Qualitative research is important because it generates data that can provide in depth insight into a question or topic. However, in order to draw conclusions from qualitative data, it is essential to quantify the data. “Qualitative researchers may criticize [the] quantification of qualitative data, suggesting that such an inversion sublimates the very qualities that make qualitative data distinctive: narrative layering and textual meaning. But assessment in the university (and the policy implications that flow from it) demands that the data are presented within a scientific construct.” (1) In addition, “until we know more about how and why and to what degree and under what circumstances certain types of qualitative research… can usefully or reliably be quantified, it is unlikely that program planners or policy makers will base decisions on studies generally regarded as ‘qualitative.’” (2)
Therefore, it is important to quantify the data obtained from qualitative research. Quantitative analysis of qualitative data “involves turning the data from words or images into numbers. This can be done by coding ethnographic or other data and looking for emerging patterns.” (3) If qualitative data is in the form of responses to standardized questionnaire surveys, this data may also be quantified. Simple frequencies and relationships between variables can be calculated either manually, or by using qualitative software, such as EZ Text. For example, a researcher studying smoking habits utilized a frequency table to describe the smoking that occurred in specific contexts. The definitions of these "contexts" were derived from interview data generated from in-depth interviews with youth.(4)
There are three main steps to conducting a quantitative analysis of qualitative data: organizing the data, reading and coding it, and presenting and interpreting it.
First, the researcher should organize the data. The data can be organized in groups which relate to particular areas of interest. For example, a study on tobacco farmers might group data into the following sections: history of tobacco farming, other crops grown, role of women in tobacco farming, reasons for tobacco farming and environmental consequences of tobacco farming. (5)
The next step is to read all of the data carefully and construct a category system that allows all of the data to be categorized systematically. The categories should be internally homogeneous and externally heterogeneous. "Everything in one category must hold together in some meaningful way and… the differences between categories need to be bold and clear.” (6) If there is a lot of data that doesn’t fit into the category system, it usually means that there is a flaw that requires the system to be reorganized. Good qualitative research will have a label for all data, and every attempt should be made so that each segment fits in only one category. Lastly, the classification system should be meaningful and relevant to the study. Once a system created for organizing the data, each category should be assigned a number, and then transcriptions of interviews or survey results can be coded. (7)
Included below is an example of a coding system, followed by a coded interview: (8)
1. The smoking period
1.1. When started and stopped
1.2. Reasons for starting smoking and continuing to smoke
2. Circumstances in which smoking takes place
3. Influences on smoking behavior
3.1. Home environment
3.3. Work environment
4. Reasons for stopping
4.2 Cultural or religious
4.4 "Significant other" pressure
4.5 Consideration for others
5. Ways to stop
5.1 Based on experience of respondent
5.2 Based on opinion of the respondent
This is an example of a portion of an interview with the categories assigned to segments of the text. (9)
Interviewer (I): How did you start smoking?
After the data is coded, the data should be displayed and organized so that it can be interpreted. Often simple matrices or charts can be used to compile interview data so that patterns can be determined among respondents. Causal network diagrams and flow charts are also often helpful to assess the cause and effect of relationships that appear in the data. (10) In order to analyze the data, the use of a computer-assisted qualitative data analysis program is suggested. Such programs link code with text in order to perform complex model building and help in data management.(11) For example, EZ-Text is a software program which is useful when working with responses to open-ended questions in a standardized survey. When the same questions are asked in every interview, EZ-Text can be used to quantify the results of an analysis, indicating the frequency of particular responses to each question. This is just one example of a computer program, and there are many other available options that depend on the exact nature of the research and the size of the database. The coding and analysis of data in qualitative research is done differently for each study and depends on the research design, as well as the researcher’s skill and experience. Regardless of the study, it is always essential to clearly document how the data was coded and interpreted, and it is important to quantify it in order to draw conclusions. (12)
(1) Ward, T. “Quantifying Qualitative Data.” Accessed on 9 December 2010.
(2) Green, E. “Can Qualitative Research Produce Reliable Quantitative Findings?” Field Methods. 13.1 (2001): 3-19. Accessed on 8 December 2010.
(4) “Module 6: Qualitative Data Analysis.” Accessed on 9 December 2010.
(10) “Module 6: Qualitative Data Analysis.” Accessed on 9 December 2010.
(11) Ward, T. “Quantifying Qualitative Data.” Accessed on 9 December 2010.
(12) “Module 6: Qualitative Data Analysis.” Accessed on 9 December 2010.