Data analysis in mixed methods research consists of analyzing the qualitative data using qualitative methods and the quantitative data using quantitative methods; therefore, knowing the steps in both forms of analysis is necessary in mixed methods research (Creswell & Plano Clark, 2007). For both forms of analysis, researchers go through the following similar set of steps: preparing the data for analysis, exploring the data, analyzing the data, representing the analysis, and validating the data. These procedures in quantitative research and qualitative research are different from each other as stated by Creswell and Plano Clark, (2007).
Preparing the Data for Analysis
In responding to my research questions and hypotheses, the way I will prepare the data for analysis qualitatively will involve organizing documents by transcribing text from interviews and observations into a word-processing file. During the transcription process, the researcher checks the transcription for accuracy and then enters it into a qualitative data analysis software program. SurveyMonkey web-based research tool will also be used for the research survey questions and the feedback that follows.
From the feedback, answers and responses the researcher gets from qualitative research questions and surveys, quantitative research questions or hypotheses will be formulated after the completion of the initial qualitative phase. The reason for collecting qualitative data initially is that instruments are not available, variables are not known, there is little guiding theory or few taxonomies (Creswell & Plano Clark, 2007).
Exploring the Data
In exploring the data to respond to my research questions and hypotheses, I will first consider the qualitative research method, which will involve reading through all of the data so as to develop a general understanding of the database. This will mean recording initial thoughts by writing short memos in the margins of transcripts or field notes. Additionally, during general review of the data, all forms of data are reviewed, such as observational field notes, journals, and minutes from various meetings, pictures and transcripts of interviews.
Making these memos becomes an important first step in forming broader categories of information, such as codes of themes (Creswell & Plano Clark, 2007). A qualitative codebook will be developed at this time as well. The codebook is a statement of the codes for a database and it is generated during a project and may rely on its codes from past literature as well as codes that emerge during an analysis. The process of generating this codebook helps organize the data and it facilitates agreement on the contents of the transcripts as new codes are added and other codes removed during the coding process.
In quantitative analysis, exploring the data involves and entails visually inspecting the data and conducting a descriptive analysis, such as, the mean, standard deviation, and variance of responses to each item on instruments or checklists to determine the general trends in the data.
I will explore the data to see the distribution of the data and determine whether it is normally or non-normally distributed so that proper statistics can be chosen for analysis. Descriptive statistics are generated for all major variables in the study, especially the main ones, such as independent and dependent variables.
Analyzing the Data
Analyzing the data will allow me to examine the database to address the research questions or hypotheses. In both quantitative and qualitative analysis, we see multiple levels of analysis. In quantitative research, the researcher analyzes the data based on the type of questions and hypotheses and uses the appropriate statistical test to address the questions or hypotheses. The choice of a statistical test is based on the type of questions being asked, the number of independent and dependent variables, the types of scales used to measure those variables, and whether the population is normally or non-normally distributed. Another aspect of quantitative analysis is the necessity of keeping in mind that although hypothesis testing is useful, journal editors look for evidence of practical results, reported as effect sizes and confidence intervals.
Qualitative analysis begins with coding the data, dividing the text into small units, such as phrases, sentences, paragraphs, and assigning a label to each unit. This label can come from the exact words of the participants. If the researcher codes directly on the printed transcript, the transcript pages need to be typed with extra-large margins so that codes can be placed in the margins. In this hand-coding process, researchers assign code words to text segments in the left margins and record broader themes in the right margin. A more practical approach today is to use one of the many qualitative data analysis software programs (Creswell & Maietta, 2002). These programs all contain some combination of the following features.