Chapter 4 & 5 – Data Preparation & Getting Started (Dedoose)
Chapter 4 discusses good transcription practices and discusses the auto-processing which is possible when importing survey data into particular software programs. In Chapter 5 to help get you started, we talk about productive things you can do in the early stages of setting up a ‘project’ in software. Experimenting with these processes where relevant in your chosen software will help to become familiar with useful entities in the software; experiments to import data will help to confirm you have prepared data in the right way. See all coloured illustrations (from the book) of software tasks and functions, numbered in chapter order.
Creating the software project
A framework of memos
Getting Started with Dedoose – overview & building blocks
Dedoose is a web application or Software as a Service (SaaS). This means that rather than purchasing a license for software that you install and maintain on your personal computer, you create an account as a Dedoose user and pay a subscription fee for access to the service for particular periods of time. This is the future of software as more and more we access free or for fee services that are delivered via the Internet. Watching movies, listening to music, reading news, finding instructions for how to use your microwave, banking, education, communicating with friends and loved ones via social media, and now, with Dedoose, managing, sharing, analyzing, and presenting qualitative and mixed method research data. Any Internet connected computer (PC, Mac, Linux…) with the free Adobe Flash Player is all you need to access Dedoose.
Setting up a Dedoose user account can be done quickly via the Dedoose informational website. Click ‘Get Started,’ enter the required information (no billing information is needed), agree to the Terms of Service, and submit.
Your account will be set up immediately and when you log into Dedoose, you will find two projects connected to your account – an empty project with the title you had provided when setting up your account and a copy of our demo project so you can explore Dedoose features while building your own database.
To get started, it is helpful to understand the basic organization of data in Dedoose. Most aspects should be familiar to qualitative or mixed method researchers, but the nature and use of descriptors is unique in Dedoose. The following schematic shows the fundamental database relationships in Dedoose. Media are stored as text, video, or audio, descriptors are stored in a spreadsheet format and related to media, excerpts are identified within media, and codes and codes weights are associated with excerpts. All of these objects are available for searching, sorting, and analyzing in Dedoose
When you log into Dedoose you will be presented with the dashboard of the project you most recently accessed.
Dedoose is organized into a number of workspaces which you can access via the main Dedoose menu bar. While you can carry out particular tasks in a number of places in Dedoose – including the home dashboard, here are the basics:
- ‘Home’ – project dashboard
- ‘Analyze’ – for access to all interactive charts, graphs, and tables
- ‘Excerpts’ – view, sort, filter, and export excepted qualitative data
- ‘Descriptors’ – view, create, upload, sort, filter, and export descriptor data
- ‘Codes’ – view and create codes and modify the code tree structure
- ‘Media’ – view, upload, sort, and filter qualitative data sources
- ‘Memos’ – create and organize memos that are free floating or linked to any number of objects in the database
- ‘Training Center’ – create, take, and view the Cohen’s Kappa and Pearson’s correlation results of tests designed to help research teams build and maintain inter-rater reliability
- ‘Security Center’ – where you can create security groups with different access privileges in your project and assign different users linked to your project to the group appropriate for their role and responsibilities
- ‘Account’ – essentially the billing center where you can view, add, or remove other Dedoose users for whom you are paying, view the status of your account, submit payments, get receipts, and change your user account password
- ‘Projects’ – view the different projects to which your user account is linked, move from project to project by double clicking on a project title, create new projects, rename or delete projects. Note that, by default, when you log into Dedoose the most recent project you had been working on will load
- ‘Data Sets’ – create, save, and activate global filters based on any number of Boolean operators
Setting up a New Dedoose Project
From the ‘Projects’ click the ‘Create Project’ button, provide a title, and the new project will be immediately ready for use.
There are three basic building blocks to a Dedoose project before you begin creating excerpts and then analyzing your data:
- Qualitative data – video, audio, text, and other media
- Descriptor data – demographic, scale score, and other survey types of data which are important in characterizing and distinguishing the sources of your qualitative data from a more quantitative perspective
- Code System – the conceptual framework you will use to understand, organize, and communicate the meaning within the qualitative data and how they speak to your research questions
Establishing these building blocks can be done in a variety of ways. Create each within Dedoose manually or upload one or more of these database aspects from .csv, .xls, .xlsx, .txt, .rtf, .doc, .docx, .mp3, and/or .mp4.
Uploading Qualitative Data
Clicking the ‘+’ sign in the upper right corner of the ‘Media’ panel on the Dedoose dashboard or in the ‘Media’ workspace will activate a pop-up with options to upload video, audio, or text files or create a new document within Dedoose. Select the type of media to upload, select any number of files to upload, submit, and they are brought into the project. Two things to note: a) video and audio files must be processed within Dedoose before they can be viewed and excerpted and warnings will inform when they have uploaded and are ready and b) the time required for upload will depend on the size/type of file and the quality of Internet connectivity.
Creating or Uploading Descriptor Data
A set of descriptor data includes two facets. First are the definitions of the fields (variables). Fields can be one of four types: text, number, date, or option list (categorical). Each field definition includes a title, description, type – if option list, a set of two or more values (ex. values for eye color could include brown, blue, and green) – and, if appropriate, a setting to designate the field as dynamic which allows for the analysis of change over time in longitudinal research. The other facet of a descriptor set are the data themselves. Descriptor data are what drive many of the interactive Dedoose analytic charts, tables, and graphs and can also be used in sorting and filtering the work you’ve done with the qualitative data.
Descriptor field definitions and data can be imported directly into Dedoose from a .csv or Excel file. Yet, from a more manual approach you can upload your qualitative data, create and enter your descriptor data, create and structure your code system.
