Which Of The Following Are Not Research Data: Complete Guide

Which of the Following Are Not Research Data?
So *The short version is: not everything you collect counts as research data. Knowing the line saves you time, headaches, and sometimes even ethics‑board trouble.

Ever stared at a spreadsheet full of numbers and wondered, “Do I really need to treat all of this as research data?That said, ”
Maybe you’ve been handed a batch of interview transcripts, a pile of sensor logs, and a stack of customer invoices. You’ll quickly discover that some of those files belong in a different folder—maybe even a different department No workaround needed..

Below we’ll walk through the kinds of material that look like data but, in most research contexts, are not considered research data. That's why we’ll also flag the gray zones where the answer depends on your discipline, funding agency, or institutional policy. By the end you’ll have a handy mental checklist for cleaning up your project files before you hit “Submit” on a grant or a journal manuscript.

What Is “Research Data” Anyway?

First, let’s get on the same page about what counts as research data. In plain language, research data are the factual records you collect, generate, or observe to answer a specific research question. They can be numbers, text, images, audio, video, or even physical samples—anything that you analyze to draw conclusions Small thing, real impact..

What makes something research data isn’t the format; it’s the purpose. If you’re measuring soil pH to test a hypothesis about crop yield, that pH reading is data. If you’re logging the time your coffee machine takes to brew because you’re curious about caffeine consumption trends, that log could be data—if you plan to analyze it scientifically.

Quick note before moving on.

The “research‑data” box

• Collected for a systematic investigation
• Documented with enough context to be understood later
• Stored in a way that preserves integrity (e.g., version control, backup)
• Intended for analysis, sharing, or replication

Anything that falls outside one of those criteria is a strong candidate for “not research data.”

Why It Matters

You might think, “It’s just a file, why does it matter if I call it data or not?”

Because the label determines how you treat it. Research data trigger:

1. Ethical oversight – Human‑subjects data often need IRB approval.
2. Data‑management plans – Funders demand a plan for storage, sharing, and preservation.
3. Metadata requirements – You have to describe the data so others can reuse it.

If you mistakenly flag a routine admin record as research data, you could end up writing a whole data‑management plan for something that never needed one. Conversely, if you overlook a piece that is data, you might breach compliance or lose reproducibility Not complicated — just consistent. Practical, not theoretical..

How to Spot What’s Not Research Data

Below is the meat of the article. We’ll break down the most common culprits, explain why they’re usually excluded, and note the exceptions.

### Administrative paperwork

Think of budget spreadsheets, purchase orders, and staff schedules Small thing, real impact. Still holds up..

• Why they’re not data: They serve operational purposes, not analytical ones.
• When they become data: If you’re studying organizational efficiency and those spreadsheets become the unit of analysis, they cross the line.

### Raw sensor “noise”

You’ve got a temperature logger that spits out a stream of readings every second. Somewhere in that stream are glitches—spikes caused by a loose cable, for example Simple, but easy to overlook..

• Why they’re not data: Noise isn’t information you intend to analyze; it’s a by‑product.
• When they become data: In engineering reliability studies, the frequency of sensor failures is the phenomenon under study.

### Marketing collateral

Brochures, ad copy, and press releases are often gathered for brand‑analysis projects.

• Why they’re not data: They’re produced for communication, not as evidence.
• When they become data: If you code the language of the brochures to test a hypothesis about messaging impact, the text becomes data.

### Personal notes and brainstorming sketches

Your coffee‑stained notebook full of ideas, doodles, and half‑finished equations.

• Why they’re not data: They’re a researcher’s private thought process, not a systematic record.
• When they become data: In qualitative research on the creative process, you might actually analyze those sketches as artifacts.

### Publicly available statistics (unless re‑analyzed)

Census tables, World Bank indicators, or published GDP figures.

• Why they’re not data: They’re already published, and you’re not generating new measurements.
• When they become data: If you combine them with your own observations in a novel model, the combined dataset is research data, but the original published tables remain secondary sources.

### Software logs that aren’t part of the experiment

Server uptime logs, error messages from a web app that’s not the object of study.

• Why they’re not data: They’re operational diagnostics.
• When they become data: In a study of software reliability, those logs become the primary evidence.

