sources are digital files that are "entered into" (registered with) the application. The supported file types are presented in Screen shot (SS) 1. File types or data sources that are not supported such as .PPT or .PPTX or web pages can be referenced as an external source and a link entered into the source listing. Unsupported file types can often be converted into PDFs and entered as an internal source.

4 NVivo 10 has a detailed, step-by-step help file, which should be addressed for process instructions.

Page 7: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

7

Whatever the file type that is used, its content has to be digitized in such a fashion that coding can be conducted over the chosen unit of analysis. For example, PDFs can be of two types: image or text. An image is a picture even if it is a picture of written text. The words are not searchable. A text file may have pictures in it but at least some of the written text is searchable. Searchable files are much preferred.

Classifying  Cases  Coding is usually conducted across two levels of content, case. The first level is the case itself; the second occurs across the unit of analysis. Case coding involves the assigning of attribute values that describe the characteristics of the entire source or the source component in use. In NVivo, attributes are attached to classifications of which there are three: source, node, and relationship types. We will be using source classification here. One purpose of source classification is to allow the sorting of cases into types that might show different characteristics across units of analysis or across the coding of units of analysis. Let's set up an example ("Oh, goody," you say). One of the ways that I stay current with the directions taken in communication literature within the domain of empirical methods is to conduct an annual analysis of 100 journal articles—that have an empirical foundation—taken in units of 25 from four journals: Journal of Communication, Critical Studies in Media Communication, Journal of Broadcasting and Electronic Media, and Journalism & Mass Media

Page 8: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

8

Quarterly. I make no argument for these journals except that they represent a reasonable range across topic and typical approach. My interest in these articles is both as a theorist and a methodologist. I have nearly no interest in them for the claims they make but rather in how they make their claims, the forms of evidence used. and how they are situated in communication theory. The attributes I use in classifying sources, then, relate to those interests plus a couple of housekeeping details. SS 2 shows the eight attributes I use and most of the attribute values that can be applied (and a good test of your eyesight).

To open up the power of attribute classification, the analyst must first create a source classification. On the Create Tab select source classification. The source classification dialogue box opens; select create new classification; enter a name; click OK. To create the attributes for this classification, click on Classifications in the navigation pane (the window on the left); click on source classifications at the top of the pane and your newly created classification will appear in the source classification pane; click on the icon at the far left. Attribute should now be available on the Create tab (SS 3). Check the Help file for the different types available.

Because my example uses only one kind of source (journal article) I need only one classification. If I import my sources one by one, I can assign the classification and the attributes in the new document dialogue box. My preference in this case is to import all of my articles at once. I then click on Sources>Internals to list the sources; select all the sources; right click and classify all the sources at once. To quickly assign the attribute values, with a source or sources selected, right click and select Open Classification Sheet. All of the sources and attributes will appear in the lower pane with "Unassigned" as the attribute value. Hovering the cursor reveals the menu of values that can be assigned. You can add attributes and attribute values at any time.

Page 9: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

9

The  Power  of  Attributes  Attributes allow the analyst to sort and then to compare and contrast. In our example one of the attributes is "Methodological family." It has three values: metric, interpretive and critical/rhetorical/empirical (CRE) hybrid. Once classified I can sort sources across these values and compare the appearance of codes, frequency, common code, unique codes and the like. I can further breakdown one classification attribute by another. Consequently, I could examine the relative proportion of CRE hybrid studies across focal interest, for example. Massie in his study of university home page images through the use of attribute classification was able to examine all his images coded by power across gender. He was able to arrive at an interesting conclusion about the role gender played in the imaging of power relations. (No spoiler here; read the study.)

Understanding  Nodes—The  Codes  of  NVivo  QSR trainers tell the story that one of the original developers of NVivo was a coding engineer. A node was a location in a textual network that had special meaning. Code was a bit of programming. Consequently, the term node made its way into the QSR culture (and they have been explaining it ever since). True or not, the story helps one understand this odd bit of terminology. In the application, a node is a code that is attached to some unit of text. It does not matter if it is a single character or the entire source. NVivo does not distinguish codes according to any type but it gives the analyst the power to do so through node classification (discussed later). NVivo provides three separate operations of coding: manual coding, automatic coding (don't get excited; it doesn't think for you) or in vivo coding. We'll walk through those next.

