They use full white and full black for all modes. As we know, dark theme is typically white on dark gray (or equivalent), not pure white on pure black. I can't think of any dark mode that uses pure black. (AOSP, IOS, macOS, etc.)
This alone invalidates the study, because it doesn't reflect the real world. Science is only as good as it is rigorous, and if it fails to model the real world, then it's not very good.
Science, these days, seems to have become a faith based system, because most people don't usually review the content of the studies.
That (and bad science, like this study) is a shame.
1: https://journals.sagepub.com/doi/abs/10.1177/001872081351550...
> That (and bad science, like this study) is a shame.
Tangential, but this has indeed been a slow drift since the 1970s roughly. We are witnessing the consequences of that, now that entire generations of academics have been molded as such.
I think we need a vocal and principled reaffirmation of Positivism in science, to meaningfully move away from all the faith-based and broadly political conduct of scientific affairs at large (and the reporting thereof).
[0] https://www.litmus.com/wp-content/uploads/2020/04/bildschirm...
All of those studies are studies of how other people experienced dark mode vs. normal mode.
All of my studies are studies of how I experience dark mode vs. normal mode.
I feel that my sample more closely aligns to the population I'm trying to study.
Also it mentions spectrum-shifting software like 'night light' or f.lux or whatever Android and Windows call it when they do this by default.
And the thing about dark-on-white being more visible than white-on-dark is just blatantly wrong and the military anecdote makes no sense, camo against dark background in the dark is going to be low contrast because the camo you wear in the dark is dark, muted colours.
This whole "those studies are studies of how other people experienced X" is a typical argumentation scheme of esoteric fields like homeopathy, where science, logic and data is completely rejected and replaced with pure faith. Real studies can be criticized without relying on that trap.
We are literally discussing my subjective experiences here. And you're telling me I'm wrong. About my subjective experiences.
Ohhh-kay.
My notebook has a matte screen and barely reflects ambient light(s) back at me. I say barely to avoid saying "none" because surely someone will nitpick that it is reflecting a small amount of light etc.
The phone has a glossy screen, and it reflects every single thing it can back my eyes. Dark mode makes it even more "mirror like". And while I prefer the colours (specially if you're using the phone at night, with no ambient lights to bother), I acknowledge that it is more straining during most of the day. Even at night when it autoswitches to dark mode, I'll have some ambient illumination unless I'm at the cinema (not for a couple of months since lockdown) or basically trying to sleep (and I know I shouldn't be checking the phone if I'm trying to sleep).
Another data point: reading PDFs of books and conference papers on my computer (black on white) is tiresome, whereas reading the same on a kindle feels amazing. But then again, the kindles have won this battle a long time ago and nobody should be trying to read books or other media made for plant-based-paper on a screen anyway (and here I give e-ink a bit of leeway and consider it more plant-paper-like than a screen).
Anyway, to cut a long story short, you could be like the smoker who prefers to have another cigarette even though it's bad for him. Not only that, it's even possible to find examples of people being wrong about their own subjective experiences. For example, people sometimes have cyclic personal preferences and one of the arguments against them is that their subjective feelings are wrong.
The hidden assumption in that framing is that there's an objective measure of "what is best for them" which somehow isn't based on what people subjectively prefer.
1. Easier to read
2. Eye Strain
3. Battery savings
Your subjective "studies" about that are irrelevant if you put them in contrast to proper scientific studies for those things - if you don't measure them properly. And that's not how you feel about it. That's what I was saying.
And it does not matter for that whether you like dark or bright mode, whether you follow the article or speak against it :) That might be part of the confusion.
I have pretty bad eye floaters, and they're more visible in bright light. Having crud bouncing around my entire field of vision does not make for an easy or strain-free reading experience, so I enable dark mode whenever it's an option. I found checking Zulip at work pretty unpleasant until I realized I could enable dark mode.
If you're running a scientific study and basing your conclusions on the average of all participants, you might not pick up on things like that. And if you're running a scientific study on people with normal vision, a sufficiently strict definition of "normal vision" guarantees that you won't pick up on that, although the studies in question probably didn't use so strict a definition.
Even if dark mode is "scientifically worse" for most people, whatever that means, it's a useful accommodation. In fact, the Nielsen Norman Group article that this post links to argue against dark mode recommends dark mode as an accommodation for people with vision impairments.
