1 of 39

The future is now - research experiments

via online platforms - a brief update

PsyBSc 10: EXPRA - Group 3

Winter term 2021

Peer Herholz (he/him)

Habilitation candidate - Neurocognitive Psychology at Goethe-University Frankfurt

Research affiliate - NeuroDataScience-ORIGAMI lab at MNI, McGill & BRAMS

Member - BIDS, ReproNim, Brainhack, UNIQUE, CNeuroMod

@peerherholz

13/01/2022

2 of 39

A few things before we start...

https://giphy.com/embed/4Z9jjKGJVM9Ck8E4cN

I hope y’all had a nice and relaxing break!

It’s great to see y’all again and I’m looking forward

to the last 4 weeks of this course!

3 of 39

A few things before we start...

https://giphy.com/embed/JRmenOACUaRFCCgSqp

As it’s been a while, we covered a lot of content already and transition

to a new topic, we will spend the first few minutes on bringing things back

to our memory as well as re-situate certain aspects.

4 of 39

EXPRA - 101 - the life cycle of a scientific project from ideas to results

Please make sure to check https://undraw.co/

analyses

results

data acquisition

project creation

not sufficient

not sufficient

not sufficient

not sufficient

sufficient if done right

communication

not sufficient

A holistic perspective: everything is important

5 of 39

EXPRA - 101 - the life cycle of a scientific project from ideas to results

Every steps needs to have/generate FAIR outcomes

Findable

  • data and metadata easily findable for both humans and machines

(machine readable metadata essential for automatic discovery)

Accessible

  • data can be easily obtained by humans as well as by machines,

through well-defined and ideally standardized protocols

  • conditions must be clearly established (license, reuse rights, etc.).
  • when the data is inaccessible, the metadata should remain so in a way that the dataset remains discoverable

Interoperable

  • possible to integrate different data into a single interface
  • use of a common programming language, the use of controlled vocabulary, as well as non-proprietary formats
  • can also be applied to metadata as well as references to

other (meta)data

Re-Usable

  • data and metadata following best practices -> optimized for reuse
  • data and metadata must be richly detailed, as well as user licenses
  • any information specific to the field of interest of the data in question must be made available as well.

https://ogsl.ca/wp-content/uploads/Fair-rectangle-en.png

6 of 39

EXPRA - 101 - the life cycle of a scientific project from ideas to results

Every steps needs to follow open-science principles

assessment

outreach

publication

writing

analysis

  • adding alternative evaluation, e.g. with altmetrics
  • communicating through social media, e.g. Twitter
  • sharing posters & presentations, e.g. at FigShare
  • using open licenses, e.g. CC0 or CC-BY
  • publishing open access, ‘green’ or ‘gold’
  • using open peer review, e.g. at journals or PubPeer
  • sharing preprints, e.g. at OSF, arXiv or bioRxiv
  • using actionable formats, e.g. with Jupyter or CoCalc
  • open XML-drafting, e.g. at Overleaf or Authorea
  • sharing protocols & workfl., e.g. at Protocols.io
  • sharing notebooks, e.g. at OpenNotebookScience
  • sharing code, e.g. at GitHub with GNU/MIT license
  • sharing data, e.g. at Dryad, Zenodo or Dataverse
  • pre-registering, e.g. at OSF or AsPredicted
  • commenting openly, e.g. with Hypothes.is
  • using shared reference libraries, e.g. with Zotero
  • sharing (grant) proposals, e.g. at RIO

search

Bianca Kramer & Jeroen Bosman https://101innovations.wordpress.com

7 of 39

EXPRA - 101 - the life cycle of a scientific project from ideas to results

life-cycle of a scientific project

modules

content �&

aspects

FAIR

outcome

execution

finalization

introduction & prerequisites

data acquisition & processing

results communication

scientific method

theory building

literature/data surveys

validity & reliability

hypothesis generation

power analysis

preregistration

preregistration

DMP

bibliography

OSF project

(open lab notebook)

data structures

data management

(version control)

understanding data

preprocessing

inferential measures

results diagnostic

preregistration (updated)

DMP (updated)

reproducible analyses

walkthroughs

(open lab notebook)

presenting results

design posters/talks

write articles

publication models

peer review

beyond publication

outreach

poster

talk

report

reviews

(open lab notebook)

idea & conceptualization

We will divide the life-cycle of our scientific projects into 3 modules/blocks and incorporate the respective FAIR/open-science resources/tools .

