Matt Tenan is Mostly on LinkedIn These Days
@TenanATC
Followers
3,775
Following
672
Media
854
Statuses
11,581
Physiologist & Data Scientist. Fan of good trouble.
North Carolina, USA
Joined July 2013
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
無料クーポン
• 129868 Tweets
Jets
• 119052 Tweets
Yankees
• 84735 Tweets
連休明け
• 83603 Tweets
リンク1
• 59273 Tweets
緊急入院
• 39461 Tweets
Rogan
• 39355 Tweets
Aaron Rodgers
• 35351 Tweets
Hail Mary
• 27940 Tweets
अब्दुल कलाम
• 27780 Tweets
当選確率2倍
• 26689 Tweets
Josh Allen
• 25603 Tweets
#BUFvsNYJ
• 22349 Tweets
पूर्व राष्ट्रपति
• 19306 Tweets
#ファミマのティにゃミスメロンパンハウス
• 14093 Tweets
#APJAbdulKalam
• 12914 Tweets
ハイタッチ
• 12891 Tweets
Stanton
• 11852 Tweets
うさぎくじ
• 10806 Tweets
महान वैज्ञानिक
• 10784 Tweets
Pinned Tweet
I deleted Twitter of my phone over the holidays. I can see I did not miss much.
I've been using
@LinkedIn
more for social media. You can follow me there or, if I know you or you send me a custom invite message, I'm happy to accept requests:
0
0
4
Dear Sport Scientists:
Make friends with Statisticians.
Make friends with Engineers.
Make friends with Computer Scientists.
Make friends with Dietitians.
Make friends with Psychologists.
Recognize others' expertise and utilize it . You don't need to do it all.
30
235
1K
I have a dirty secret to tell you
My PhD is in Neuroscience but I don't understand most of the nervous system
Your statistician friend doesn't understand all of statistics
Your physician doesn't understand all of medicine
Being an Expert means knowing a small area really well
18
278
1K
In case you're wondering, computer scientists are now correcting all the statisticians out there: Logistic Regression is NOT regression, it's classification.
No hubris at all 🙄
23
66
561
I apparently have a bit of a reputation as someone who is anti-machine learning or anti-AI when it comes to human research. This is a bit of misrepresentation of my views, and (I'd argue) a misrepresentation of the issues "statistics people" take with AI/ML as a whole.
(THREAD)
4
74
215
No one device, not a force plate, an accelerometer, a markerless movement screen... NOTHING.
Is validated to predict and mitigate injury.
I think I just saved your organization $20,000 a year.
You're welcome.
11
23
171
If you use Dashboards in Sport, I WROTE THIS PAPER FOR YOU in
@SportsMedicineJ
! Doesn't matter if you use
@MovellaHS
,
@Smartabase
@Teamworks
,
@KitmanLabs
or
@catapultsports
, if you handle missing data by 'doing nothing', your interpretations are wrong.
2
35
162
Reminder: If you ever need a scientific manuscript, just contact the author. Even if the lead author is a Chair/Dean/Chancellor, I have a 90% response rate within 24 hours.
Everyone likes to know their research is going to be read/used.
1
29
161
Thrilled to announce I've joined
@WestVirginiaU
@WVURNI
as Program Director for Data Science & Human Performance!
Along w/research in the PM&R Department, I'll be developing the analytical ecosystem for
@WVUSports
. Look out
@Big12Conference
, we're coming
@WVUhoops
@WVUfootball
.
14
2
145
Dear researchers performing statistical analysis in SPSS or Matlab:
Your p-value is never = .000
</RANT>
9
7
143
If you have the opportunity to decide HOW you are going to eat (low carb, paleo, vegan, vegetarian, high carb, organic, etc), you are a person of privilege. Many people have to worry IF they get to eat their next meal.
Just remember that this holiday season.
0
29
140
Exercise & Sport Sciences has a poor track record developing "Novel Statistical Methods" & a new one was recently published:
As a reviewer on this paper now working to have it retracted, here is my perspective on how these methods get popularized
[THREAD]
7
57
143
I've noticed a rash of people with +100 publications suggest that public post-publication review of scientific manuscripts is:
1. Disrespectful
2. Mean-spirited
3. Invalid
...etc.
Publicly questioning science shows clinicians we are serious about solving their clinical problems.
3
28
130
Are confidence intervals "better than p-values"?
No, and the premised statement doesn't make any sense. Thanks for heading this up
@ExPhysStudent
and publishing it
@JPhysiol
8
21
126
If you're going to develop a predictive algorithm read this paper by
@Richard_D_Riley
@f2harrell
@GSCollins
et al.