Manual  Coding  Callie Spencer Alright, hike your pants up, take a seat, sit up straight, and take a deep breath, because here we go. This is where the rubber meets the road; this is the fun stuff. In order to show you how to code, I have imported a source into NVivo (following the above steps) that will hopefully provide you with a little comic relief. I suggest that our example project will look at all of the spoofs/jests on the Ph.D. acronym. This is a list compiled on one website which we might have imported as one case. So, here we go.

Page 10: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

10

Figure 1

Step 1: (Figure 1) While you are still on the “sources” view (1), double click on the little yellow symbol (2) in order to open your case (3). There is a header that runs across this new tab that you just opened that says, “Click to edit.” If you click that tab, it allows you to change the text itself within NVivo. Step 2: Click on “Nodes” on the lower left hand of your screen under “Sources” (1). Now you will be able to view the nodes that you create as you go. Note, “node” is just a fancy name for “code.” We are doing science here so it is important for things to sound cryptic and fancy. Step 3: Creating your first node. Yeehaw! First, highlight whatever piece of the text that you wish to code first. This is called your “coding unit.” It can be as big or small as you like (ie. paragraphs, sentences, phrases, or individual words). In our case, we are going to code each asterisk (*). Note: you can always code text several times with different levels of specificity. So, after we code at each asterisk, we might want to re-code each asterisk breaking it down further word-by-word so that we can separate spooks on the letter “P”, “H” and “D” respectively. At this point, you have several options which I will outline one-by-one.

Page 11: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

11

Figure 2

Option #1 (Using “right-click,” In Vivo coding, and uncoding): (Figure 2) After highlighting the text, right click it. A box will pop up giving you several options. For this first scenario, select “Code at In Vivo” (1). “Code at In Vivo” means that your node will have the same name as the text that you have highlighted. Notice that after you click this button, a bright blue ball appears under the Nodes heading (3) the name of your node is “Patiently hoping for a Degree” and you can now see the name of the user that created the node, the number of sources in which there is text coded at this node, as well as the number of individual occurrences of coding at this node. Because it doesn’t make much sense for us to create nodes that are named in such a way, let’s un-code this selection as an In vivo code. In order to do this, re-highlight the text. Again, right click it and this time, select “Uncode Selection” (2). Another menu bar will now present you with three options: “Uncode Selection at Existing Nodes,” “Uncode Selection at This Node,” “Uncode Selection at Current Node.” Select the first option. This option gives you the most choice. A new dialog box will now open. All of the nodes in which the selection is coded will appear and you can choose from which nodes to uncode the selection. In this case, select “Patiently hoping…” and “OK.” You will notice that the previous values of “1” under “sources” and “references” are now “0.” Before we continue, let’s delete the node that we just created. In order to do this, right click on that pretty blue ball (3) and click “delete.”

Page 12: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

12

Figure 3

Now, let’s do some manual coding, which is probably the bulk of what you will be doing. Again, highlight the text (Figure 3) and this time, click “Code Selection” (1) and then select “Code Selection at New Node” (2) as we have not created any nodes as of yet. The following box will pop up (Figure 4):

Figure 4

Choose a name for your node. We are going to call this one “length of time” as this node includes words such as “patiently” which allude to a long time spent waiting. We might also code this again at “loss of agency” as it seems as if the student is awaiting something over which they have little control. The possibilities are rather endless as you can see. It is important to either include a thorough description of the node at this point or to link a memo to the node so that you can write notes to yourself explaining to yourself at a later date what you were thinking in coding this selection here and in naming this node as such. In order to link a memo (Figure 5), right click that bright blue ball beside the node and choose “memo link.” Clicking link to new memo will create a new memo for you (as we have no existing ones at this point).

Page 13: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

13

Figure 5

Next, the following box will pop up (Figure 6) :

Figure 6

Note: you cannot name the memo the same exact name as a node (as I have done here). I have therefore changed the name of this memo to “Length of Time Memo.” Feel free to enter a brief description of the memo in the description box. The new memo will then appear as a new tab and the symbol for memo is a bright green tablet. If you open the memo, you will effectively have an open document where you can jot notes to yourself as you are coding. I know that this seems tedious, but I promise that you will be exceptionally happy if you do this now. I promise you your memory is not as good as you think it is (I mean, come on, we Ph. D. students have so much crammed into our brains at any given

Page 14: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

14

moment, how could we possibly remember our thoughts while coding) so save yourself time and stress later by recording your brilliant insights and thought processes as you go. Option #2: (Using the toolbar at the bottom of the page) (Figure 7) Your next option for coding is to use the handy-dandy toolbar located at the bottom of the page.

