Abstract
Two computation-enabled matrix-based analytics techniques have become more available for the analysis of text data, including from online surveys. These two approaches are (1) the qualitative matrix coding query and (2) the qualitative crosstab matrix, both in NVivo 12 Plus. The first approach enables insights about the coding applied to qualitative data, and the second enables the identification of data patterns based on case (ego or entity) attributes of survey respondents. The data analytics software has integrations with multiple online survey platforms (Qualtrics and Survey Monkey currently), and the automated coding of the data from these respective platforms and other software features enable powerful data analytics. This chapter provides insights as to some of what may be discoverable using both matrix-based techniques as applied to online survey data.
TopIntroduction
Matrices, as exploratory data structures, have long been used in qualitative research to identify data patterns, surface fresh insights, and achieve other research and data analytics aims. They have been used to elicit information from survey respondents in visual ways (aka “graphic elicitation techniques”) (Copeland & Agosto, 2012, pp. 514 – 517, pp. 519 - 524). For quantitative, qualitative, and mixed methods research, matrices have been a staple. In early days, they were completed manually, and in more recent years, they have been populated using computational means. Matrices are a basic data structure form: a rectangular (including) table consisting of row and column headers and then overlapping or intersecting cell data. For binary matrices, the cells are 1s (present) or 0s (not present); for intensity matrices, the cells are numbers, with higher numbers indicating higher intensities of counts or frequencies. Depending on their focuses and respective purposes in research, the different matrices have different names.
Since the early 1990s, qualitative data analytics suites have extended the power of computational matrices. These tools are referred to as a category of Computer Assisted Qualitative Data AnalysiS (CAQDAS). Software may be used to convert coding from qualitative data for statistical analysis in a mixed methods approach:
Such integration (of mixed methods studies data) is seen as occurring: (a) when text and numeric data are combined in an analysis; (b) when data are converted from one form to another during analysis; or (c) when combination and conversion occur together iteratively or in generating blended data for further analyses. (Bazeley, Spring 2006, p. 64)
These enablements broaden the types of available knowledge and askable questions in qualitative research. They complement the theory-based top-down-coded research by enabling reproducible research with objectively supported data.
This work introduces two core matrix applications in the NVivo 12 Plus software tool: (1) the qualitative matrix coding query and (2) the qualitative crosstab matrix. It also includes some references to some other lesser-known matrix queries (like Coding Comparison queries (based on a similarity matrix). [QSR International (Qualitative Solutions and Research) is the maker of NVivo, which was initially known as NUD*IST or “Non numerical Unstructured Data Indexing Searching and Theorizing software.” NVivo originated in 1999.]
TopReview Of The Literature
Matrices have over a 150 years of historical priors. An early precursor would be a mathematical array of numbers, represented in a rectangular form. These mathematics-based matrices have been around since the 1850s. They have been harnessed for quantitative data analysis for a number of applications, and these applications are integrated in any number of quantitative data analytics packages. These data matrices or data arrays may involve millions of variables that are managed through spreadsheet programs and statistical packages (and some databases).
At core, a matrix is just a basic structure. There are column headers representing columns of data, and there are row headers representing rows of data, and the intersecting cells between the row and the columns indicate intersections or overlaps between the particular column and row. (Table 1) These intersections show “associations” and suggest at a kind of relationship between the two variables (in the Column and Row), but much more needs to be understood for further understandings and further definitions of relationship (Positive association? Negative association? Curvilinear association? Causation? Precursor? Leading indicator? Lagging indicator? Orthogonal non-relationship? Others?)
Table 1. Key Terms in this Chapter
Similarity Matrix: A data table that compares the amount of similarity and difference between the coding of two different coders or respective coding teams.
Binary Matrix: A rectangular data table that records the absence or presence of a particular phenomenon in the cells indicating overlaps between the respective row and column headers.
Matrix: A data table with fixed numbers of rows and columns, with each representing a variable or attribute or other phenomenon, and the overlapping cells capturing the incidence or intensity of a phenomenon.
Qualitative Matrix: A data table structure involving qualitative and/or mixed methods research data (often in textual format).
Matrix Coding Query: An exploratory query that involves defining row and column variables for a constructed matrix, which may include a combination of coding nodes, source contents, folder contents, and other contents in the project file.
Relational Matrix: A data table with similar variables in the column headers and the row headers to enable the identification of the presence of a relationship between variables or not (in a binary matrix) or the intensity of a relationship between variables or not (in an intensity matrix); resulting data can be represented as a relational network graph.