As an example, one part of your mind could be extremely conditioned to want to please and be with someone who is (objectively) abusing you and so you think you're having a good time when they do show you attention, while another part of your mind hates them for hurting you. There are countless examples of how different parts of a single mind can be at odds with each other, and depending on when you think deeply about an experience, can both have enjoyed and not enjoyed an experience.
It's not unreasonable to think that similar mechanics are at play with more subtle, less social-based mechanics. Food (and drugs) have a lot of this going on too.
She looked askance, said "Well, its a little salty".
"Yeah, I guess it is."
"And not really very hot"
"Oh, yeah. Hm."
"And there no spice in it. Kinda plain."
"Dang, you're right"
So was I enjoying it? Or the experience of being out with the wife and not having to cook? Or just hungry and anything would do? And I'd mis-attributed what I was enjoying. Clearly my 'subjective experience' was not very, well, objective.
Or did you just let your wife bully you into not liking something that you, personally, actually liked?
Sure I was enjoying myself. This is kind of the point of this thread. Your enjoyment is not a real measure of what's good or bad.
Feeling efficient and being efficient are both valid bases for choosing, but the point of the article is that they don't have high correlation.
Dark mode could be more aesthetically pleasing, and yet could also be less legible. I have personally experienced this. Dark mode is hard to read and find my place in compared to light mode. The white space helps me absorb the information and use it effectively. I say this while I still think dark mode looks better for many apps. I just choose light mode because of utility and usability.
It's possible one can enjoy heroin, while it's simultaneously not beneficial for them to do so.
Just because something is subjectively positive, doesn't mean it's objectively positive.
regardless of the sample population, the difference is asking about perception vs. designing an experiment that provides consistent measures and accounts for bias. That's the definition of science. What you're arguing is "I know what I know", i.e. faith.
1. Start with a hypothesis, that "dark mode is less comfortable for me than light mode".
2. Design an experiment: find the sites that you use most. Use them for 10 minutes each in light mode and dark mode. Write down your impressions.
3. Build a conclusion: was your hypothesis correct? Was it partially correct for some sites?
Then you can adapt your behaviour according to your new, Scientifically-proven knowledge.
Science isn't faith-based, but there are a ton of problems with Academia and the way Science is practised and published at the moment. Science is also not connected with Academia - the scientific method doesn't need a university grant to work. You can literally conduct your own experiment in a couple of hours (as above) to get better information than any study. Not publishing it doesn't make it any less scientific.
It makes no sense to me to trust an unknown number of studies over my personal experience. I prefer dark mode. I'll stick with that. You do you.
Also, asking people their impressions can be helpful in the human aspect of the research, but quantitatively you need some sort of metric you can evaluate their experience on. For example, you'd assign a task and see how well people did comparing the two modes, while also making sure the difference is statistically significant (meaning it wasn't just as likely to be chance).
This is just the beginning of where good study design starts. You'd also do things like assigning the modes themselves randomly, so to go back to the audiophile example, I might notice if it's always x and then y, but not if it's scrambled. You'd could go further and try to stratify the groups, so for example making sure one group isn't all elderly people and the other young. It goes on and on...
So while the scientific method is nice, especially for introducing science in educational contexts, the methodology and rationale behind research is much more deliberate and involved. By all means they can try out things themselves, but no, they will not "get better information than any study".
One of the more important parts of the studies referenced in the NNGroup article the blog post in the OP is referring to is that they did experiments between people rather than subjecting people to different conditions consecutively. When the initial light-dark contrast isn't present, people's fatigue ratings were about equal for both modes.
Moreover, people perceived both modes equally hard to read, but were actually more efficient in reading lightmode text.
That advantage is driven largely by the sheer amount of light - some researchers did an experiment where they cranked up the brightness on a dark theme so the experienced brightness was the same and it was just as readable as a light theme. But then you're using a crapton of electricity since you're turning dark-hued pixels to be super bright.
Your personal experience is incapable of judging your performance in visual-acuity tasks and proofreading tasks, or determining the level of eyestrain caused when using light or dark mode.
Meanwhile, the scientists aren't trying to change your mind over which one is "more comfortable." They are trying to determine why you find something more comfortable and how well your the various options perform.
Maths has proofs. Logical reasoning from first arguments that a thing is true or not. It either is, or is not, proven. End of.
Physics as applied maths, pretty solid. We need to conduct experiments to verify that nature agrees with our mathematical constructs, but on the whole these are simple and unequivocal. e.g . gravity. Particle physics gets dodgy because of the vast range of collision results - statistics starts creeping in. Rather than being able to say that particle A collides with particle B to produce particles D an E, we now have a statistical chance that somethng might happen.
