Working with Computers: Essential Skills for Researchers

Introduction

This tutorial introduces essential computer skills for researchers, students, and anyone who uses a computer for academic or professional work. It covers keeping your machine healthy and performant, organising files and data systematically, protecting your work with good security practices, managing research data responsibly, and working productively — including the responsible use of AI tools that are increasingly embedded in research workflows.

The goal is not to turn you into a systems administrator, but to give you the practical knowledge and habits that prevent the most common and most costly computing problems researchers encounter: lost data, slow machines, security breaches, and hours spent searching for files that should be easy to find.

Learning Objectives

By the end of this tutorial you will be able to:

1. Explain why computers slow down and apply a regular maintenance routine to prevent it
2. Organise files and folders using a systematic, sustainable naming and hierarchy system
3. Describe the main categories of security risk and apply appropriate protections
4. Create strong passwords and use a password manager
5. Enable and use multi-factor authentication
6. Organise a research project folder in a reproducible, shareable way
7. Apply the 3-2-1 backup rule to protect your research data
8. Describe best practices for working with AI tools such as ChatGPT, Copilot, and Claude
9. Identify and avoid common computing mistakes before they happen

No technical expertise required — this tutorial is designed for everyone.

Citation

Schweinberger, Martin. 2026. Working with Computers: Essential Skills for Researchers. Brisbane: The Language Technology and Data Analysis Laboratory (LADAL). url: https://ladal.edu.au/tutorials/comp/comp.html (Version 2026.05.01).

---

Part 1: Understanding the Problems

Section Overview

What you will learn: The most common computer problems researchers face, why they occur, and the reassuring fact that almost all of them are preventable

Common computer problems

Most researchers encounter the same set of recurring problems.

Performance problems: The machine runs slowly, programs take a long time to start, the computer gets very hot, the fan runs loudly, or the system freezes and crashes. These problems are almost always caused by accumulated clutter, too many background processes, dust blocking ventilation, or hardware that is due for an upgrade.

Organisation problems: Files cannot be found when needed. Duplicate versions exist with names like final.docx, final_FINAL.docx, and final_FINAL_use_this_one.docx. Projects are scattered across multiple locations. Hours are lost searching for documents that should take seconds to find. These problems stem from not establishing a system early and not maintaining it consistently.

Security problems: Sensitive data is not protected. Passwords are weak or reused. Phishing emails are acted on. A single data breach compromises many accounts. These problems are entirely preventable with the right habits and tools.

Data loss: Files are accidentally deleted, overwritten, or lost when a hard drive fails. Without backups, years of work can disappear permanently. This is the most catastrophic and most easily preventable category of problem.

Good news

All of these problems are preventable with the right knowledge and habits. This tutorial will show you exactly what to do.

Exercises: Understanding the problems

Q1. A researcher stores all her working files on her desktop and has not restarted her computer in three weeks. She notices her machine is slow, her boot time has increased from 30 seconds to over 5 minutes, and she cannot find a manuscript draft she wrote last month. Which of the following best describes the combination of problems she is experiencing?

A security problem — her accounts have probably been compromised
Performance problems (caused by startup clutter and no restart to clear temporary files) and organisation problems (caused by storing files on the desktop without a systematic naming or folder structure). The slow boot and heat are classic signs of too many startup items and accumulated temporary data; the missing manuscript is a file organisation problem.
A hardware failure — she needs a new computer
A data loss problem — the file has been permanently deleted

---

Part 2: Keeping Your Computer Healthy

Section Overview

What you will learn: Why computers slow down; the essential maintenance tasks that keep them running well; how to monitor performance; and how to manage startup programs, storage, antivirus protection, and physical cleanliness

Why computers slow down

A useful analogy: your computer is like a physical workspace. A tidy desk with clear surfaces and organised drawers lets you work efficiently. A desk buried under stacks of paper, with drawers jammed full of old documents and a coffee ring on every surface, slows everything down. Computers work the same way.

The main causes of performance degradation are:

• Too many startup programs — each program that loads when you turn on the computer adds time. A clean startup takes 20–30 seconds; a cluttered one can take 5 minutes or more.
• Accumulated temporary files — the operating system and applications create thousands of temporary files over time. These take up space and slow file access.
• Malware and background processes — hidden programs consume CPU, memory, and network bandwidth.
• Overheating — dust blocks airflow, causing the processor to automatically reduce its speed to avoid damage (called thermal throttling).
• Outdated software — older versions are less efficient and contain security vulnerabilities.
• Insufficient RAM — if you are running more applications than your machine can hold in memory, it starts using the hard drive as overflow, which is dramatically slower.

Essential maintenance tasks

1. Restart regularly

Sleep mode is not the same as a restart

Many people put their computer to sleep at the end of the day and never restart it. This is one of the most common causes of performance problems.

Sleep mode keeps everything in memory, accumulates temporary data, does not install system updates, and maintains open connections that can represent security risks.

A restart clears temporary files, installs pending security updates, refreshes system resources, and closes any processes that have become stuck or bloated over time.

Restart at least once a week. Restart immediately when prompted to complete a security update.

How to restart properly

On Windows: Start menu → Power → Restart (not Sleep or Shutdown, which do not always apply updates).

On macOS: Apple menu → Restart → confirm.

On Linux: Applications menu → Restart, or from the terminal: sudo reboot.

2. Keep software updated

Software updates are not just about new features — the majority are security patches that close vulnerabilities actively being exploited by attackers. Delaying a security update is a genuine risk.

Checking for updates

On Windows 10/11:

Settings → Windows Update → Check for updates

On macOS:

System Settings → General → Software Update

On Linux (Debian/Ubuntu):

sudo apt update && sudo apt upgrade

What to update and when

Critical or security updates should be installed immediately. Operating system, antivirus, and browser updates are the highest priority. Research software (R, Python, SPSS) should be kept reasonably current. Enable automatic security updates for your operating system and browser at minimum.