Figure 7

First, as per usual, highlight a chunk of text that you would like to code. Then, in the “code at” box, either choose an existing node (1) (you can also click the arrow (2) to activate a drop-down menu that allows you to see all of your nodes) or you can create a new node by simply clicking in the blue box, deleting the name that is currently in the box (in this case, Length of Time) and typing in a new name. After you have successfully selected the node at which you would like to code your selected text, you MUST click the green check button (4) to actually code the text at that node. Clicking (5) the red X will UNCODE the selection at your chosen node and (6) the blue ball icon will do In Vivo coding for you, naming the node the same thing as the text highlighted. The gray box (3) opens up a new box which allows you to code the selected text at multiple nodes. Option #3: (Using the toolbar at the top of the page) (Figure 8)

Figure 8

A third option for coding is to use the toolbar at the top of the page. After you have selected your text, go to the “Analyze” tab. Here, you can code your selection at a new node, existing node, In Vivo, or you can choose to Auto Code. At this point, you might want to punch me in the face for not telling you about this feature earlier in the process; however, don’t get too worked up, it doesn’t do what it sounds like it does. Auto code does not do all of your coding for you. The “Auto Code” feature is useful if you have organized your text into distinct paragraphs using tools such as hierarchical headings for individual interview questions and responses. The following is an example of this organization given in NVivo “help” (Figure 9).

Page 15: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

15

Figure 9

Using the auto code, you must select your paragraph style and the location for these codes (at new node or existing nodes). Then, new nodes will be created for each paragraph and you can code each question as a different node. Checking the “aggregate nodes” box will combine nodes into parent and child nodes. From the above example, the H2 node will be located inside of the H1 node.

How  to  visibly  see  what  you  have  coded   So now that you have gone along and coded some text, how do you see what you have already coded and in which node(s) it is coded? I have two tidbits to help you answer this question. First: turn on coding stripes (Figure 10). In the toolbar at the top of the page, under the “View” tab (1), choose “Coding Stripes.”

Page 16: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

16

Figure 10

A drop down menu will appear offering you several options. I find it most useful to select “Selected Items” (2) at which point a box will pop up with all of your nodes. You can then check the ones you would like to appear in the coding stripes, or, alternatively, you can simply click “Select All” and all of your codes will be visible as coding stripes. The coding stripes will appear on the right side of your text. They will tell you two pieces of information. First, the colored block (3) will directly correspond with the actual line of text that you have coded at the indicated node. Second (4) the grayscale will let you know how often you have coded text at that specific node within this case. The lighter the gray, the less coding at the node, the darker, the more frequent the coding is at that specific node. Another way to view what you have coded is to look at each individual node and pull up a list of all of the coded text within that node. (Figure 11)

Page 17: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

17

Figure 11

To do this, double click the node that you would like to view (1). A new tab will pop up (2). This tab will list all of the text coded at that node. At the top, it will show you the total percentage of the text that is coded at this node (3). It will then list each individual reference including percentage coverage for each individual coded unit (4) along with the actual text of each unit (5).

How  to  generate  a  report  of  what  you  have  coded NVivo can generate a printable “Coding Summary Report” which is rather helpful in taking a “birds-eye view” of your progress (Figure 12). In order to create such a report, choose the “Report” box in the lower left hand side of the screen. Next, choose “Coding Summary Report.” At this stage, you have several options of what you would like NVivo to include in the report. I went with the default and clicked “OK” which gave me everything. A new tab appears containing the report which conveys the name of each node, the number of coders that coded at each node, the percentage of text which resides in each node (coverage), whether or not the node is an aggregate node (parent and child hierarchy), and the number of coding references at each node.

Page 18: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

18

Figure 12

Finally, I could not re-iterate this enough, if you have coding questions, the NVivo help tab is actually very helpful and is organized in a concise manner that makes good sense. It is located under the “File” tab. Good luck and remember to take breaks!

Tips  for  Importing  Source  Files  into  NVivo  10  Heather Stone

Before  Importing  

Clean Up File Names. It is easier to change file names in the Windows file system than in NVivo. You can do it in NVivo, but you have to right click on each source to select “Properties” to bring up a dialogue box, and then type a new name in the “Name” field.