8 of 39

EXPRA - 101 - the implementation

A few more details

You will be conducting two scientific research experiments:

demo experiment

guided under close supervision

work in groups

same experiment for everyone

lots of feedback from instructor & classmates

your own experiment

(more) independent work

work in groups

hypotheses vary between groups

less direct feedback

Along the way you’ll: - receive input and background information from the instructor for all steps

of a given scientific research experiment

- implement the respective sub-tasks & work together with your classmates on

certain problems & provide feedback

9 of 39

EXPRA - 101 - how to get those credits

A few more details

Successful participation and your grade will be based on different aspects:

demo experiment

guided under close supervision

work in groups

same experiment for everyone

lots of feedback from instructor & classmates

your own experiment

(more) independent work

work in groups

hypotheses vary between groups

less direct feedback

report, one per group,

~15 pages, no grading,

deadline: 15/12/2021 12 PM CET

report, individually,

~15 pages, 100% of your grade,

deadline: ~-09/03/2022 12 PM CET

course work

participation/readings

reviews/feedback

journal club presentation

poster presentation

in groups/individually, no grading,

deadline: throughout semester

10 of 39

Objectives for this session

https://media.makeameme.org/created/look-at-the-25f4b725ed.jpg

  • learn how to monitor online experiment and get respective data
  • journal club presentations
  • Ask and answer questions
  • Have a great time

11 of 39

Outline for this session

  • Online experiments via meadows.com
    • monitoring experiments
    • getting data
  • Journal Club presentations - I
    • Study outlines from each group
    • general feedback/discussion
  • Outro/Q&A

https://twitter.com/OfficeMemes_/status/1298982572848869380/photo/1

12 of 39

Outline for this session

  • Online experiments via meadows.com
    • monitoring experiments
    • getting data
  • Journal Club presentations - I
    • Study outlines from each group
    • general feedback/discussion
  • Outro/Q&A

https://twitter.com/OfficeMemes_/status/1298982572848869380/photo/1

13 of 39

Online experiments via meadows.com - monitoring experiments

documentation

Please also consult these pages in addition to this session.

researcher login

Will be provided!

14 of 39

Online experiments via meadows.com - monitoring experiments

Main/landing/overview page

The page you’ll see after the log in.

your experiments

Experiments you created in the past/that are currently running.

Account information

Info related to a researcher’s account.

Researcher account

Everyone will be using the same account for billing reasons.

15 of 39

Online experiments via meadows.com - monitoring experiments

Experiment overview page

The page you’ll see after the clicking on your experiment, e.g. EXPRA_2021_MusicListening.

name & description

experiment settings

structure & tasks of experiment

duration & settings for each task

experiment duration est. & preview

16 of 39

Online experiments via meadows.com - monitoring experiments

How to monitor your experiment

You can monitor your experiment via two options: Activity or Versions.

evaluate/check different versions of your experiment and respective participants’ data in detail

get a general overview how much “traffic” (online minutes spents) wrt your experiment

17 of 39

Online experiments via meadows.com - monitoring experiments

How to monitor your experiment

You can monitor your experiment via two options: Activity or Versions.

  • indicating minutes per day
  • helpful to get an idea of general data acquisition trends
  • not very informative regarding details

18 of 39

Online experiments via meadows.com - monitoring experiments

How to monitor your experiment

You can monitor your experiment via two options: Activity or Versions.

all versions of your experiment

total number of versions differs between experiments based on problems with SONA

19 of 39

Online experiments via meadows.com - monitoring experiments

How to monitor your experiment

You can monitor your experiment via two options: Activity or Versions.

tasks and structure of specific version of experiment

participants who started the experiment

20 of 39

Outline for this session

  • Online experiments via meadows.com
    • monitoring experiments
    • getting data
  • Journal Club presentations - I
    • Study outlines from each group
    • general feedback/discussion
  • Outro/Q&A

https://twitter.com/OfficeMemes_/status/1298982572848869380/photo/1

21 of 39

Online experiments via meadows.com - getting data

How to get data from your experiment

You can get from your experiment via two options: browse all -> for each participant or download all to get everyone who started.

current “batch” of participants, i.e. round of data acquisition

check each participant in detail

download all participant’s data

currently running until 31/01/2022 or 40 participants, we’ll discuss strategy next week

22 of 39

Online experiments via meadows.com - getting data

all participants who started

the experiment

random participant ID

participant status

download all data

started experiment but didn’t finish/timed out yet

started experiment but didn’t finish within 24 hours

How to get data from your experiment

You can get from your experiment via two options: browse all -> for each participant or download all to get everyone who started.

23 of 39

Online experiments via meadows.com - getting data

How to get data from your experiment

You can get from your experiment via two options: browse all -> for each participant or download all to get everyone who started.

Y’all should download this version of the data as it will be identical/comparable to the data we analyzed within the demo experiments and thus feasible for the R analyzes script.

24 of 39

Online experiments via meadows.com - getting data

How to get data from your experiment

You can get from your experiment via two options: browse all -> for each participant or download all to get everyone who started.

information on all tasks, including duration

25 of 39

Online experiments via meadows.com - getting data

How to get data from your experiment

You can get from your experiment via two options: browse all -> for each participant or download all to get everyone who started.