And, hey, if you are going to use a predictive algorithm, maybe check with developers to make sure they know about this paper as well!
10
37
125
Methods: The schools were ordered by standardized scores and the separated into tertiles. Independent samples t-test were used to test for group differences.
Results: Significant differences were found between groups (p <0.001).
6
10
123
I'm thinking of starting a list of standard statistical procedures that are mathematically equivalent to machine learning or that are actually statistics but now called machine learning
That way I can use sexy terms like Artificial Intelligence and get my work funded. Thoughts?
12
15
112
This was the decade of tech rediscovering things that already existed:
1. Computer scientists rediscover statistics (largely) and call it machine learning
2. WeWork rediscovers office leasing
3. Slack rediscovers AOL instant messenger
4. TikTok rediscovers Twitter's Vine
2
29
113
Have you ever had a Reviewer say something like:
"The authors seem to have run 200 unadjusted t-tests, so I'm not entirely surprised they found 6 significant results..."
That was me. I was that Reviewer.
9
7
109
Happy to see our report on conceptual issues with Acute:Chronic Workload Ratio finally published in
@IJSPPjournal
. We briefly discuss the issues of assumed causality and unjustfied ratio metrics. Thanks
@francoimpell
for the invitation to contribute.
3
36
108
<Rant>
Dear IRBs & Reviewers,
Exploratory research, by definition, will not have a strong a priori hypothesis. That does not make exploratory research invalid or useless. Stop requiring arbitrary "hypotheses" that don't make sense.
Thanks,
Matt
</Rant>
1
3
107
As a former researcher in the area, I find it VERY troubling to see a speaker at
#SSAC23
saying that menstrual cycle affects injury risk, recovery, and advocating for 'training around the cycle'.
There is ZERO evidence for this.
Don't do it.
Thanks for coming to my TED Talk
5
17
104
Early Career Researchers are undervalued
More should be funded by national agencies
More should be on Editorial Boards
More should be on government committees
We shouldn't talk about how they haven't paid dues.
We should talk about how they can change the future of the field
2
18
106
If you're getting a PhD, I'll tell you the opposite of what your advisor says:
You should be spending more time on Twitter
Learn from and interact with experts in your field and abroad. It's both better and worse than any professional conference, use Twitter wisely.
5
12
103
I feel like this is something that self-aware "experts" know and aren't bothered by. However, I often think that people following or otherwise looking for guidance in an area don't quite grasp how large, difficult, and all-encompassing many disciplines are.
2
0
101
Translation: the only people citing MDPI papers are the ones writing MDPI papers.
Seems about right.
2
29
98
Folks. MDPI Journals are predatory. They publish garbage research at more than $1,000 a pop for authors.
And yet, you keep tweeting MDPI papers. They're bad. Really bad.
I'm done seeing it.
11
16
97
Me: Do you want the model/algorithm for Inference, Prediction or Causal Inference?
Client: Yes, all of those things
Me: As we've previously discussed...
6
8
95
Things I can say because I'm not in academia:
A Research 1 institution in an urban environment paying their Assistant Professors $50k/year is shameful.
3
2
93
@EpiEllie
Last night over dinner:
Son: Did you know when Dad was a boy, people got the chicken pox!?
Daughter: What's a chicken poc? (sic)
These kids have no idea what life was like before vaccines (I know my parents probably said that about Hep-B for our generation)
41
3
93
Here's your semi-annual reminder that needing to retract a paper or write an erratum due to Honest Error is not a moral failure or demonstrate you aren't a good researcher.
In fact, the opposite is true. Better to be right eventually than wrong for posterity.
2
11
87
Accuracy, predictability and reliability data or I call BS.
4
11
83
Two MAJOR problems using machine learning techniques in a study with less than 100 subjects and a ton of predictors:
1. Model overfitting
2. Study generalizability
Why do journals keep glazing over these huge issues?
1
23
84
"Much to some practitioners’ disappointment, GPS data does not hold all the answers to sport’s mysteries."
@JasonAWeber
0
3
28
1
21
79
This brings me to a REALLY important review paper just published. It ties all HPG & sport dysfunctions together in a cohesive picture.
RED-S
Athlete Triad
Overtraining Syndrome
Sex Differences
Hypogonadism
It's all there and more. Check it out.