Creating and Working with Code Systems
To add or modify a Dedoose code system, click the ‘Edit’ codes icon in the ‘Codes’ panel
To add a new root code, click the ‘+’ sign in the panel header and enter the desired information: a title, description (which will appear as a smart tip in Dedoose when floating over the code with your cursor), and, if appropriate, you can enable and enter the settings for a code weight system. Making use of code weighting can be a powerful part of a qualitative or mixed methods database. The system can be anything one can represent by a dimension. For example, in the Dedoose demo project that is created when a Dedoose account is established, rating scales of 0-5 were used to represent the relative quality of reading by parents to young children and other aspects of a home literacy environment, 0-7 to indicate the number of days per week the parent read, and 0-90 to indicate the number of minutes in a reading episode. Other uses of weight systems include ‘importance,’ ‘quality,’ or ‘centrality.’ Weights/ratings are set whenever you apply particular codes and can be activated at any time. Often one must become familiar with the qualitative data before a clear dimension emerges and weight/rating systems can be applied whenever the time is right.
To change the hierarchical structure of a code tree when in editing mode, codes can be dragged and dropped onto other codes to create children or dropped into the space at the bottom of the tree to convert a child code to a parent.
Codes can also be merged or deleted. To edit an existing code you must first be in code tree editing mode. Then when floating your cursor over a code three icons appear: memos, edit code, and add child code (another way to make use of a code hierarchy). Click the edit code icon and you can change the title or description, activate/deactivate code weighting, delete or merge the code
Creating and Working with Memos
Dedoose memos are organized in a folder system. Click the ‘Memos’ icon in the main menu bar to enter the memo workspace. Create a first folder by clicking the ‘Edit’ icon ’ in the memo folders panel click the ‘+’ sign to add a new folder, enter a title, and submit.
To add new memos, click the ‘+’ sign in the larger ‘Memos’ panel, enter all desired information, designate the folder of choice, and save.
You can also view and edit the items linked to the memo whenever a memo is open and active, though it is more common to create and link memos when dealing directly with particular documents, excerpts, or codes and there are ‘memo’ buttons and icons throughout Dedoose for just this purpose – you’ll find them when editing excerpts or codes, viewing descriptors or documents, and they can also be linked to locations in documents, videos, or audios, by right-clicking at the specific location and choosing to ‘Create/Link Memo.’
Excerpting and Coding
Excerpts are created in Dedoose by blocking text and then either clicking the ‘create excerpt’ button in the lower left corner of the document viewer or right clicking on the blocked text and selecting the ‘create excerpt from selection’ in the short cut menu. You will see that the excerpted text becomes colored, a bracket will appear in the left margin marking the boundaries of the excerpt, and the header in the ‘Selection Info’ panel will turn orange – your indication you are in excerpt editing mode. You can then apply codes by dragging and dropping them into the selection info panel or by double clicking the codes of choice. If activated, you can also set code weight/rating values. While in excerpt editing mode, codes can be deleted by clicking the ‘x’ next to any code that was applied and the entire excerpt can be deleted by clicking the ‘x’ next to the excerpt icon. Clicking anywhere else in the document will cause you to exit editing mode and you can re-enter editing mode by clicking on the bracket that marks the excerpt boundary.
Excerpting audio or video is a similar process. Clicking anywhere on the stream timeline will cause the play head to jump to that location and there are two excerpt brackets that always surround the play head. Pause the video or audio, set the start and end points by dragging the excerpt boundary brackets, and click the ‘create excerpt’ button. You’ll see a bracket appear in the timeline and, when orange, codes are added and removed in the same way as when excerpting text.
Preparing and importing data
Preparing and Importing e.g. Word files (as in Case study B) the focus group data.
There are no special changes or requirements of textual data. The other documents to be uploaded to the project need no special preparation. The process starts by clicking the ‘+’ sign in the Media panel, selecting ‘Import Text,’ locating the files on the client computer, and then submitting.
Preparing and Importing ‘Survey’ Data to Dedoose
In situations where project data are all stored or prepared in a spreadsheet, ex. exported from SurveyMonkey, Qualtrics, or other online data collection service, they can be imported directly into Dedoose if the Excel file is properly formatted.
What is required, and recommended to maximize analytic features, the following guidelines should be followed prior to the import.
- Replace missing data points for any categorical variables so all respondents will be included in Dedoose charts and tables
- Use brief and informative column headers
- Numerical proxies for values should be changed to their proper label…as in the sample data ‘Q.1. JOB-SECURE’ column. Ex. 0 = ‘Very Poor,’ 1 = ‘Poor’… 4 = ‘Very Good’
- Where respondents can select more than one value to a survey question, create columns for each value and then use values like ‘true/false’ or ‘yes/no’ as the data points and then unpack the data for each participant…as in the sample data ‘Fault’ column
- For any narrative/open-ended responses, you add a ‘_ddqual_’ prior to the code in the column header. Ex. ‘How Affected?’ becomes ‘_ddqual_How Affected?’
And, with that, the file is ready for import via the Dedoose survey import wizard. All demographic and scale data will be converted to descriptors, the fields will take the name of the column headers and the valid values will all be defined, and each case will be populated. For the qualitative data, each column with qualitative data is extracted and a document created for each participant with all their text responses and properly linked to the descriptor data. The column headers for these data become codes in Dedoose and each response is excerpted and tagged with the appropriate code.
A new empty project:
And seconds later after clicking the survey import icon and selecting the file…
The other documents to be uploaded to the project need no special preparation. The process starts by clicking the ‘+’ sign in the Media panel, selecting ‘Import Text,’ locating the files on the client computer, and then submitting.
The next exercises which are connected with topics in Chapter 6 – concentrating on work that can be done at data level