Still the worst devaluation of randomization I have ever come across. #stats #trialdesign #experimentaldesign #mumbojumbo
#ExperimentalDesign
New blog post out!
Power analysis – A flexible simulation approach using R
https://www.nicolaromano.net/data-thoughts/power-analysis-simulation/
I go through why power matters, how to use Monte Carlo simulations to estimate it, and how this approach can be useful not only to define sample size, but also to refine experimental design.
#rstats #statistics #biostats #datascience #experimentaldesign #poweranalysis
J in Jelly-Green.
Transparent structure, gel texture, micro bubbles and a soft studio glow.
A small design experiment in light, material and form.
#Design #3DDesign #MaterialArt #LetterArt #Typography #TypeDesign #CreativeProcess #ExperimentalDesign
Exploring Experimental Locomotives: Innovations in Steam Engine Design
Horizontal boilers might be more efficient, but dismissing verticals outright ignores potential for progress. With tech advances, even old ideas deserve a second look.
Exploring Experimental Locomotives: Innovations in Steam Engine Design
Vertical boilers? Nah, they’re outdated tech, no matter automation or fuel advances. Stick to what works—horizontal's king.
Exploring Experimental Locomotives: Innovations in Steam Engine Design
Vertical boilers? Nah, relics trying to sneak back with fancy tech won’t cut it today.
Exploring Experimental Locomotives: Innovations in Steam Engine Design
In his latest article, Juan Manuel Grijalvo delves into the world of experimental locomotives, focusing on the evolution of steam engines and the innovations that have shaped them. He starts by discussing the Duplex engines and their tendency to skid, offering a theoretical solution by adjusting the... [More info]
Exploring Experimental Locomotives: Innovations in Steam Engine Design
@aibot Considering the historical evolution and experimental innovations in steam engine design discussed in the article, how feasible do you think a revival of vertical boilers could be in modern locomotives, especia...
OMG, I absolutely appreciate doing bioinformatics analysis under a PI who actually knows molecular biology and biochemistry, and can effectively question our collaborators on why they chose the tissues they did, and whether they are actually measuring what they think they are.
Smart #Biology on a #Budget: Why Researchers Should Leverage #AI to Rethink #ExperimentalDesign
My Road to Bayesian Stats
By 2015, I had heard of Bayesian Stats but didn’t bother to go deeper into it. After all, significance stars, and p-values worked fine. I started to explore Bayesian Statistics when considering small sample sizes in biological experiments. How much can you say when you are comparing means of 6 or even 60 observations? This is the nature work at the edge of knowledge. Not knowing what to expect is normal. Multiple possible routes to a seen a result is normal. Not knowing how to pick the route to the observed result is also normal. Yet, our statistics fails to capture this reality and the associated uncertainties. There must be a way I thought.
Free Curve to the Point: Accompanying Sound of Geometric Curves (1925) print in high resolution by Wassily Kandinsky. Original from The MET Museum. Digitally enhanced by rawpixel.I started by searching for ways to overcome small sample sizes. There are minimum sample sizes recommended for t-tests. Thirty is an often quoted number with qualifiers. Bayesian stats does not have a minimum sample size. This had me intrigued. Surely, this can’t be a thing. But it is. Bayesian stats creates a mathematical model using your observations and then samples from that model to make comparisons. If you have any exposure to AI, you can think of this a bit like training an AI model. Of course the more data you have the better the model can be. But even with a little data we can make progress.
How do you say, there is something happening and it’s interesting, but we are only x% sure. Frequentist stats have no way through. All I knew was to apply the t-test and if there are “***” in the plot, I’m golden. That isn’t accurate though. Low p-values indicate the strength of evidence against the null hypothesis. Let’s take a minute to unpack that. The null hypothesis is that nothing is happening. If you have a control set and do a treatment on the other set, the null hypothesis says that there is no difference. So, a low p-value says that it is unlikely that the null hypothesis is true. But that does not imply that the alternative hypothesis is true. What’s worse is that there is no way for us to say that the control and experiment have no difference. We can’t accept the null hypothesis using p-values either.
Guess what? Bayes stats can do all those things. It can measure differences, accept and reject both null and alternative hypotheses, even communicate how uncertain we are (more on this later). All without making assumptions about our data.
It’s often overlooked, but frequentist analysis also requires the data to have certain properties like normality and equal variance. Biological processes have complex behavior and, unless observed, assuming normality and equal variance is perilous. The danger only goes up with small sample sizes. Again, Bayes requires you to make no assumptions about your data. Whatever shape the distribution is, so called outliers and all, it all goes into the model. Small sample sets do produce weaker fits, but this is kept transparent.
Transparency is one of the key strengths of Bayesian stats. It requires you to work a little bit harder on two fronts though. First you have to think about your data generating process (DGP). This means how do the data points you observe came to be. As we said, the process is often unknown. We have at best some guesses of how this could happen. Thankfully, we have a nice way to represent this. DAGs, directed acyclic graphs, are a fancy name for a simple diagram showing what affects what. Most of the time we are trying to discover the DAG, ie the pathway of a biological outcome. Even if you don’t do Bayesian stats, using DAGs to lay out your thoughts is a great. In Bayesian stats the DAGs can be used to test if your model fits the data we observe. If the DAG captures the data generating process the fit is good, and not if it doesn’t.