5/FIN
0
8
76
Current Peer Review Process:
*submit paper using Bayesian analysis*
Reviewer 2 R1: Do a traditional statistical analysis
Me: No, and Bayes is actually older than NHST
Reviewer 2 R2: Authors have still not addressed my concerns
Editor: I see both sides, ignore reviewer 2
🎉🎉
5
0
76
Un-trendy Opinion:
Neither Machine Learning, Directed Acyclic Graphs, Bayesian Statistics, nor the elimination of p-values will solve the root problems within science.
WHY?
All of the above can be mis-used inadvertently or abused purposefully, just like less-trendy methods.
8
18
74
I heretofore propose that all p-values between 0.1 and 0.05 shall be described as:
"trending away from significance"
8
5
74
Dear Sport & Exercise Science Researchers,
If you publish a "novel statistical method" paper in an exercise science journal and don't cite a single non-exercise methods paper, your method is either wrong or not novel (maybe both!)
Google Scholar indexes beyond our field, use it
2
5
72
STOP USING BAR CHARTS. THEY ARE MISLEADING.
STOP.
New: Boutcher and colleagues examine the effect of an 8-wk sprint interval training program on body composition and aerobic fitness of overweight postmenopausal women.
0
8
26
4
12
73
Dear Physical Medicine & Rehabilitation Researchers,
I just discovered you have a thing called a "Clinical Prediction Rule"... which everyone else in science just calls a multivariable model.
Please stop inventing new terms. Also, collaborate with statisticians.
Thanks!
3
13
73
Tell me your grant doesn't have a Statistician without telling me your grant doesn't have a Statistician:
"We are going to use a linear mixed model, but prior to analysis, we are going to check to make sure the assumption of normality is not violated"
*SIGH*
4
4
70
Did I allow us to submit a paper where we continually say "Casual Analysis" and not "CAUSAL Analysis"?
Yes, I did.
No, a reviewer or editor did not catch it.
11
1
70
Once a week I send a co-authoring clinician the Datasaurus Dozen as a reason why I won't publish a correlation paper with only coefficients and now visualization.
This week is no different.
2
14
68
I was asked today, by a clinician, to run X analyses "so we can start formulating our research questions"
6
3
68
I just had a reviewer describe my statistical analysis as 'bland' without any suggestion for an 'exciting' procedure.
If the statistics section in empirical work is exciting/progressive, it's because the conventional analysis didn't "work out".
Here's your daily reminder that in empirical science, you don't need "progressive statistics" you just need "appropriate statistics".
I know it's boring, but sometimes doing things well is boring. That's okay.
5
2
24
8
8
68
I have paragraphs in my papers (and in my doctoral dissertation!) that look like this...
After addressing Reviewer 3's comments the manuscript now looks like...
2
16
226
3
3
67
Can we just stop using MCIDs? They're useless and non-reproducible.
Just... stop.
4
14
67
Don't buy any predictive analytics product where they can't provide validity metrics.
If you don't feel comfortable asking those questions, hire an independent consultant, pay them $1-3k, and get their opinion.
This will save you a ton of money in the long run.
4
15
67
I'm excited to announce that I have accepted a position supporting the
@MOTIONetwork
as a Senior Biostatistician. Pumped to do statistics "full time" and increase the quality of healthcare for the Warfighter at the same time. Onward!
12
0
67
If any PhD's out there have a background in Exercise Physiology and are handy with statistics, let me know if you're interested in working with U.S. Special Forces. DM me and I'll set you up with the recruiter:
4
29
66
I graduated from undergrad 15+ years ago and if I email a former professor
@IthacaCollege
, I always get a response within 48 hours.
Rankings are important but that non-quantified information is the really impressive stuff that should drive college selection.
0
2
67
You're never too "senior" a scientist to not spend 6 hours figuring out how why something won't plot correctly.
2
1
63
I don't share a ton of personal stuff on Twitter, but last night my freshman year roommate at
@IthacaCollege
won an Oscar.
A Freaking Oscar.
Excuse me while I re-evaluate my life.
1
0
66
It's really tough to talk about sports and analytics when one country is sending hundreds of thousands of Warfighters into another country completely unprovoked and without reason.
1
0
63
It took me nearly exactly 6 years but by golly this thing looks good framed.
5
0
63
Hot off the press! Measurement error in wearables is poorly understood & it does not help when companies say they measure one "thing" but really measure "another".
We quantified if measuring HRV with a wearable is the same as with an ECG
3
15
65
Issue 1: CLAIMING EVERYTHING IS MACHINE LEARNING.
Just because AI/ML may use algebra or linear regression, doesn't mean it is AI/ML. Same goes for Nonlinear regression, correlation, logistic regression, or everything else that IS STATISTICS (or information theory, etc.)