Create a Folder Structure for Sources. If the folder exists before import, you can open the appropriate folder before you push the import button and NVivo will automatically store the source files in the right spot for you. Otherwise, you have to drag and drop them into the correct folder manually.

Do All Necessary File Conversion. NVivo only supports certain file types, and it doesn’t do any file format conversion. Make sure files are in a compatible format or they won’t import with the proper tool, if at all. For example, if you click the “External Data-->Videos” button, and your video file is not in a supported file type, it won’t even show up in the folder as an option for importing.

NOTE: See NVivo online help for a list of supported file formats.

Page 19: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

19

Hints  1. Import  only  the  files  you  need.  Viewing  and  using  files  in  NVivo  is  harder  than  viewing  and  using  files  

in  Windows  File  Manager  Utilities.  For  example,  if  you  import  an  entire  directory  of  pictures,  

intending  to  only  use  the  best  ones,  then  it  will  be  difficult  in  NVivo  to  figure  out  which  photos  are  in  

fact  the  best  ones.  You  are  better  off  opening  the  pictures  in  Windows  File  or  Picture  Management  

Utilities  to  identify  the  photos  you  intend  to  use  and  delete/archive  the  rest  before  you  import  any  

of  the  photos  into  NVivo.    

2. Name  source  files  something  identifying,  rather  than  simply  using  consecutive  numbers.  Quick  and  

easy  scanning  of  thumbnails  is  not  possible  with  NVivo.  If  your  file  names  don’t  clearly  identify  your  

source  material,  you  will  waste  a  lot  of  time  opening  files  for  editing  just  so  you  can  tell  what  they  

are.    

Working  with  Multimedia  Sources  in  NVivo  10     One of the most powerful features of NVivo 10 is how easily the software lets you manipulate sources other than MS Word documents or PDFs. You can import and code videos, audio files, transcripts, photos, and even data sets such as MS Excel spreadsheet files. These various source files can be organized into folders and given source- and node-level classifications just like the more common text sources. NOTE: See the section entitled Starting a Project and Entering Cases in NVivo 10 for more detail about supported file formats.

Importing  Video  Files   TIP: If the video file you want to work with is not in a supported format, use video conversion software such as Free Recorder to convert the file BEFORE attempting to import it into NVivo. Use the “Videos” button on the “External Data” tab to import video files:

When you see the “Import Internals” dialogue box, click the “Browse” button to find and select the video:

Page 20: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

20

You can import several files at once by using the CTRL key to select multiple files, or you can simply import files one at a time. If you import files one at a time, you are prompted with the Video Properties dialogue box:

If you import several files at once, you will not see the Video Properties dialogue box. You can set properties later by right-clicking on the imported source file. Use the “General” tab on this dialogue box to give your video file a name and description. Use the “Video” tab to specify whether you want your video imported by reference or embedded in NVivo. Use the “Attribute Values” tab to specify a source-level classification and attributes for the file. You can of course wait to classify files until after you’ve imported them. Files imported by reference are not actually stored within NVivo and are instead accessed from an external directory whenever you manipulate the video within NVivo. Embedded files are actually stored within NVivo. Either method is supported fully by QSR. However, be careful not to delete or change the path of files imported by reference because doing so will break the NVivo link to the file and your work will be lost.

Page 21: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

21

On the other hand, if you have many very large files, you might overtax NVivo if you embed them all. The safer strategy may be to link by reference and rely on the Windows file management system for actual storage.

Importing  Other  File  Types   Follow the instructions in the previous section to import audio files, transcripts, photos, or data sets. Select the appropriate type of source and follow the prompts on the screen.

Coding  Video  Files  To code videos at a source-level, right click on the video file in the “Internals” window. The following menu pops up:

Choose “Code Sources” to display a smaller pop-up menu. From this menu, choose “Code Sources at Existing Nodes” if you want to use a node that you have previously created. Choose “Code Sources at New Node” to create a new node.

Page 22: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

22

To code at a node level, double-click the video name to open it within NVivo. The video will display on a divided screen, as shown, with the accompanying audio track above it. The “Media” menu will be selected at the top of the screen, showing the video playback tools provided to make coding easier. NOTE: If the video does not appear on a split screen as shown, change your screen display settings using the “View” tab.