Spoiler: most participants timed out and stopped at the sound test!

26 of 39

Online experiments via meadows.com - getting data

How to get data from your experiment

A note on participants in general and specifically online settings.

A Comparison of Crowd Types: Idea Selection Performance of Students and Amazon Mechanical Turks (Banken, 2021)

27 of 39

Online experiments via meadows.com - getting data

How to get data from your experiment

A note on participants in general and specifically online settings.

https://giphy.com/embed/l0MYC0LajbaPoEADu

… to “real-life” research! Seriously, that’s a classic and very expectable. Quite often participants don’t pay attention, just stop, etc. . That’s not you, that’s just how it is and there’s only so much we can do about it at this point.

Any ideas how to get more participants to finish the experiment?

28 of 39

https://giphy.com/gifs/Friends-episode-12-friends-tv-the-one-with-embryos-WOIGpnJ3ye445BUQl4

https://giphy.com/gifs/season-7-the-simpsons-7x20-3o6MbnqpZBwWdWg3gk

Suggested task for y’all: check your experiment once per day and download data as soon as possible. The same holds true for exploring the data following the outline of the demo experiment and respective R script, as well as your data management plans and pre-registration.

(Regarding all these points: try to remember the different project management options we discussed.)

Research experiment groups - monitoring experiments & getting data

29 of 39

Outline for this session

  • Online experiments via meadows.com
    • monitoring experiments
    • getting data
  • Journal Club presentations
    • Study outlines from each group
    • general feedback/discussion
  • Outro/Q&A

https://twitter.com/OfficeMemes_/status/1298982572848869380/photo/1

30 of 39

https://giphy.com/gifs/Friends-episode-12-friends-tv-the-one-with-embryos-WOIGpnJ3ye445BUQl4

https://giphy.com/gifs/season-7-the-simpsons-7x20-3o6MbnqpZBwWdWg3gk

Task for y’all: present your study/experiment in a short presentation (up to 20 min.), outlining background & hypothesis (and methods). Within that, each group member should briefly summarize one study that was utilized to form the research questions and derive hypotheses, i.e. prior literature on the topic.

Research experiment groups - journal club I

31 of 39

https://giphy.com/gifs/Friends-episode-12-friends-tv-the-one-with-embryos-WOIGpnJ3ye445BUQl4

https://giphy.com/gifs/season-7-the-simpsons-7x20-3o6MbnqpZBwWdWg3gk

Task for y’all: before we start, each group should set up a shared doc with the following parts/headings: general impression, experiment outline/summary, hypotheses derivation & methods and send the link via Discord.

Folks who are not presenting are supposed to provide feedback via this doc.

Research experiment groups - journal club I

32 of 39

Outline for this session

  • Online experiments via meadows.com
    • monitoring experiments
    • getting data
  • Journal Club presentations
    • Study outlines from each group
    • general feedback/discussion
  • Outro/Q&A

https://twitter.com/OfficeMemes_/status/1298982572848869380/photo/1

33 of 39

https://giphy.com/gifs/Friends-episode-12-friends-tv-the-one-with-embryos-WOIGpnJ3ye445BUQl4

https://giphy.com/gifs/season-7-the-simpsons-7x20-3o6MbnqpZBwWdWg3gk

Task for y’all: provide feedback and impressions gathered so far regarding running your own experiment, initiate discussions, etc., basically consider this section an open mic/stage!

Research experiment groups - general feedback/discussion

34 of 39

Outline for this session

  • Online experiments via meadows.com
    • monitoring experiments
    • getting data
  • Journal Club presentations - I
    • Study outlines from each group
    • general feedback/discussion
  • Outro/Q&A

https://twitter.com/OfficeMemes_/status/1298982572848869380/photo/1

35 of 39

Outro/Q&A - Readings/add-on material for this session

36 of 39

Outro/Q&A - Recap for this session

  • online experiments via meadows
    • monitoring: versions, general activity vs. in-depth
      • different versions of each experiment due to SONA problems
      • general activity for overview, less precise
      • in-depth for detailed overview per participant
    • getting data: all data vs. per participant
      • all data will download data from all participants who started the experiment
      • per participant allows download of data only from participants who finished experiment
      • research experiment groups should use the .json version of the data

37 of 39

Outro/Q&A - Were your expectations for this session fulfilled?

38 of 39

Outro/Q&A - Questions you could/should ask based on this session

How to weigh pros and cons of online experiments?

How to validate online experiments and responses?

How to situate the interaction between online experiments and different research fields?

If possible, would you consider running only online experiments?

39 of 39

https://giphy.com/gifs/season-17-the-simpsons-17x6-xT5LMB2WiOdjpB7K4o