3. Monitor performance

Windows — Task Manager

Press: Ctrl + Shift + Esc

macOS — Activity Monitor

Applications → Utilities → Activity Monitor

Linux — System Monitor or htop

htop    # in the terminal

Interpreting the numbers

CPU (Processor)

• Below 20%: idle or light use — normal
• 20–50%: regular use — normal
• 50–80%: heavy use during active tasks — acceptable
• 80–100% sustained: a process is consuming too much — investigate

Memory (RAM)

• Below 70% used: comfortable
• 70–90%: acceptable, consider closing unused programs
• Above 90% sustained: you are running out of memory — close applications or consider a RAM upgrade

Disk

• Occasional spikes: normal (indexing, updates, etc.)
• Sustained 100% for more than a few minutes: a problem — check for indexing processes, malware, or a failing drive

If you see sustained high CPU or memory usage, open the Processes tab in Task Manager, sort by CPU or memory, and identify which application is responsible. Often a browser with many open tabs, or a background update process, is the culprit.

4. Manage startup programs

Many applications register themselves to start automatically at boot, even when you do not need them running all the time. Each one adds seconds to startup and consumes memory while running.

Managing startup on Windows

1. Open Task Manager (Ctrl + Shift + Esc)
2. Click the “Startup apps” tab
3. Review the list — Task Manager shows a “Startup impact” rating (High, Medium, Low)
4. Right-click any program you do not need at startup and select “Disable”

What to disable: Programs you only open occasionally (Spotify, Skype, Steam, Adobe updaters, cloud sync tools you do not need constantly).

What to keep: Antivirus software, essential drivers, and any security tools. When in doubt, search the program name to understand what it does before disabling.

Managing startup on macOS

System Settings → General → Login Items

Remove items from the “Open at Login” list that you do not need immediately on startup.

5. Clean your storage

Over time, computers accumulate significant quantities of junk: temporary files, download caches, old system files, and duplicates. Reclaiming this space improves performance and prevents “low disk space” warnings.

Windows — Disk Cleanup

1. Open File Explorer
2. Right-click the C: drive → Properties → Disk Cleanup
3. Check: Temporary files, Downloaded program files, Recycle Bin, Temporary Internet files
4. Click “Clean up system files” to also remove Windows Update leftovers

Windows — Storage Sense (automatic)

Settings → System → Storage → Configure Storage Sense

Enable this to automatically delete temporary files after 30 days and empty the Recycle Bin after 30 days.

macOS

Apple menu → System Settings → General → Storage

Options include removing watched streaming content, reviewing large files, and emptying the Trash automatically.

Linux (Debian/Ubuntu)

sudo apt autoremove      # remove unused packages
sudo apt clean           # clear package cache

Third-party cleaning tools

Tools like CCleaner can be useful for cleaning temporary files, but use them with care. Run “Analyze” before cleaning so you can review what will be deleted. Do not use registry cleaners unless you know exactly what you are doing — an incorrect registry change can break your operating system. Always back up important data before running any cleaning tool.

6. Use antivirus software

Antivirus is not optional

Millions of new malware variants appear every day. Ransomware can encrypt all your files and demand payment. Spyware steals passwords and research data silently. Phishing attacks become more sophisticated every year. A computer without active antivirus protection is almost certain to be infected eventually.

Built-in protection

Windows 10 and 11 include Windows Defender (now called Microsoft Defender Antivirus), which provides reasonable baseline protection and is automatically kept up to date. It is a good starting point.

macOS includes XProtect and Gatekeeper. While macOS is historically less targeted, it is not immune — macOS malware is increasingly common.

Additional options

• Malwarebytes Free — excellent at detecting and removing existing malware; good as a second-opinion scanner alongside your primary antivirus
• Avira Free — strong malware detection with a light system footprint
• Institutional antivirus — many universities provide enterprise-grade endpoint protection (such as Symantec Endpoint Protection or CrowdStrike) to staff and students at no cost; check with your IT department

Recommended scanning schedule

Real-time protection runs automatically and continuously. Run a quick scan weekly. Run a full system scan monthly and after any suspicious activity (unexpected slowdowns, strange pop-ups, unusual network activity).

7. Manage power settings

Appropriate power settings extend battery life, reduce heat, and prevent unexpected shutdowns during long analyses.

Windows

Control Panel → Power Options

• Balanced — the right choice for most everyday work; adjusts automatically
• High Performance — useful for computationally intensive tasks (large data analyses, video rendering); use only when plugged in
• Power Saver — reduces performance to extend battery; good for basic tasks away from power

macOS

Battery preferences are in System Settings → Battery. Enable “Low Power Mode” when away from power for extended periods, and “Prevent automatic sleeping when the display is off” when running long analyses while plugged in.

8. Keep the computer physically clean

Dust is a serious problem

Dust blocks ventilation, causes overheating, triggers thermal throttling (automatic performance reduction to prevent damage), and shortens hardware lifespan from years to months. The first signs are a louder fan, the computer running hotter than usual, and unexpected shutdowns.

Monthly: Wipe the exterior and keyboard with a microfiber cloth. Clean the screen with a proper screen cleaner (not window cleaner, which damages coatings). Check that vents are not blocked.

Every six months: Use compressed air to clean vents and fans. Hold the can upright and use short bursts — never a continuous spray, and never with the can tilted, as liquid propellant can damage electronics. Work outdoors or in a well-ventilated space.

For laptop interiors: only proceed if you are comfortable doing so; otherwise take the machine to a technician. Power off completely, unplug, remove the battery if possible, then use short bursts of compressed air while holding the fan still (spinning fans generate electricity that can damage the motherboard).