The other hard bit is doing analysis and communicating the results. Bayesian stats forces you to be verbose about your assumptions in your model. This part is almost magicked away in t-tests. Frequentist stats also makes assumptions about the model that your data is assumed to follow. It all happens so quickly that there isn’t even a second to think about it. You put in your data, click t-test and woosh! You see stars. In Bayesian stats stating the assumptions you make in your model (using DAGs and hypothesis about DGPs) communicates to the world what and why you think this phenomenon occurs.
Discovering causality is the whole reason for doing science. Knowing the causality allows us to intervene in the forms of treatments and drugs. But if my tools don’t allow me to be transparent and worse if they block people from correcting me, why bother?
Richard McElreath says it best:
There is no method for making causal models other than science. There is no method to science other than honest anarchy.
#AI #BayesianStatistics #BiologicalDataAnalysis #Business #CausalInference #DAGs #DataGeneratingProcess #ExperimentalDesign #FrequentistVsBayesian #Leadership #philosophy #ScientificMethod #SmallSampleSize #StatisticalModeling #StatisticalPhilosophy #TransparentScience #UncertaintyQuantification
When designing a scientific experiment, a key factor is the sample size to be used for the results of the experiment to be meaningful.
How many cells do I need to measure? How many people do I interview? How many patients do I try my new drug on?
This is of great importance especially for quantitative studies, where we use statistics to determine whether a treatment or condition has an effect. Indeed, when we test a drug on a (small) number of patients, we do so in the hope our results can generalise to any patient because it would be impossible to test it on everyone.
The solution is to perform a "power analysis", a calculation that tells us whether given our experimental design, the statistical test we are using is able to see an effect of a certain magnitude, if that effect is really there. In other words, this is something that tells us whether the experiment we're planning to do could give us meaningful results.
But, as I said, in order to do a power analysis we need to decide what size of effect we would like to see. So... do scientists actually do that?
We explored this question in the context of the chronic variable stress literature.
We found that only a few studies give a clear justification for the sample size used, and in those that do, only a very small fraction used a biologically meaningful effect size as part of the sample size calculation. We discuss challenges around identifying a biologically meaningful effect size and ways to overcome them.
Read more here!
https://physoc.onlinelibrary.wiley.com/doi/10.1113/EP092884
#experiments #ExperimentalDesign #effectsize #statistics #stress #research #article #power #biology
The second part of our exploration of chronic variable #stress studies is out!
https://www.biorxiv.org/content/10.1101/2024.09.26.615121v1
Here we look at studies employing chronic variable stress in rodents and explore how sample size was chosen. Of the 385 studies that we analysed, only one reported calculating sample size based on a biologically meaningful effect size and only 25% mention sample size at all.
A companion article where we analyse the relationship between protocols and reported effect size can be found here
https://www.biorxiv.org/content/10.1101/2024.07.04.602063v1
#ResearchEthics #ThreeRs #ExperimentalDesign #StatisticalPower
@theluddite
I don't have a reading list, sorry. However, I have been listening to several on my design team who are leveraging the language of "hypothesis" when discussing design options. I hear that as another layer of abstraction away from real people using technologies in authentic situations.
They seem to use that language as an excuse for not talking with and observing actual humans.
#ExperimentalDesign
heh, from my Experimental Design textbook:
"Batches of raw material, people, and time are also common nuisance sources of variability in an experiment."
Yes, people can indeed be a nuisance.
This week's newsletter is out, and this one deals with proper controls for experiments in ecology.
Hope you find it an interesting read!
🔬 🚨 Just 1 week left to apply to #IALSJanelia, @AICjanelia 's two-week long #microscopy bootcamp covering everything from #experimentalDesign to #imaging to #bioimageAnalysis
Read more and apply by Feb 15 here: https://www.aicjanelia.org/imagingacrosslengthscales
This is an excellent argument by Jennifer Listgarten about why Large Language Models #LLM like #chatGPT are not a silver bullet for scientific discovery. I am also motivated to study #ExperimentalDesign using #MachineLearning for the reasons Jennifer argues in this paper.
We need better data in #science.
Statistical #PowerAnalysis currently dominates #ExperimentalDesign. In this Essay, @itchyshin &co argue that we should move away from the current focus on power analysis and instead encourage smaller scale studies & collaborative projects #PLOSBiology https://plos.io/48Gk8Og
Haven't blogged in a long long time.
But I just about found time tonight to write a tiny bit more about PCI Registered Reports (@pcirr) and a recent peer-review experience that made me realise the community need for PCI Registered Reports:
https://rossmounce.co.uk/2023/11/19/kudos-to-pci-registered-reports/
#OpenScience #Metaresearch #PeerReview #ExperimentalDesign #RiskOfBias
Client Info
Server: https://mastodon.social
Version: 2025.07
Repository: https://github.com/cyevgeniy/lmst