Coding Exercise

Southern Wesleyan University

1. Introduction
 “Qualitative data are often extensive, and careful data reduction is key. Too much

reduction and the integrity and detail of the data are lost; too little and data overload and loss of

clarity ensue” (Cohen et al., 2018, p. 668). This coding exercise required careful examination,

reduction, and categorization of 125 responses from 400 teachers who were asked, “What major

factors led you into teaching?” This paper will examine four distinct areas: my initial analysis of

the code; the rationale for selecting the overarching and subcategories, including a methodology

that explains my specific coding approaches, a synthesis of the data, theorizing on the data in

relation to the categories shared by Denee Dill, and a summary of findings.

2. Analysis of the Data

Researchers use codes as tags to categorize specific concepts or information within a text

(Miles and Huberman, 1994; Gläser and Laudel, 2013). When I first looked over the code, I

focused on the exact wording used. What caught my eye first was the significant number of

times the reason for teaching involved children, and the related terms such as “love, like, and

enjoy” compared to words such as “(make a) difference, help, desire, and impact.” Next, I

noticed similar terms of “love and enjoyment” tied to teaching and learning, along with

particular subjects such as reading, sports, swimming, or math. Thirdly, there were a notable

number of references to both teachers in one’s family and the impact of past instructors.

Additionally, I reviewed lifestyle reasons such as “summers off, the hours, and the lifestyle it

affords.” Finally, there were career and profession-related reasons, such as “139: I felt it would

be a service to people” and “233: I have always had a positive attitude towards school, and I

wanted to be part of the process.” My initial analysis of the data was very focused on the literal

words used in the text, rather than on the higher thematic categories they ultimately represented.

3. Sort or Categorize the Data

 Glaser and Strauss recommend first examining and categorizing the data while setting

aside the relevant literature or overarching theories (1967). I chose that approach and also

adopted responsive categorization or “...developed [categories] from the material itself” (Cohen

et al., 2018, p. 668). After faculty feedback, I created six mutually exclusive categories (with

various subcategories) that covered broader thematic areas: 1. Child-related: A. Fondness for

working with children, B. Desire to support or influence children. 2. Personal fondness for

teaching or subject area: A. Enjoyment of teaching, B. Enjoyment of a particular subject. 3.

Impact of past teachers: A. Positive, B. Negative. 4. Family and lifestyle reasons: A. Parenting,

B. Lifestyle, and C. Family-related history of teachers. 5. Profession: A. Career. 6.

Unclear/Undetermined.

To begin with, terms such as “love, like, and enjoy” when related to children were

categorized as a “Fondness for working with children,” while “(make a) difference, help, desire,

and impact” fell under the broader category of “Desire to support or influence children. Similar

terms for liking and enjoyment came up for both teaching and subject areas, such as “108: I love

to instruct,” “258: Always loved school and learning,” “74: Love for reading,” and “270: Love of

sports.” Later, after working through the entire dataset, I noticed a few responses attributable to

subpar teachers, so I split the data into positive and negative subgroups. I also found it helpful to

thematically group families and lifestyles together, as these responses often occurred and related

to one another. Lastly, the impact of family was much more nuanced and could be broken down

between both advantages for parenting and the impact of having a family history of teachers.

Finally, I created a category “Unclear/Underdetermined” for four responses that were extremely

difficult to discern the meaning, including “189: Calendar,” “209: All these things I felt as a

[para?],” “299: Acid rain,” and “305: Prayer.”

Methodology

Approaches to texts can either be lightly or densely coded, “...where a single piece of text

has several codes attached to it” (Cohen et al., 2018, p. 668). I chose to use a dense coding

approach since a notable proportion of the sample included multiple categories within a single

individual's responses. For example, response number 217 included three distinct categories.