### Customer invoices and receipts

You might have a stack of receipts from a field trip or a set of invoices for lab reagents Not complicated — just consistent..

• Why they’re not data: They’re bookkeeping records.
• When they become data: If you’re researching procurement practices or cost‑effectiveness, they turn into data points.

### Training manuals and SOPs

Standard operating procedures are essential for reproducibility, but they’re not data themselves Most people skip this — try not to..

• Why they’re not data: They describe how you collect data, not the data itself.
• When they become data: When you evaluate the impact of different SOPs on outcome variability, the SOPs become variables in your analysis.

### Email threads (unless content‑coded)

A chain of emails coordinating a field study.

• Why they’re not data: They’re communication artifacts.
• When they become data: If you conduct a discourse analysis of the emails, each message is a unit of data.

Common Mistakes / What Most People Get Wrong

1. Treating every file as data because it’s digital.
   A PDF of a conference program isn’t data unless you’re studying conference structures It's one of those things that adds up. Worth knowing..

2. Assuming “big data” equals “research data.”
   A dump of click‑stream logs is only research data if you have a research question driving the analysis Which is the point..

3. Mixing raw and processed files without labeling.
   You might think a cleaned CSV is the same as the original sensor dump. In reality, the raw file is potential data; the cleaned version is derived data. Both need separate metadata.

4. Leaving out “negative results.”
   If an experiment yields no signal, the blank file is still data—just a different kind. People often discard it, which hurts reproducibility.

5. Over‑documenting admin stuff.
   You don’t need a data‑management plan for a receipt unless the receipt itself is an analytic unit.

Practical Tips – What Actually Works

• Create a “data inventory” early. List every file, its source, and the intended analytical use. Anything without a clear research purpose can be moved to an “administrative archive.”

• Use a two‑tier labeling system.

  • Tier 1: “Research data” – files you’ll analyze, share, or preserve.
  • Tier 2: “Support material” – everything else (admin, logs, drafts).

• Ask yourself the “why” question for each file.

  • Why did I collect this?
  • What hypothesis does it test?
    If the answer is “I needed it to run the experiment,” you probably have support material, not data.

• apply metadata templates. Even a simple spreadsheet with columns for “File name,” “Purpose,” “Format,” “Access restrictions,” and “Retention policy” can save you weeks later That's the whole idea..

• Separate “raw” from “derived.” Keep the untouched original in a read‑only folder. Anything you manipulate (e.g., normalized values) belongs in a “processed” subfolder with its own metadata.

• Check funder and institutional definitions. Some agencies (e.g., NIH) have very broad definitions that may pull in things you’d otherwise exclude. Align your inventory with those guidelines Most people skip this — try not to. Nothing fancy..

• Document the exclusion decision. Write a one‑sentence note for each non‑data file explaining why it’s excluded. Future auditors love that transparency.

FAQ

Q: If I’m using public datasets, do I need to treat the original files as my research data?
A: No. The public dataset remains a secondary source. Your derived dataset—what you actually analyze and possibly augment—is the research data you need to manage.

Q: Are interview transcripts always research data?
A: Usually, yes. But if the interview was just a pilot to test equipment and you never intend to analyze the content, the recordings can be classified as “support material.”

Q: How do I handle mixed‑purpose files, like a lab notebook that contains both experimental logs and personal thoughts?
A: Split the notebook into two documents or annotate sections clearly. Only the experimental logs become research data; the personal reflections go into a “research diary” archive That's the part that actually makes a difference..

Q: Do software version‑control logs count as data?
A: Only if you’re studying software development processes. Otherwise, they’re support material for reproducibility but not research data per se Not complicated — just consistent..

Q: What about photographs taken during fieldwork that also capture background scenery?
A: If the scenery isn’t part of the analysis, you can crop or annotate the image. The cropped, analysis‑relevant portion is data; the rest is ancillary.

Every time you finally close that project folder, you’ll see a tidy hierarchy: a clean “Research Data” tree, a “Support Materials” shelf, and a short note explaining why each piece landed where it did.

That clarity not only eases compliance and sharing, it also makes your own future self thank you when you revisit the work years later.

So the next time you stare at a mountain of files, ask yourself: Is this really data, or just a useful by‑product? The answer will shape how you store, protect, and ultimately publish your findings. Happy sorting!