For desktop interiors: power off, unplug, open the case, ground yourself by touching the metal chassis, and use compressed air on fans, heat sinks, and vents.

9. Avoid bundled software and malware

Free software downloads frequently bundle unwanted additions: browser toolbars, adware, changed search engines, and fake “optimisation” tools that actually slow your computer down. These are not technically viruses, but they consume resources, collect data, and are difficult to remove.

How to avoid them:

During installation, always choose “Custom” or “Advanced” installation rather than “Express” or “Recommended.” Read every screen carefully. Uncheck any pre-ticked boxes for additional software, toolbar installations, or browser changes. Decline anything you did not specifically seek out.

Download software only from official sources: the developer’s own website, your institution’s software portal, or trusted package managers (the Mac App Store, the Microsoft Store, or package managers like Homebrew on macOS or apt on Linux). Be cautious with general download aggregator sites.

Red flags that indicate a suspicious download site:

• Multiple “Download” buttons scattered across the page, most of which are advertisements
• Promises of “speed boost,” “PC optimisation,” or “free registry fix”
• Unexpected download dialogue boxes that appear without you clicking anything
• Misspelled or unusual domain names

Summary: maintenance schedule

Regular maintenance checklist

Every week (5 minutes)

• Restart the computer fully
• Check for pending updates and install critical ones
• Empty the Recycle Bin or Trash
• Glance at Task Manager / Activity Monitor for anything unusual

Every month (10 minutes)

• Run an antivirus scan
• Run Disk Cleanup / Storage management
• Review startup programs — disable any newly added ones
• Wipe keyboard and screen
• Check available storage space

Every six months (30 minutes)

• Full antivirus scan
• Deep storage cleanup
• Physical cleaning with compressed air
• Review and reorganise files
• Verify backups are working

---

Part 3: File Organisation

Section Overview

What you will learn: Why file organisation matters; why the desktop is not a storage location; how to build a systematic folder hierarchy; file naming conventions that make your work findable and sortable; and how to use cloud storage responsibly

Why organisation matters

Studies of knowledge workers consistently find that people spend between 20 and 30 minutes each day searching for files. For a researcher working 48 weeks per year, that is 80–120 hours annually — the equivalent of two to three full working weeks — spent not on research, but on searching for documents that should take seconds to find.

Beyond the time cost, poor organisation causes missed deadlines (because the right file could not be located), lost work (because the wrong file was deleted), version confusion (because multiple near-identical copies exist), and the professional embarrassment of sending a colleague an outdated draft.

All of this is avoidable with a simple, consistent system established from the beginning of a project and maintained throughout.

Why the desktop is not storage

Do not store data on the desktop

When your computer starts, the operating system loads first, and then the desktop loads — with every file and folder on it. Each item on the desktop is checked, indexed, and rendered at startup. Ten files adds about two seconds; one hundred files adds twenty seconds or more; a cluttered desktop with hundreds of files can add minutes to every boot.

The desktop is a workspace for shortcuts and currently active items — not permanent storage. Similarly, avoid storing data directly on the C: drive root (Windows) or top-level home directory; keep data in dedicated folders within your Documents or user home area.

The right approach:

• Store working files in a well-organised folder structure within your Documents folder or equivalent
• Create shortcuts on the desktop to folders you access frequently — shortcuts are tiny and do not affect startup speed
• When a project is complete, archive it off the desktop entirely

Recommended folder structure

A hierarchical folder structure should reflect how you actually think about your work. Here is a practical structure for a researcher:

Documents/
├── Projects/
│   ├── Project_2024_TopicModeling/
│   │   ├── 01_data/
│   │   │   ├── raw/
│   │   │   └── processed/
│   │   ├── 02_scripts/
│   │   ├── 03_outputs/
│   │   │   ├── figures/
│   │   │   └── tables/
│   │   ├── 04_documentation/
│   │   └── README.md
│   ├── Project_2023_CorpusAnalysis/
│   └── ...
├── Teaching/
│   ├── 2024_S1_LING101/
│   ├── 2024_S2_LING201/
│   └── ...
├── Admin/
│   ├── Grants/
│   ├── Reviews/
│   └── ...
└── Reference/
    ├── Literature/
    ├── Methods/
    └── ...

The key principles: every project lives in one self-contained folder; data, scripts, outputs, and documentation are clearly separated; numbered prefixes (01_, 02_) force folders to display in logical order in any file browser.

File naming conventions

Naming rules that actually work

The goal: a filename that tells you what is in the file, when it was created, and what version it is — without opening it.

The rules:

1. Date first, in ISO format — YYYY-MM-DD sorts chronologically in any file browser
2. Use underscores or hyphens, not spaces — spaces cause problems in terminal commands, R paths, and some operating systems
3. Be descriptive — the name should be meaningful six months from now
4. Include a version indicator — v1, v2, draft, submitted, revised
5. No special characters — avoid ! ? * " < > | \ / : ;
6. Always include the file extension

Bad names:

final.docx
finalFINAL.docx
final_FINAL_use_this_one.docx
new doc.docx
john's thing 2.r

Good names:

2024-02-08_manuscript_draft-v1.docx
2024-02-15_manuscript_draft-v2.docx
2024-02-20_manuscript_submitted.docx
2024-03-01_manuscript_revised.docx

For research data files:

2024-01-15_experiment-01_participant-01_raw.csv
2024-01-15_experiment-01_participant-02_raw.csv
2024-01-20_experiment-01_all-participants_processed.csv

The ISO date format (YYYY-MM-DD) is particularly important. Unlike 02-08-2024 (which is ambiguous — is it February 8th or August 2nd?), 2024-02-08 sorts correctly in any context, is unambiguous internationally, and looks the same to everyone.