NVivo has several powerful playback tools, including a Skip Back (F9) and Skip Forward (F10). You can use these with Start Selection (F11) and Finish Selection (F12) to easily navigate with your keyboard to exactly the portion of the video you want to code. NVivo also allows you to speed or slow down playback, to loop playback, to start a transcript file, and to synchronize a transcript with the audio and video. Though these functions still require some fairly tedious interaction, they do provide much more powerful data manipulation capability than a researcher could obtain without the software. You can also set an individual video frame as a thumbnail and create a Content log where you can make specific notes and tie those notes to a segment of the video.

Page 23: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

23

Coding  Audio  and  Photo  Files   Audio files can be manipulated much the same way as video files, except that you only have an audio track to work with rather than a combined audio/video track. Many of the same playback tools exist, and you can start a transcript or a content log for an audio file as well. Photos have their own set of manipulation tools, including rotating, compression, and brightness. You can use the mouse to select a particular region of the photo to either code or to call out with notes in a content log. Manipulation of multimedia files is a strength of NVivo 10. Renaming  Your  NVivo  10  Project   Your NVivo project file has two types of names: A  Title  which  is  a  text  string  you  assign  when  you  create  the  project.    

A  Filename  which  is  the  name  that  is  stored  on  your  computer’s  file  system.  NVivo  10  filenames  use  

the  .nvp  extension.    

Changing  the  Project  Title   Your project title is simply a 250-character or less text string you assign when you create a project. You can use numbers, letters, and even special characters in your project title. However, your title may print automatically on reports, so you would be smart to keep it short and readable. To change your project title after you’ve already created your project, do the following:

1. Click  the  “File”  menu  in  the  top  left  corner.    

2. Select  “Info-­‐-­‐>Project  Properties”  

3. When  the  Project  Properties  dialogue  box  appears,  select  the  text  in  the  “Title”  field  and  replace  

it  with  your  new  title.    

Page 24: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

24

4. If  desired,  add  text  to  the  “Description”  field  to  further  clarify  your  new  title.    

5. Change  any  other  fields  you  want  to  modify.  Note  that  you  cannot  change  the  filename  in  this  

box.    

6. Click  the  “OK”  button  to  save  your  modifications  and  close  the  dialogue  box.    

Renaming  or  Moving  the  NVivo  File   When you create a new NVivo project, you are prompted to enter a Title, a Description, and a Filename including location in your computer’s directory structure. CAUTION: Think carefully when you specify a filename and directory location. You can change filenames and locations later, but this breaks your “Recent Projects” shortcuts within NVivo and in some cases, may cause a database access error. Remember you can change your Project Title easily from within NVivo, so do that instead of renaming your file whenever possible. Also, use the “Copy Project” tool from within NVivo if a copy of your project is really what you need. Copying is safer than renaming. If you must rename or move your NVivo file, use the Windows Explorer Utility. NOTE: If you are renaming or moving a server file rather than a standalone file, consult the NVivo user’s guide for special instructions.

Sharing  or  Copying  the  NVivo  File   CAUTION: Always share or copy your NVivo file by using the Manage Projects tool within the NVivo program. Do not use the Windows Explorer utility to copy, as this may cause database access errors, make the copied file unreadable, or fail to preserve certain settings. To share or copy your NVivo file, do the following:

Page 25: Coding’Texts’using’NVivo’10’ - WordPress.com...Coding’Texts’using’NVivo’10’ James A. Anderson Callie Spencer Heather Stone Introduction* A text can be considered

DRAFT: NOT FOR CITATION OR QUOTATION

25

1. Click  the  “File”  menu  in  the  top  left  corner.    

2. Select  “Manage-­‐-­‐>Copy  Project”  

3. When  the  Copy  to  Project  File  dialogue  box  appears,  type  in  the  filename  you  wish  to  assign  to  

the  copied  file.  If  desired,  browse  through  the  directory  structure  and  select  a  specific  folder  in  

which  to  store  the  copied  file.      

4. Click  the  “Save”  button  to  close  the  dialogue  box  and  copy  the  file.  

5. When  the  green  status  bar  in  the  bottom  left  corner  says  “100%”,  your  file  copy  is  complete.    

You can use Windows Explorer to browse through your directories and access the file, or you can open it within NVivo.