3/10
2
6
62
As a scientific journal reviewer & Editor, the first thing I always check is to see if the research question matches the methods and if the methods match the reported results.
You would be shocked how many manuscripts don't meet this basic requirement.
6
4
63
COVID nearly had me saying "screw it, I won't renew my ATC credential" since I don't practice clinically these days. Glad I decided to keep representing the profession in my own little way.
1
1
63
When a Journal asks you to submit 8 potential reviewers for your manuscript, does anyone else feel like that exceeds the number of people you've met in your entire life?
8
2
60
I continue to not understand readiness scores:
-I have a miserable cold
-dog woke me up barking at 2 am
-3 year old threw tantrums at both 5 am and 7 am
... And this is my peak readiness? If so, I'm I'm trouble.
14
3
62
In the past 24 hrs, I got Moderna
#2
. I'd like to thank:
-The Chinese Scientist who sequenced the virus genome days into the pandemic and posted it publicly
-Bill Gates
-Dolly Parton
-Scientific ingenuity
-5G
-Alien DNA coursing through my veins
-Mind control microchips
4
3
59
I personally think that AI/ML has a lot to bring to the table to enhance science, health and human performance. The problem is that the AI/ML crowd are over-selling their wares and often being disingenuous about what is current state-of-the art
2/10
2
9
59
I've been doing it ad hoc for a couple of years now, but IT'S OFFICIAL:
My company Optimum Performance Analytics Associates LLC is open for business! We consult & provide analytical solutions to Professional & Collegiate Sports Teams, Defense Depts & Political Candidates.
6
7
58
Sport Science is not "data rich."
Sport Science is "good question poor."
2
5
59
#MachineLearning
Hype in Action:
Title: "Machine Learning Outperforms Regression Analysis!!..."
Reality: Logistic regression and their 84(!!) machine learning algorithms have nearly identical prediction performance
5
16
57
I'm really excited about my paper that was published in
@ExpPhysiol
where we use Bayesian decision theory to understand how variable human VO2 is during exercise. It was pretty cool {THREAD}
5
10
57
Is there a good GIF for p-value hacking?
Please share!
27
14
55
2nd manuscript in a row where I, as a reviewer, raise SIGNIFICANT concerns about the statistical practices (where I provide citations) but the Editor decides to accept manuscript.
It is hard to be optimistic about the direction of science when Handling Editors don't seem to care
6
7
57
Probably my favorite thing about
@JAT_NATA
is that it is a Society journal which is completely open access to reader, free to the author. They keep copyright (I think?), but at that point, I don't care.
It is literally a Journal to benefit the AT profession. 👏👏
Nice work
@JAT_NATA
with this FREE ACCESS special issue on training load in various pop’s;
Sports ⚽️ ⚾️ ,the arts 🎭💃🕺,
Fire fighters👩🚒👨🚒
0
18
74
0
6
57
I call bullshit 1 million times over.
9
5
56
Reasons Why Journals Should Have Statistical Editors:
1. Ensure analyses are adequately described & can be reproduced
2. Verify that Analyses match the goal of the paper/hypothesis.
3. Verify that Conclusions match Analysis Results
4. Remove Causal Language when unjustified
3
9
54
I was recently privately criticized about some of the things I've said on Twitter over the past year. It was suggested that what I saw here might hurt me professionally. While potentially true, I believe this dialog has been a net positive. Here is a thread why:
1/12
5
15
55
In closing, careful how you present yourself in a public forum. Done right, it can enrich your life personally & professionally. You may make missteps or have people who don't like you/agree with you, but that happens off Twitter too.
Be critical, be willing to be wrong.
12/fin
3
0
56
Happy to suggest accepting a manuscript today & I am really excited about manuscript. It fell in a funny middle ground where the outcome was "non-significant" and the equivalence testing was "non-significant", but the study was interesting/deserves publication. Thks
@JSCRonline
.
1
3
53
The more I connect with professional-level Sport Science and Strength & Conditioning coaches, I'm amazed at the life they've sign up for.
It is seemingly a new team & geographic location every 2-3 years. I don't understand how you establish a record of success or have a family.
8
6
56
Proposal: ALL Publicly funded research be available to the public free of charge and without paywalls.
If you're a taxpayer, you've already paid for the research so you should get to see its results.
Proposal: all NIH funded RCTs should mandate a preprint of primary trial results at same time as any press release/presentation of trial results.
Please comment.