Cloud storage

Cloud storage provides automatic off-site backup and convenient access from multiple devices. The main options differ primarily in how much free storage they provide and how well they integrate with different software ecosystems:

Service	Free storage	Best for
OneDrive	5 GB (1 TB with Microsoft 365)	Office documents, Windows integration
Google Drive	15 GB	Mixed file types, collaboration
Dropbox	2 GB	Cross-platform sync
iCloud Drive	5 GB	macOS/iOS integration
Institutional storage	Often generous	Sensitive or large research data

Sensitive data does not belong in commercial cloud storage

Commercial cloud services (Google Drive, Dropbox, OneDrive personal) are not appropriate for personally identifiable information, medical or health data, financial records, unpublished confidential research data, or exam materials and grades. The terms of service for these platforms grant the provider rights to process your data, and their security standards may not meet institutional requirements.

For sensitive research data, use your institution’s approved research data management infrastructure. Most universities provide purpose-built secure storage for research data — check with your research computing or IT team.

Best practices for cloud storage:

• Establish your folder structure locally before connecting cloud sync — organise first, then sync
• Use selective sync: only sync active projects, not your entire archive
• Review and archive completed projects regularly rather than letting cloud storage grow indefinitely
• Understand what your sync client does when you delete a file (most have a recovery window, but not indefinitely)

Exercises: File organisation

Q2. A researcher names her files as follows: notes.docx, notes2.docx, notes_final.docx, notes_FINAL2.docx, notes_use_this.docx. Three months later she needs to find the most recent version. What is the fundamental problem with this naming system, and how could she have avoided it?

She used the wrong file format — she should have used .txt instead of .docx
The files are too large and should be split into smaller documents
The names contain no date or version number that sorts correctly, making it impossible to determine which file is most recent without opening each one and checking its modification date. The fix is to include the date in ISO format (YYYY-MM-DD) at the start of each filename and use sequential version numbers: 2024-01-10_notes_v1.docx, 2024-01-15_notes_v2.docx, 2024-02-01_notes_v3.docx. This way, the most recent version is always the last file when sorted alphabetically.
She should have stored all versions in the cloud so the modification date would be tracked automatically

---

Part 4: Security and Privacy

Section Overview

What you will learn: How to encrypt sensitive files and drives; how to create and manage strong passwords with a password manager; how to use multi-factor authentication; and how to browse the web safely

Encryption

Encryption scrambles data so that it is unreadable without the correct key. Without encryption, anyone with physical or remote access to your storage can read your files. With full-disk encryption, even a stolen laptop reveals nothing.

When to encrypt:

• Sensitive research data (interviews, clinical data, personally identifiable information)
• Unpublished manuscripts and grant applications
• Exam papers, grades, and assessment materials
• Any device that leaves your desk (laptops, USB drives, external hard drives)

Encrypting individual files and folders

Windows — EFS (Encrypting File System)

1. Right-click the file or folder → Properties → Advanced
2. Check “Encrypt contents to secure data” → OK
3. When prompted, back up your encryption key to a USB drive and store it securely, separate from the computer

Back up your encryption key — this is critical

If you encrypt files with EFS and lose the encryption key (for example, if you reinstall Windows or the user account is deleted), the files are permanently inaccessible. No recovery is possible. Back up the key immediately when prompted, store it in a separate secure location, and test that it works.

Windows — BitLocker (full disk encryption)

Settings → Update & Security → Device encryption

or

Control Panel → BitLocker Drive Encryption

Requires Windows 10/11 Pro and a TPM chip. Save the recovery key to your Microsoft account or print it — store it somewhere other than the encrypted drive.

macOS — FileVault (full disk encryption)

System Settings → Privacy & Security → FileVault → Turn On FileVault

FileVault encrypts the entire drive. Save the recovery key securely. Encryption takes several hours initially but runs transparently thereafter.

Linux — LUKS

Most Linux distributions offer full-disk encryption as an option during installation. For an existing system, individual folders can be encrypted with tools like VeraCrypt (cross-platform, also available for Windows and macOS).

Portable encryption: VeraCrypt

VeraCrypt creates encrypted containers that work on Windows, macOS, and Linux. It is a good option for encrypting USB drives or specific project folders that need to be portable. The free and open-source tool is available at veracrypt.fr.

Password management

The password problem

The average person has over 100 online accounts requiring passwords. Research consistently shows that most people reuse the same password across many accounts. The consequence: when one service is breached and your password is stolen, every account using that password is compromised.

Billions of username and password combinations from past breaches are freely available to attackers. You can check whether your email address appears in known breaches at haveibeenpwned.com.

Creating strong passwords

What makes a password weak:

• Common words or phrases (password, iloveyou, sunshine)
• Personal information (name, birthday, pet’s name)
• Short length — an 8-character password can be cracked by a modern computer in minutes
• Predictable patterns (Password1!, Summer2024)

What makes a password strong:

Length is the single most important factor. A 20-character passphrase of random common words (correct-horse-battery-staple) is far stronger than an 8-character string of mixed case, numbers, and symbols. This is because the number of possible combinations grows exponentially with length.

Good examples:

CatsAndDogsAreAwesomeIn2024!
correct-horse-battery-staple
My$onTurn3d5inJan2024

Use a unique password for every account. If one password is compromised, no other account is affected. In practice, this means using a password manager.

Password managers

Password managers: the practical solution

A password manager stores all your passwords in an encrypted vault. You remember one strong master password; the manager generates, stores, and auto-fills a unique strong password for every account.

This is not just convenient — it is the only practical way to use a unique strong password for every account.

