Content analysis helps you answer a defined question about communication material through a documented coding procedure. Before collecting hundreds of texts, decide exactly what you will examine, how you will classify it and what the resulting evidence can establish.
Your coding decisions need to be visible in the written project. UniDraft helps organise your confirmed method, source records and findings into an editable Word document shaped by your supervisor's instructions. Start your project document with UniDraft; carry out the analysis yourself and bring the evidence you can explain.
Choose the question before the coding sheet
Columbia University's methods overview describes content analysis as coding communication material to examine concepts and their relationships. It distinguishes recording a concept's presence from counting its occurrences. Those choices answer different questions.
“How often do campus notices name a responsible office?” is different from “How do notices describe responsibility?” The first might use a predefined feature code. The second may need an interpretive approach with a clear account of how categories developed. Agree your approach before treating every text as a row in the same spreadsheet.
For a different qualitative workflow, compare the existing guide to thematic analysis. A different label alone does not establish a different method; explain the actual procedure you use.
Make a register of the material
Give every item a stable identifier. Record the source, publication date where available, retrieval date, relevant version and inclusion decision. Keep enough context to trace a coded extract back to the item.
| Register field | What to record |
|---|---|
| Item ID | A consistent identifier such as N01 |
| Source and version | Where the item came from; whether it was revised |
| Inclusion rule | Why this item belongs in the defined corpus |
| Exclusion reason | Why a candidate was left out |
| Access/context note | Missing pages, restricted access or uncertain date |
These are suggested record fields, not an institutional form. Public availability does not remove privacy, copyright or research-ethics responsibilities. For personal or restricted material, follow your institution's research ethics and consent process before collection or disclosure.
Separate the item from the part you code
A whole notice may be the selected item, while a sentence or explicitly defined feature is what you classify. Write down both decisions. If you count sentences in one item and whole documents in another without explanation, your totals become difficult to interpret.
For an interpretive study, preserve the surrounding meaning. Erlingsson and Brysiewicz's methodological paper explains qualitative content analysis through meaning units, codes and categories, with repeated movement back to the original text. Their interview-based exercise is a methodological illustration, not a rule that all undergraduate projects must use interviews or follow a fixed linear sequence.
A fictional coding-frame exercise
Everything in this exercise is invented for teaching. These are not real university notices, research findings or a recommended sample size.
Suppose the question is: “Which submission instructions are explicitly stated in the selected notices?” Use three features: required action, named receiving office and deadline. For this exercise only, give each feature one of three values:
- Explicit: the item itself states the feature clearly.
- Unclear: wording appears relevant but does not allow a definite classification under the rule.
- Absent: the feature is not stated in the item examined.
| Fictional extract | Feature | Proposed decision | Reason |
|---|---|---|---|
| “Submit the signed form to the Faculty Office by 12 May.” | Receiving office | Explicit | The destination is named |
| “Return it there before the closing date.” | Receiving office | Unclear | “There” has no identifiable referent in this invented extract |
| “Read the revised handbook before applying.” | Deadline | Absent | No deadline is given in this sentence |
The second row is deliberately a boundary case: opening the surrounding notice might resolve the referent. Do that before final coding. Absence from an isolated sentence is not necessarily absence from the complete document.
For each category, add inclusion rules, exclusion rules and at least one boundary example to your codebook. Do not silently turn “unclear” into “absent” because a binary chart looks cleaner.
Try the rules and record revisions
Use a manageable trial subset to find ambiguous instructions before applying the planned procedure across the corpus. Record the version, problem, decision and affected items. If a rule changes, identify earlier material that needs reconsideration.
Where your design includes multiple coders, agree how coding will be done, how disagreements are documented and which quality checks fit the method. Do not invent a second coder, report an agreement statistic you never calculated or copy a universal “passing” percentage into the methodology. Discuss the appropriate check with your supervisor.
Report what was actually examined
Keep selection decisions and analytic decisions distinct. A useful report can explain:
- What material was eligible and how you obtained it.
- What was included, excluded or unavailable, with reasons.
- What you coded and which version of the rules you used.
- How you checked ambiguous cases and revised decisions.
- What the resulting observations support, and their limitations.
If presenting counts, label the denominator and unit. “Items with a named office” and “mentions of an office” are different measures. Multiple features may occur in one item, so feature totals need not equal the corpus size. Do not manufacture example findings to fill the results chapter.
Use the results-chapter guide for presenting your verified evidence. Keep statements about selected texts separate from claims about an entire population, author intentions or how readers behaved.
Before you turn the analysis into a document
Check that another reader can follow the route from question to selected item, extract, coding decision and conclusion. Keep your register, codebook versions and illustrative evidence available, subject to confidentiality and permission. Explain material limitations instead of hiding missing documents or changing your original selection story.
When that evidence is ready, prepare your project document with UniDraft. Supply your confirmed methods, findings and instructions, then review every factual statement and reference. The document supports your explanation; it does not replace the analysis you must defend.
Sources
- Columbia University: Content Analysis, for the general coding and presence/frequency distinction.
- Erlingsson and Brysiewicz: A hands-on guide to doing content analysis, African Journal of Emergency Medicine, 2017, DOI 10.1016/j.afjem.2017.08.001, for its qualitative methodological explanation.