“My mother owned a... day care... I saw the impact she made on... children... I wanted to have

children of my own, and teaching was a profession where I could make an impact and spend

quality.” Although there are two sentences/statements related to teachers' family histories, this

instance is coded only once. I consistently coded all responses by assigning each instance type

only once, since there were often explanations or additional details that could be tied back to an

original comment. Although there are two sentences on family history, they convey a single

overarching idea and should be demarcated only once in the coding scheme. Additionally, both

categories of parenting and desire to support or influence children are also coded. I consistently

use this coding pattern for all responses that include multiple categories within an individual

response.

4. Synthesis of data

Table 1
Total Number of Responses and Percentage of Coding Categories
____________________________________________________________________________

Overarching Categories and Sub-Categories Total Number of Responses Percentage

____________________________________________________________________________

1. Child-related Total 75 38.37%

A. Fondness for working with children 41 20.92%
B. Desire to support or influence children 34 17.35%

2. Fondness for teaching or subject area Total 26 13.27%

A. Enjoyment of teaching 12 6.12%
 B. Enjoyment of a particular subject 14 7.14%

3. Impact of past teacher Total 29 14.80%

A. Positive 23 11.73%
B. Negative 6 3.06%

4. Family and lifestyle reasons Total 39 19.90%

A. Parenting 13 6.63%
B. Lifestyle 13 6.63%
C. Family history of teachers 13 6.63%

5. Profession Total 23 11.73%

A. Career 23 11.73%

6. Unclear/Undetermined 4 2.04%

Grant Total 196

____________________________________________________________________________

After coding all 125 responses, there was a grand total of 196 distinct codes across six

distinct categories, including: 1. Child-related (75/196) 38.37%, 2. Fondness for teaching or

subject area (26/196) 13.27%, 3. Impact of past teacher (29/196) 14.80%, 4. Family and lifestyle

reasons (39/196) 19.90%, 5. Profession (23/196) 11.72%, and 6. Unclear/Undetermined (4/196)

2.04%. I divided the total number in the distinct category (or subcategory) by the total number of

distinct codes, or 196. The most cited reason for entering teaching, in 38.27% of the sample,

pertained to children, with slightly more responses related to fondness for children (41/196 =

20.92%) than to a desire to support or influence youth (34/196 = 17.35%). The second-highest

category was related to family and lifestyle reasons, with (39/196) 19.90%, and an even split of

13 responses (6.63%) for parenting, lifestyle, and family history of teachers. Impact of past

teachers (29/196), at 14.80%, had slightly more responses than fondness for teaching or subject
area (26/196), at 13.27%. Finally, the lowest category outside of unclear/undetermined (4/196 =

2.04%) was related to profession/career, with only (23/196) 11.73%.

5. Theorize based on the Data

Thomas (2006) describes the value of using blind parallel coding or having another

researcher code the raw data, and comparing shared areas and inconsistencies. I received a

response from my classmate Denee Dill, who preliminarily coded her data as follows: 1.

Mission/social impact, 2. Lifestyle/family and schedule, 3. Affinity, 4. Personal influence:

family/mentor/ teacher (positive or negative), 5. Passion for teaching and learning, 6.

Content/subject area aptitude, and 7. Career change/redirection.

I believe Denee’s “Mission/social impact” corresponds with my category “Desire to

support or influence children,” “Personal influence: family/mentor/teacher” overlaps with

“Impact of past teachers,” “Lifestyle/family” with “Family and lifestyle reasons,” and finally,

“Content/subject area aptitude and passion for teaching and learning” with “Fondness for

teaching or subject area.” Our main differences include that there are no similar categories for

“affinity” and that, instead of a general area of “Profession/career,” Denee has focused more on

“Career change/redirection.” Some other subtle discrepancies also relate to “personal influence”,

as I don’t have a “mentor” category. Differences in our data set categorization could be due to

our familiarity with qualitative coding, our coding approach (e.g., light or dense), years of

education, or professional background. Overall, our categories strongly overlap, with significant

agreement. I do not have precise information on which categories have the most responses, but

based on my data, I can infer that “Mission/social impact” most likely has the highest number of

responses.
 Denee later updated her coding to include additional categories, sub-categories, and more

nuanced information: 1. Altruistic Mission & Social Impact (Desires to effect change, serve

society, or improve the future lives of students). 2. Pragmatic Lifestyle and incentives (Benefits

such as academic calendar, coaching, and family compatibility). 3. Relational (Personal

enjoyment of working with/being around children) 4. Interpersonal Influence & Legacy