Recommended options:

Manager	Free tier	Cost	Notes
Bitwarden	Full features	$10/year for premium	Open-source; recommended for most people
1Password	14-day trial	$36/year	Strong family/team sharing
KeePassXC	Full features	Free	Open-source; local storage only; no cloud sync
ProtonPass	Basic features	$48/year	From the makers of ProtonMail; strong privacy focus

Getting started with Bitwarden (or any manager):

1. Create an account and set a master password — at least 20 characters, a memorable passphrase
2. Install the browser extension and mobile app
3. Begin adding accounts — prioritise email, banking, and institutional accounts first
4. When you log in to any site, let the manager save the password; over time, update to manager-generated passwords
5. Write your master password down and store it in a physically secure location. If you lose your master password, you lose access to all stored passwords. Do not store it digitally.

Multi-factor authentication

Multi-factor authentication stops 99.9% of automated attacks

Even if an attacker obtains your password, multi-factor authentication (MFA) prevents them from logging in. MFA requires a second proof of identity beyond your password — something you have (a phone or hardware key) or something you are (a fingerprint).

Enable MFA on every account that supports it. Prioritise: email, institutional accounts, banking, cloud storage, and password manager.

Types of MFA, from weakest to strongest:

• SMS codes — convenient but vulnerable to SIM-swapping attacks where an attacker redirects your phone number; better than nothing, but not the best option
• Authenticator apps — generate time-based one-time passwords that expire every 30 seconds; works offline; recommended options include Microsoft Authenticator, Google Authenticator, and Authy
• Hardware security keys — a physical USB or NFC device (such as a YubiKey); the most secure option; immune to phishing; recommended for high-value accounts
• Biometrics — convenient but cannot be changed if compromised; use as an additional factor rather than a replacement for others

How to enable MFA on any account:

1. Go to account settings → Security (or Privacy & Security)
2. Look for “Two-factor authentication,” “Two-step verification,” or “Multi-factor authentication”
3. Choose an authenticator app (recommended) and scan the QR code with your phone
4. Save the backup codes provided — store them in your password manager or printed in a secure location
5. Test the login process before closing the setup page

Safe browsing

Check links before clicking. Hover over any link to see the actual URL before you click. Look for https:// (not http://). Check that the domain is correct — g00gle.com (with zeros) is not Google. Be particularly suspicious of shortened URLs (bit.ly, t.co) in unexpected emails.

Be cautious on public WiFi. Open WiFi networks (cafés, airports, conference venues) are not encrypted — anyone on the network can intercept unprotected traffic. Use your phone’s mobile hotspot instead when possible, or use a VPN (Virtual Private Network) on untrusted networks. University eduroam networks are encrypted and safe.

Recognise phishing. Phishing emails impersonate trusted organisations to steal credentials. Warning signs: unexpected requests for login credentials or payment, urgent language (“your account will be closed”), links that do not match the claimed sender, emails from slightly wrong domains. When in doubt, go directly to the website by typing the address rather than clicking the link.

Useful browser extensions:

• uBlock Origin — blocks ads and many malicious scripts; significantly reduces exposure to malvertising
• Privacy Badger — blocks trackers that follow you across sites
• Modern browsers (Chrome, Firefox, Edge, Safari) now include HTTPS enforcement by default; no extension needed for this

Exercises: Security

Q3. A colleague tells you she uses the same strong 12-character password for all her accounts, including email, banking, and institutional login. She argues that since it is a strong password, this is safe. What is the flaw in her reasoning, and what should she do instead?

There is no flaw — a strong 12-character password cannot be cracked
The flaw is that 12 characters is not long enough; she should use 16 characters instead
The flaw is that password strength only matters if each password is unique. When a service she uses is breached — which is a matter of when, not if — attackers obtain her email and password combination and immediately try it on other services (called credential stuffing). Because she reuses the same password, all her accounts are compromised at once. She should use a password manager to generate and store a different strong password for every account, so a breach of one service cannot cascade to others.
The flaw is that she does not use multi-factor authentication, which makes her password irrelevant

---

Part 5: Data Management for Research

Section Overview

What you will learn: How to structure a research project folder; the principle of never editing raw data; how to write a useful README file; version control basics; and how to back up your data using the 3-2-1 rule

Research project folder structure

A well-organised research project folder makes your work reproducible, easier to share with collaborators, and comprehensible to your future self. The key insight is that a project folder should be self-contained: everything needed to understand and reproduce the work — data, scripts, outputs, documentation — should be inside it, in a predictable location.

ProjectName_YYYY/
├── 00_admin/
│   ├── ethics/
│   ├── consent_forms/
│   └── correspondence/
├── 01_planning/
│   ├── proposal.docx
│   ├── methodology.docx
│   └── timeline.xlsx
├── 02_literature/
│   ├── pdfs/
│   ├── notes/
│   └── bibliography.bib
├── 03_data/
│   ├── raw/              ← NEVER edit these files
│   │   ├── README.txt    ← Document the data source
│   │   └── original_data.csv
│   ├── processed/
│   │   ├── 2024-02-01_cleaned.csv
│   │   └── 2024-02-05_analyzed.csv
│   └── metadata/
│       └── codebook.xlsx
├── 04_analysis/
│   ├── scripts/
│   │   ├── 01_clean.R
│   │   ├── 02_analyze.R
│   │   └── 03_visualize.R
│   └── notebooks/
├── 05_outputs/
│   ├── figures/
│   ├── tables/
│   └── reports/
├── 06_manuscript/
│   ├── drafts/
│   ├── submitted/
│   └── published/
├── 07_presentations/
└── README.md             ← Essential project documentation

The raw data principle

Never edit your raw data files

Raw data files are your primary source of truth. Once they are collected, they should never be modified. All cleaning, processing, and transformation should happen in scripts that read from the raw folder and write to the processed folder. This means:

• If you discover a cleaning error, you fix the script — not the raw data
• You can always re-run your entire analysis from scratch
• You can always verify what the original data looked like
• Reviewers and collaborators can check your work end-to-end

Write-protect raw data files (right-click → Properties → Read-only on Windows; chmod 444 filename on macOS/Linux) to prevent accidental editing.