19
14
113
1
7
53
Reviewer told me that studying menstrual cycle physiology wasn't important because "women are not a general population"
*I asked the Editor to ignore the sexist comments, they agreed and it was published in
@EJAP_official
What's the best reason why you've had a paper rejected?
I'll go first:
'There's a split-infinitive on page 6. As the paper has been so poorly proofread it's not worth my time reviewing'.
114
59
536
2
3
52
The ease of measuring things in the modern era has lead people to believe that everything you can give a number has value.
We've got a lot of people wasting a lot of time and resources measuring things that likely don't impact sport performance.
4
5
49
*STATISTICS REQUEST FOR HELP*
@avigotsky
,
@ExPhysStudent
& I are working to prove the invalidity of a "novel method to determine responders/non-responders"
We have written a concise forum post:
& are asking for input & correction from esteemed colleagues
5
18
51
I'm excited to announce that our R package for assessing motor unit synchronization is now available on CRAN!!
Notice bugs, issues or requested fixes?
Please spread the word
@Dr_DeFreitas
@mredumartinez
@RogerEnoka
@Dusty_Grooms
4
14
50
I'd like to invite you to submit to a Special Issue Journal called "Matt Reviewed This".
You send me the manuscript along with $2,000. Within 48 hours, you get a review. Upon acceptance, I 'publish' it with a citable DOI.
Courtesy of Matt's Digital Publishing Initiative.
4
0
50
I'm in the process of publishing my second paper that is a fully Bayesian analysis.
This is the second time a reviewer has asked about a power analysis for a fully Bayes paper.
What is wrong with scientists? This is basic.
7
4
50
My paper with
@Coach_Alejo
is out!
Hear what we think about the tension between the "Medical Model" and "High Performance Model" in sport.
Athlete Health and Human Performance Will Not Improve Without Transdisciplinary Collaboration and Data Sharing
4
8
49
Whoops, paper rejected.
Clinical journals just hate exploratory analyses (What do you mean there's no hypothesis!?) and papers that aren't ANOVAs or a straight Pearson correlation can be a hard sell.
Onward and upward!
#AcademicTwitter
2
2
50
Just spent 1 hour trying to get a manuscript from 1986 that seemingly can be found NOWHERE (even for purchase). Eventually realized the author was still on faculty. E-mailed him for reprint.
Author responded to e-mail within 5 minutes providing manuscript.
Lesson learned.
4
1
50
Life of a researcher at 4:55 pm on Friday:
"Ahhh, almost the weekend! Good to spend some time with the family."
"HERE ARE YOUR MANUSCRIPT PROOFS! YOU MUST RESPOND WITHIN 48 HOURS OR ELSE!"
3
0
48
Since I keep seeing people posting about the acute:chronic workload ratio, I'll take this moment to remind everyone about this excellent applied paper by
@LukeBornn
and
@OSPpatrick
showing it to be nonsense
3
13
48
Thanks to
@erikMeira
for the endorsement as an Athletic Trainer to follow on Twitter.
I'll take this opportunity to demonstrate to my Sports Med colleagues why you shouldn't trust the recently
@US_FDA
Authorized Q-Collar to protect against TBI
1/17
6
14
49
.
@ExPhysStudent
,
@avigotsky
& I have worked w/authors of a deeply flawed paper to issue a retraction
Authors refuse to retract despite clear evidence their error rates are far higher than reported
This *THREAD* is an example from MY OWN PUBLICATIONS how science should work
1/10
2
13
48
"...we believe that poor statistical education and practice are symptoms of and contributors to problems in science as a whole"
Best statistics introductory paragraph ever.
1
18
46
I've decided to re-publish every one of my research papers with Bayesian analyses. The only difference is that the title will start with "A Bayesian Analysis of..."
I look forward to doubling the length of my CV 🍻
4
1
45
Just saw this posted by an "influencer".
Red light therapy has no demonstrated Human Performance effects in humans, much less 40k papers showing it does.
I just saved your team/athletic department $150,000. You're welcome.
5
0
46
As a scientist, my guiding principles are as follows:
1. Science should be a meritocracy (it isn't)
2. Science is an iterative process of progress and backsliding
3. Publicly funded science shouldn't be owned by third parties
4. Dialog about science should be open and public
1
3
45
Finally, many "stats people" are interested in applying AI/ML techniques and seeing where it can compliment our backgrounds and current work. We just get turned off by the bravado, hype and salesmanship that accompanies AI/ML. So yeah, we'll keep giving it a hard time.
10/FIN.
4
1
44