(External inspiration from family members, former teachers, or mentors (both positive and

negative). 5. Content& process/engagement (Passion for subject matter or the "lightbulb"

moment) 6. Vocational Aptitude (Perceived natural talent or prior successes in teaching-related

roles like tutoring or babysitting) 7. Economic Necessity & Redirection (factors such as job

burnout, lack of other prospects, or a desire for intrinsic value) 8. No response. This set of codes

shows a significant level of improvement, indicating both a much more detailed understanding of

the data set and a much higher degree of similarity to my personal coding categories. However,

one unique difference is Denee’s expanded vision for “Vocational Aptitude” and “Economic

Necessity & Redirection,” which ties to my less detailed “Profession/Career.” Seeing Denee’s

additional categories and subcategories makes me realize that I could have broken down

“Profession/Career” into multiple subcategories related to aptitude, economic reasons, and

redirection.

6. Summary of Findings

Although Denee and I have a significant degree of overlap in our coding categories,

individual researchers will still differ in their categorization, codes, and data analysis. “Codes

can be at different levels of specificity and generality when defining content and concepts...

Some codes are very general; others are more specific” (Cohen et al., 2018, p. 669). For

example, in my initial review of the data, I focused very heavily on specific words such as
“(make a) difference, help, desire, and impact” which were originally distinct categories. Later, I

learned the value of coverage and mutual exclusivity, so I combined them into the overarching

category of “Desire to support or influence children.” Denee instead chose a much broader, and

perhaps more abstract, thematic category of “Mission/social impact.” She currently works as the

Associate Director of the Cerone Center for Learning, so in her leadership role, terminology such

as “Mission/social impact” may be more familiar or accessible. Our professional roles and

backgrounds can introduce bias and influence our coding decisions (Smith & Noble, 2025).

From an ethical standpoint, blind coding with multiple participants can help reduce bias arising

from personal backgrounds or professional affiliations” (Cohen et al., 2018). I'd like to see how

Tiffany coded her data and whether her categories are more similar to mine, as a fellow English

teacher, or to Denee, who works in educational leadership. Ultimately, qualitative research and

coding are more influenced by interpretation and potential bias, underscoring the researcher’s

need to be particularly aware of their personal coding rationale, decisions, and categorizations

(Smith & Noble, 2025).

7. Conclusion

This assignment provided a rigorous and hands-on opportunity to review best practices in

qualitative coding, carefully examine a large data set, and justify the rationale for coding and

respective categories. At first, my categories were overly specific and overlapping, but with

faculty feedback and careful review of the readings, I successfully defined six overarching

categories. The methodology section helped me to both apply and explain my approach to

coding. Consistently applying coding is imperative for accurate data tabulation. Comparing my

data with Denee's helped me see both areas of overlap and discrepancy. In contrast, the summary

of findings helped me reflect on differences between researchers and potential sources of bias.
Ultimately, this assignment provided an impactful opportunity to reflect upon, categorize, and

analyze qualitative data.

References

Cohen, L., Manion, L. & Morrison, K. (2018). Research Methods in Education (8th edition).
 New York, NY: Routledge.

Gläser, J. and Laudel, G. (2013) Life with and without coding: Two methods for early‐stage data

analysis in qualitative research aiming at causal explanations. Forum Qualitative

Sozialforschung/Forum: Qualitative Social Research, 14 (2), pp. 1–24.

Miles, M. B. and Huberman, A. M. (1984). Qualitative Data Analysis. Beverly Hills, CA: Sage.

Smith, J., & Noble, H. (2025). Understanding sources of bias in research. Evidence-based

nursing, 28(3), 137–139. [Link]

Thomas, D. R. (2006). A general inductive approach for analyzing qualitative evaluation data.

American Journal of Education, 27 (2), pp. 237–46.