README files

A README file is a plain-text document that explains a folder’s contents to anyone who encounters it — including your future self six months from now. Every project folder should have a README at the top level, and data folders should have their own README explaining the data.

What a project README should contain:

# Project Title

## Overview
What this project is about. Research question. Status (ongoing/completed).

## Compilers / Authors
Name, affiliation, contact email.

## Contents
Brief description of each subfolder and what it contains.

## Data Sources
Where the data came from. Collection dates. Any licensing restrictions.

## How to Reproduce the Analysis
What software is needed (including versions).
In what order to run the scripts.
Any setup steps required (installing packages, setting working directory, etc.).

## Ethics and Data Governance
Ethics approval reference number (if applicable).
Data access restrictions (who can use the data and how).

## Version History
Date — brief description of changes.

A README in the raw data folder should additionally document: exactly where the data came from, the date it was collected or downloaded, what each column means, what units are used, and any known issues or limitations.

Version control with Git

For code and scripts (R, Python, shell scripts), a version control system is the professional standard. Git records every change you make, lets you see exactly what changed and when, lets you revert to any previous state, and makes collaboration straightforward.

The basic workflow:

git init              # start tracking a project
git add .             # stage all changes
git commit -m "Clean data: remove duplicates and standardise date format"
git log               # see history of commits

Platforms like GitHub, GitLab, and Bitbucket host Git repositories online, providing a remote backup and enabling collaboration. Many universities provide institutional GitLab instances.

For researchers new to Git, the LADAL Reproducibility with R tutorial covers Git and GitHub in the context of R projects.

Version control for non-code files

Git works well for text-based files (scripts, markdown, CSV, plain text). For Word documents, spreadsheets, and other binary files, use the file naming convention described in Part 3 (date + version number in the filename) and keep all versions in a dedicated drafts/ subfolder.

The 3-2-1 backup rule

If your data is only in one place, it is not backed up

Hard drives fail. Laptops are stolen. Fires and floods happen. Ransomware encrypts everything it can reach. Any single storage location — including cloud sync — can fail. The industry-standard protection is the 3-2-1 rule.

3 copies of your data (original + two backups)

2 different types of media (for example, internal drive and cloud, or internal drive and external drive)

1 copy stored off-site (cloud storage counts, as does an external drive kept at a different physical location)

The reason for multiple media types is that a single failure mode (a power surge, a ransomware attack, a house fire) often affects multiple devices of the same type in the same location simultaneously.

Practical backup schedule:

• Daily (automatic): Active project files sync to cloud storage (OneDrive, Google Drive, or institutional storage)
• Weekly: All active projects backed up to an external hard drive
• Monthly: Full system backup; verify that previous backups can actually be restored

Backup tools:

• Windows: File History (continuous incremental backup); System Image Backup (full system snapshot)
• macOS: Time Machine (continuous backup to an attached drive); works automatically once configured
• Cross-platform cloud backup: Backblaze ($99/year) provides continuous off-site backup of your entire computer; useful for researchers with large data sets

Cloud sync is not the same as a backup

OneDrive, Google Drive, and Dropbox are sync services. If you accidentally delete a file, it is deleted from the sync service too (usually with a recovery window of 30 days or less). They are a useful component of a backup strategy but not a complete backup on their own.

A true backup involves a separate, independent copy that is not automatically synchronised with deletions.

Spreadsheet best practices

Many researchers store data in spreadsheets. Poorly structured spreadsheets create problems for analysis, sharing, and reproducibility. The core principle is to structure spreadsheets for machines, not for human readability in the moment.

The rules:

• One row per observation, one column per variable — no exceptions
• One header row only (no merged cells, no multi-row headers)
• No merged cells anywhere
• No colour coding to encode information (use an explicit column instead)
• Consistent date format: YYYY-MM-DD in every cell
• No blank rows or columns within the data
• Include units in column headers (Height_cm, Reaction_time_ms)
• Use NA (not blank cells) for missing data, consistently

Bad structure:

Name	Test Scores
	Math	English
John	85	90
	Science: 88
Sarah	92	87

Good structure:

StudentID	StudentName	Math_Score	English_Score	Science_Score	Date_Tested
S001	John Smith	85	90	88	2024-02-01
S002	Sarah Jones	92	87	91	2024-02-01

The good structure can be loaded directly into R, Python, or any statistical software without manual manipulation. The bad structure cannot.

Exercises: Research data management

Q4. A researcher keeps her only copy of a large interview corpus on her laptop’s internal drive, which she also syncs to Google Drive. Her laptop is stolen at a conference. She argues she is fine because she has the Google Drive copy. Is her backup strategy adequate, and why or why not?

Yes — she has a cloud backup, which is sufficient for any data loss scenario
Her backup strategy is not adequate because she only has two copies (laptop + cloud sync) and both are effectively the same storage type — one is just a mirror of the other. The 3-2-1 rule requires three copies, two different media types, and one off-site copy. She has one off-site copy (Google Drive), but no second independent backup. If the Google Drive sync had an error, or if the account were compromised, she would have nothing. The correct strategy adds a third copy: a weekly backup to an external hard drive kept at home or the office, separate from the laptop.
No — Google Drive is not a backup because it is not approved for research data
Yes — as long as the files are synced, the location of the physical laptop does not matter

---

Part 6: Productivity and AI Tools

Section Overview

What you will learn: Essential keyboard shortcuts; practical productivity habits; and how to use AI writing and coding assistants responsibly — understanding what they can and cannot do reliably

Keyboard shortcuts

Learning keyboard shortcuts is one of the highest-return productivity investments available. Research on expert users consistently finds they save 20–30 minutes per day compared to mouse-only workflows — that is 80–120 hours per year.

