Case Studies

Case Study 1: Grounded Theory

When fieldwork, netnography or observations are used, a particularly apt way to analyse collected data may be grounded theory, which was developed by Glaser and Strauss (1967) and later by Strauss and Corbin (1990). The analysis entails “classifying and categorizing text data segments into a set of codes (concepts), categories (constructs), and relationships” and then “the interpretations are ‘grounded in’ (or based on) observed empirical data” (Bhattacherjee 2012, 113). The coding strategies of grounded theory and explain with the graph below:

Open coding
  • A process aimed at identifying concepts or key ideas within textual data, related to the phenomenon of interest.
  • The researcher examines the raw textual data line by line to identify discrete events, incidents, ideas, actions, perceptions, and interactions of relevance that are coded as concepts.
  • Each concept is linked to specific portions of the text (coding unit).
  • Similar concepts are later grouped together into higher order categories, which tend to be broad and generalizable.

Axial coding
  • Categories and subcategories are assembled into causal relationships or hypotheses that can tentatively explain the phenomenon of interest.
  • A coding scheme may be used to understand which categories represent conditions, actions/interactions, and consequences.
  • With the coding scheme, researchers can start explaining why a phenomenon occurs, under what conditions, and with what consequences.

Selective coding
  • Identifying a central category or a core variable and systematically and logically relating this central category to other categories.
  • Data is selectively sampled to validate the central category and its relationships to other categories.
  • The researcher must watch out for other categories that may emerge from the new data that may be related to the phenomenon of interest (and engage in open coding).
  • Selective coding limits the range of analysis.

Based on: Strauss and Corbin (1998), discussed in Bhattacherjee (2012)

This process requires the researcher to deeply engage with the text, increasing the depth of understanding by a reiterative process of searching for relationships and interpretations. For example, if we have a set of interviews and our research question is “How people from different socio-economic groups experienced the pandemic?” we may use the category ‘feelings’, which can have such characteristics as ‘isolation’, ‘loneliness’ or ‘anger’, which we can further dimensionalize as severe, medium, or low. Using interviews to explore topics that are understudied or new (such as Covid-19 was at the beginning of the pandemic) may be particularly useful, considering that no previous research or data may be pertinent to this research question.

The next important step is integrating these categories and developing an answer to our research question. There are three major integration techniques, further explain in the graph below: storylining, memoing, or concept mapping.

Storylining

  • Categories and relationships are used to explicate and/or refine a story of the observed phenomenon

Memoing

  • Memoes are theorized write-ups of ideas about concepts and their theoretically coded relationships as they evolve during ground theory analysis.

  • Memoing is the process of using these memos to discover patterns and relationships between categories (using two-by-two tables, diagrams, or figures, or other illustrative displays).

Concept mapping

  • Graphical representation of concepts and relationships between those concepts (e.g., using boxes and arrows).

  • The major concepts are typically laid out on one or more sheets of paper, blackboards, or using graphical software programs.

Based on: Strauss and Corbin (1990), discussed in Bhattacherjee (2012)

Case Study 2

The module on Surveys / questionnaires proposes to collect quantifiable and standardized information, usually with a large sample (large-N). Hence, quantitative data analysis would be most useful for analyzing the gathered data. Let’s review one simple example:

Research topic: Students’ attitude in online learning

Timeframe of the survey: 2021 October-November

Survey results: see the table below

Respondent Gender Age  Did you participate in higher education during the pandemic? Are you satisfied with the quality of online education at your institution during the Covid Pandemic on a scale from 1 (not satisfied at all) to 5 (very satisfied)
1 M 18 Yes 5
2 F 19 Yes 4
3 M 19 Yes 5
4 F 19 Yes 3
5 F 20 Yes 2
6 F 21 Yes 1
7 F 20 Yes 5
8 M 19 Yes 4
9 F 18 Yes 3
10 F 18 Yes 5
11 F 19 Yes 3
12 M 19 Yes 3
13 M 19 Yes 2
14 F 19 No 3
15 F 20 Yes 1
16 F 20 Yes 5
17 F 20 Yes 4
18 F 20 Yes 4
19 M 21 Yes 3
20 M 50 Yes 3

What are some of the useful information to mention for our analysis?

First, we need to describe the student population that participated in this study. For that, we need to state the following:

  • There were 20 anonymous respondents to our survey, N=20
  • We need to give a gender distribution of the survey if we consider that gender might have played a role in how students viewed online education; in our case, there were 7 male respondents. Gender is an important factor as women were reportedly more burdened by house chores and childcare duties than men, which likely had an effect on their perception of online education. All other relevant factors must be included in the survey by the researcher.
  • We need to characterize the age of the respondents. Here, we can give for example give the average age of the respondents, that is 20.9. However, a perceptive researcher will also search for outliers – in our case there is one student who is 50 years old, clearly standing out from the rest of the students, who are between 18 and 21 years of age. Researchers might decide to exclude the outlier (respondent no. 20 who is 50 years old) and clearly communicate that when analysing the data. If we exclude the outlier, the average age is 19.4.
  • We note that all but one student attended a higher educational institution during the pandemic; the student who did not should be excluded from analysis, as their answer is not valid.
  • Finally, we look at students’ answers regarding their satisfaction with online education. Again, we can calculate average, which is 3.42 (after excluding the student who did not attend)
  • We can choose to look at further findings, such as whether female or male students were more satisfied with online education. For that, functions in Excel can be used, and for more complex data statistical software might be the best option.
  • Finally, to present findings, graphs or tables can be used to provide a better visual representation of findings. For example, a simple pie chart can show the share of votes regarding satisfaction (similarly using the Excel). We can also present the same information in a table, showing the percentage of respondents who voted for individual values (from 1 to 5). Depending on data, other types of charts can be used. Percentages might be better illustrations of findings from large-N studies, but in this example we used percentage for illustrative purposes.

Scale value Percentage
1 11%
2 11%
3 32%
4 21%
5 26%

There are multiple resources students can use to improve their skills in analysing statistical data, such as survey data. Below is a short list of source:

The Thematic website has a short article entitled “How to analyze survey data: best practices for actionable insights from survey analysis” available at https://getthematic.com/insights/analyze-survey-data-survey-analysis/ 

The Winston-Salem State University has a short handout on quantitative design available at https://www.wssu.edu/about/offices-and-departments/office-of-sponsored-p...