Skip to content

PITCHME.md

Latest commit

 

History

History
196 lines (169 loc) · 5.79 KB

PITCHME.md

File metadata and controls

196 lines (169 loc) · 5.79 KB

Reproducible Research

@AmandaMiottoGU and @JulieToohey

Griffith University

What is Reproducible Research?

image Note: https://thenib.com

Why is it Important?

image Note: https://www.nature.com/news/polopoly_fs/7.36718.1464174471!/image/reproducibility-graphic-online3.jpg_gen/derivatives/landscape_630/reproducibility-graphic-online3.jpg

Extracts from Retraction Watch (Australian papers)

A paleontology journal has retracted a recent paper after discovering it had published the uncorrected version of the manuscript.

The mistake occurred after the

authors submitted revisions to the manuscript without tracking the changes,

prompting the publisher to believe nothing had been changed and publishing the previous version. The journal initially told the authors it planned to publish an erratum that described the mistake as a production error, but then retracted the paper—seemingly without consulting the authors. However, the authors said they were happy with the outcome.

A grad student was caught in the crossfire of fraud — and fought back

In March, 2013, a graduate student joined the lab of a prominent researcher in Australia...

Reflections by authors who had discovered they had made a mistake: https://dynamicecology.wordpress.com/2016/12/02/on-finding-errors-in-ones-published-analyses/ http://ecoevoevoeco.blogspot.com/2016/12/wrong-lot.html

Today we are going to learn 9 lessons to safeguard us against these situations

Let's look at some scenerios....

What if a key person from your lab disappeared one day (family or personal emergency, no longer contactable)?

Could you all continue your work?

Would you know where all your data is stored?

Could you keep running effectively for 1 month?

1 year?

image Note: Image from the Upturned Microscope

Let's look at a few tools we can use to protect our research:

Documentation

  • Naming Conventions
  • Folder Structures
  • Automation
  • Version Control

---?include=Step1Documentation.md

Let's look at a few tools we can use to protect our research:

  • Documentation

Naming Conventions

  • Folder Structures
  • Automation
  • Version Control ---?include=Step2Naming.md

Let's look at a few tools we can use to protect our research:

  • Documentation
  • Naming Conventions

Folder Structures

  • Automation
  • Version Control

---?include=Step3FolderStruct.md

Let's look at a few tools we can use to protect our research:

  • Documentation
  • Naming Conventions
  • Folder Structures

Automation

  • Version Control

---?include=Step4Automation.md

Let's look at a few tools we can use to protect our research:

  • Documentation
  • Naming Conventions
  • Folder Structures
  • Automation

Version Control

---?include=Step5Version.md

Great!

By implementing these steps, we've already improved our processes!

Now if someone leaves our lab, we're covered!

Next Scenerio....

Imagine you're travelling and lose your laptop bag with your external hard drives?

Or your office is robbed? Could you continue your work? Is your data backed up? Encrypted?

Let's look at a few tools we can use to protect our research:

  • Backing up to the cloud
  • Computer Security
  • Deidentifying your data

image Note: PHDcomics

---?include=Step6CloudBackup.md

Let's look at a few tools we can use to protect our research:

  • Backing up to the cloud

Computer Security

  • Deidentifying your data Note: While Computer security isn't really about reproducibility, it is important. Plus I wanted to highlight encrypting hard drives ---?include=Step7CompSecurity.md

Let's look at a few tools we can use to protect our research:

  • Backing up to the cloud
  • Computer Security

Deidentifying your data

---?include=Step8SepId.md

So now we have some new skills up our sleeve

If our laptop goes missing, we are better prepared

Now let's look at after a project has finished

Next Scenario....

Someone has published contradicting results to your published paper, and you've been asked to provide your data and methods. Could you?

image Note: PHDcomics

We're already covered some things that can help us

  • Storing our data in the cloud
  • Documentation
  • Versions of software

New tool

Digital Object Identifier (DOI) and Persistent identifier (PiD)

---?include=Step9Identifiers.md

Wow, that's 9 Reproducible Research things!

What if we look at another Scenario?

if a research partner organization believes your "sensitive" data has been made available to others (ie a data breach). Could you show that steps were taken to avoid this or show that it couldn't happen?

Turns out, we can!

  • Owncloud or Research Space – tracing who you have shared a file (Step 6 Cloud Backups)
  • Computer encryption (Step 7 Good Computer Safety)
  • General good computer safety – unique passwords and use Multi factor Auth when possible (Step 7 Good Computer Safety)
  • Separating identified variables (Step 8- Deidentifying your data)

Amazing! We have so many new skills!

What changes are you going to make?

image Note: the nib.com Repeat after me comic

Thank you for coming!