Universal (Windows: Ctrl; macOS: Cmd)

Shortcut	Action
Ctrl/Cmd + C	Copy
Ctrl/Cmd + V	Paste
Ctrl/Cmd + X	Cut
Ctrl/Cmd + Z	Undo
Ctrl/Cmd + Y	Redo
Ctrl/Cmd + S	Save
Ctrl/Cmd + F	Find
Ctrl/Cmd + A	Select all
Ctrl/Cmd + P	Print
Ctrl/Cmd + W	Close window/tab

Windows-specific

Shortcut	Action
Win + D	Show/hide desktop
Win + E	Open File Explorer
Win + L	Lock computer
Win + S	Open Search
Win + V	Clipboard history
Alt + Tab	Switch between open windows
Ctrl + Shift + Esc	Open Task Manager
Win + Shift + S	Screenshot (region select)

macOS-specific

Shortcut	Action
Cmd + Space	Spotlight search
Cmd + Tab	Switch between applications
Cmd + Shift + 4	Screenshot (region select)
Cmd + Option + Esc	Force Quit applications
Ctrl + Space	Switch input source

Text navigation (works in most applications)

Shortcut	Action
Home / End	Start / end of line
Ctrl/Cmd + Home	Start of document
Ctrl/Cmd + End	End of document
Ctrl/Cmd + →	Jump word by word right
Ctrl/Cmd + ←	Jump word by word left
Shift + any movement key	Select while moving

General productivity habits

Close what you are not using. Every open application and browser tab consumes memory and attention. Close tabs and applications you are not actively working with. Use a read-later tool (such as Pocket or a dedicated folder in your reference manager) for pages you want to return to.

One task at a time. Switching between tasks has a cognitive cost. Batch similar work (email, reading, writing, analysis) rather than interleaving it. Turn off notifications during focused work periods.

Save frequently. Many applications now save automatically, but not all. The habit of Ctrl+S after every significant change takes one second and prevents the loss of work to unexpected crashes. Enable autosave in Microsoft Office and Google Docs.

Use search efficiently. Both Windows (Win+S) and macOS (Spotlight, Cmd+Space) have powerful built-in search. Learning to use them effectively — including searching by file type, date range, and content — is often faster than navigating folder structures.

Working with AI tools

AI language models — including ChatGPT (OpenAI), Copilot (Microsoft), Claude (Anthropic), and Gemini (Google) — have become practical tools for researchers. They can accelerate many tasks: drafting, editing, summarising, explaining concepts, generating and debugging code, and brainstorming. Used well, they save significant time. Used carelessly, they introduce errors that can be difficult to detect.

What AI tools do well

• Writing assistance: Drafting emails, restructuring paragraphs, improving clarity and tone, suggesting alternative phrasings
• Code generation and debugging: Writing boilerplate code, explaining error messages, suggesting fixes, translating between programming languages
• Summarisation: Condensing long documents or explaining complex passages in plain language
• Brainstorming: Generating lists of ideas, suggesting alternative approaches, identifying gaps in an argument
• Explaining concepts: Providing accessible explanations of unfamiliar methods or terms

What AI tools do poorly

AI tools have well-documented limitations — know them before you rely on them

Hallucination: AI models generate plausible-sounding but factually incorrect information, including fabricated citations, invented statistics, and non-existent studies. They do this confidently and without warning. Any factual claim generated by an AI must be independently verified before use.

Knowledge cutoffs: Models are trained on data up to a certain date. They do not know about research published after their training cutoff, recent software updates, or current events.

No access to your data: Unless you explicitly provide context, the model does not know your specific situation, field conventions, or institutional requirements. Generic advice may be inappropriate for your context.

Inconsistency: The same prompt asked multiple times may produce different answers. Confidence in the response does not correlate with accuracy.

Privacy risks: Text you type into a cloud-based AI system may be used to train future models (depending on the service and your settings). Do not enter personally identifiable information, unpublished research data, sensitive correspondence, or confidential institutional information.

How to use AI tools responsibly in research

Verify everything. Treat AI output as a starting point, not a finished product. Check every factual claim. Every citation must be looked up and confirmed to exist and say what is claimed.

Disclose use appropriately. Most journals, conferences, and institutions now have explicit policies on AI use in research and writing. Read the relevant guidelines for your context before using AI in any work you will submit. Transparency about how AI was used is increasingly expected.

Keep sensitive data out. Do not paste interview transcripts, participant data, unpublished manuscripts, confidential grant materials, or student data into public AI interfaces. Many services offer privacy-preserving enterprise tiers, or you can use locally-run open-source models (see the LADAL Local LLMs with Ollama tutorial for running models on your own machine).

Use it for efficiency, not as a substitute for expertise. AI can help you write faster or understand something new faster — but it cannot replace your judgement about what claims are valid, what methods are appropriate, or what your results mean. The researcher remains responsible for all decisions and outputs.

Practical workflow for AI-assisted writing:

1. Draft your own outline or key points first — this keeps your ideas central
2. Use AI to expand, rephrase, or improve clarity on specific passages
3. Read every AI-generated sentence as if you wrote it — you are responsible for it
4. Verify any fact, figure, or reference that the AI provided
5. Disclose AI assistance in accordance with the relevant guidelines

Exercises: AI tools

Q5. A PhD student asks ChatGPT to list five key references on corpus linguistics methodology. The model provides five plausible-sounding citations with authors, titles, journals, volumes, and page numbers. The student copies them directly into her literature review. What is the risk, and what should she have done?

There is no risk — AI models have access to academic databases and provide accurate citations
The risk is that the citations might be from outside her field
The risk is that AI language models frequently generate fabricated citations — real-sounding but non-existent papers, or real papers with invented details (wrong year, wrong journal, wrong page numbers, wrong authors). This is called hallucination. Every citation in the literature review must be independently verified: searched for in a database (Google Scholar, Web of Science, the publisher's website) and confirmed to exist and to say what is claimed. If a reviewer or examiner checks references and finds that some do not exist, the consequences for the student's academic integrity can be serious.
The risk is a copyright violation for copying the citation text

---

Quick Reference

Maintenance checklist

Every week

• Restart the computer fully
• Check for pending updates and install critical ones
• Empty Recycle Bin / Trash
• Glance at Task Manager / Activity Monitor

Every month

• Run antivirus scan
• Run disk cleanup
• Review startup programs
• Clean keyboard and screen
• Check available storage and backup status

Every six months

• Physical cleaning with compressed air
• Full file organisation review
• Update passwords for high-value accounts
• Review security settings and MFA status
• Full system backup and verify restore works

Security levels

Progressive security — start at the top

Level 1 — Foundation (do these first)

1. Enable full-disk encryption (BitLocker or FileVault)
2. Use unique, strong passwords for every account
3. Enable automatic security updates
4. Use active antivirus protection

Level 2 — Good practice

5. Install a password manager
6. Enable multi-factor authentication on email and institutional accounts
7. Enable MFA on all other accounts that support it
8. Implement the 3-2-1 backup rule

Level 3 — Strong security

9. Use a VPN on public WiFi
10. Use a hardware security key for high-value accounts
11. Encrypt sensitive files and folders
12. Review third-party app permissions regularly

Level 4 — Research-specific

13. Use institutional approved storage for sensitive research data
14. Write README files for all research projects
15. Never edit raw data files
16. Use version control (Git) for analysis scripts

---

Troubleshooting

Computer running very slowly

1. Check Task Manager or Activity Monitor — is CPU or memory near 100%?
2. Close unused applications and browser tabs
3. Restart the computer
4. Run an antivirus scan
5. Run disk cleanup
6. Disable unnecessary startup programs

Computer overheating

1. Check that vents are not blocked — laptops need space beneath them
2. Clean vents with compressed air
3. Reduce the number of running programs
4. Check fan operation (unusual noise may indicate failure)
5. Consider a laptop stand or cooling pad
6. If problems persist, consult a technician

Cannot find a file

1. Use built-in search: Win+S on Windows, Cmd+Space on macOS, or find / locate in a Linux terminal
2. Check the Recycle Bin / Trash
3. Check cloud storage — the file may not have synced to the current device
4. Sort the folder by date modified to surface recently changed files
5. Check whether it was saved with a different name

Forgot a password

1. Use the “Forgot password” / reset link on the login page
2. Check your password manager — it may have saved the password
3. Contact the service’s support team
4. Lesson learned: set up a password manager before this happens again

---

Citation & Session Info

Schweinberger, Martin. 2026. Working with Computers: Essential Skills for Researchers. Brisbane: The Language Technology and Data Analysis Laboratory (LADAL). url: https://ladal.edu.au/tutorials/comp/comp.html (Version 2026.05.01).

@manual{schweinberger2026comp,
  author       = {Schweinberger, Martin},
  title        = {Working with Computers: Essential Skills for Researchers},
  note         = {https://ladal.edu.au/tutorials/comp/comp.html},
  year         = {2026},
  organization = {The Language Technology and Data Analysis Laboratory (LADAL)},
  address      = {Brisbane},
  edition      = {2026.05.01}
}

AI Transparency Statement

This tutorial was written with the assistance of Claude (claude.ai), a large language model created by Anthropic. Claude was used to substantially revise and expand an earlier version of this tutorial, remove institution-specific references, add learning objectives, check questions, and a new section on AI tools. All content was reviewed and approved by Martin Schweinberger, who takes full responsibility for its accuracy.

sessionInfo()

R version 4.4.2 (2024-10-31 ucrt)
Platform: x86_64-w64-mingw32/x64
Running under: Windows 11 x64 (build 26200)

Matrix products: default


locale:
[1] LC_COLLATE=English_United States.utf8 
[2] LC_CTYPE=English_United States.utf8   
[3] LC_MONETARY=English_United States.utf8
[4] LC_NUMERIC=C                          
[5] LC_TIME=English_United States.utf8    

time zone: Australia/Brisbane
tzcode source: internal

attached base packages:
[1] stats     graphics  grDevices datasets  utils     methods   base     

other attached packages:
[1] checkdown_0.0.13

loaded via a namespace (and not attached):
 [1] digest_0.6.39       codetools_0.2-20    fastmap_1.2.0      
 [4] xfun_0.56           glue_1.8.0          knitr_1.51         
 [7] htmltools_0.5.9     rmarkdown_2.30      cli_3.6.5          
[10] litedown_0.9        renv_1.1.7          compiler_4.4.2     
[13] rstudioapi_0.17.1   tools_4.4.2         commonmark_2.0.0   
[16] evaluate_1.0.5      yaml_2.3.10         BiocManager_1.30.27
[19] rlang_1.1.7         jsonlite_2.0.0      htmlwidgets_1.6.4  
[22] markdown_2.0       

---

Back to top

Back to LADAL home

---

Additional Resources

General computer help

• Microsoft Support — Windows and Office documentation
• Apple Support — macOS documentation
• Ubuntu Documentation — Linux help

Security

• Have I Been Pwned — check whether your email appears in known data breaches
• Bitwarden — recommended open-source password manager
• Two Factor Auth — directory of services that support MFA
• Privacy Tools — privacy-focused software recommendations
• Security Planner — personalised security advice

File organisation and data management

• The Turing Way — community handbook for reproducible data science
• Data Carpentry — data management lessons for researchers
• UK Data Service — Research Data Management — comprehensive guidance

AI tools

• Claude — Anthropic’s AI assistant
• ChatGPT — OpenAI’s AI assistant
• LADAL Local LLMs tutorial — running AI models locally for privacy-sensitive work