Chemistry as applied physics, again, pretty solid. Everything has to be tested by experiment, and experiment frequently throws up surprises, but if a reacts with B in Chicago, it probably does in Moscow too.
Biology as applied chemistry. Mostly solid. It gets massively complex, and so the temptation to resort to statistics is overwhelming and most biological papers start talking about statistical probabilities rather than actual results. But the basic biology is mostly the same for most subjects, and if the conclusion is simple (virology and the effectiveness of vaccines, for example) then we're all good.
Any social science as applied biology: not much. This is really dodgy territory where the experiment design totally dominates the result, and the result is statistical data that has to be massaged into a definitive statement. This is p-hacking territory, where experiments are largely unreproducible, very subject to bias, cultural references, and academia politics. E.g. whether creativity shares a limited resource pool with willpower - highly subjective, highly variable between individuals, hihgly suspect if your paper cites this as a proven result.
Not all science is worthy of the same level of trust. The scientific method is trustworthy. Academia is not.
Arguing with this is like disputing the fact that high heels make for terrible running shoes. Research in question with be quite similar.
Sampling. It probably feels good hand-waving away the problem of "does this study apply to everyone?" by introducing the "representative sample" construct, but it has its problems and I doubt they will be ever solved. You can't speak for everyone when you have several thousands samples -- out of ~8 billion people.
This is kind of akin to that artificially absurd example of "on average, every human on Earth has one testicle". (And don't get nitpicky here, please; it might be "median" or another term, and that's not the point.)
What your parent poster says is that this article is kind of hiding behind science to be able to claim a generalisation it makes is true. Which it still isn't. Most people I asked said they prefer dark mode. Some said they like light mode. This article changes none of that.
No. If I claim everyone has 0.5 testicles, that's bad science. If I randomly choose a large enough number of subjects and report that approximately 50% have two testicles and approximately 50% have no testicles, with error bars, p-values, etc., saying that you don't trust this because I am trying to make a career and you prefer to follow your own experience, which after looking between your legs clearly shows that 100% of people have two testicles, would be quite stupid. If I chose only 3 people, or all of them are male or female, or I make any other mistake, point the mistake, but don't attack me personally, and much less science in general.
For example, I followed the link to the study by the Nielsen Norman Group and then the reference to the study of Piepenbrock. They explain well their sampling method, with different groups by age and depending on vision problems. You can clearly see the individual results and the variance, and it is obvious that generalizing to the 100% of population would be wrong, but there are some very clear trends. Calling these researchers "someone trying to p-hack their way to a publishable result that's sensational enough to advance their career" without any proof whatsoever is insulting.
To be clear, I do not intend that anyone changes habits because of these studies. I agree that this is subjective enough to make it a personal decision. But as a scientist trying to make a career, I found the above comment very disrespectful.
Especially in the social sciences, there's been a whole discussion recently about how our current system of evaluating and rewarding scientists is not benefitting Science. There's even been high-profile commentators disputing whether the social sciences are actually Science at all.
As you're a scientist I won't bother explaining this to you. I'm sure you're aware of the problems here.
So my point is that for a study like this, there's lots of room for playing statistical games in order to achieve a more "sensational" result that is more publishable and more likely to get cited. We know this happens and we know this is especially rife in this area of study. So I have become much more sceptical of social-science studies showing broad generalised results about a subject applying to the whole human race. Especially if those studies contravene some commonly-held view about the subject. My default position has moved from "well, they know what they're doing so there must be something to it", to "I'm going to assume that they p-hacked their way to a sensational result until I have evidence to prove otherwise".
I might be wrong in taking that stance. I will change it if I have better evidence.
I get easily triggered when science is presented as a matter of faith, but in fact I totally agree with your skeptical point of view.
That it's conducted on humans (a.k.a. very small sample, not generalizable, non-objective metrics, hard to remove observer bias etc.)? I think regardless of personal opinion, just based on a Bayesian / base rate mental model, you should default to disbelieving any published social studies (I refuse to call it "science") that haven't been rigorously replicated.
Maybe, but most published studies are close to pure garbage and even in clinical trials which is supposed to be the holy grail of Science there's cherry picking, improper design, and lack of repeatability across the board.
"Science" is only as good as the humans conducting it. Unfortunately us humans are pretty bad at doing Science.
The "real" science has no absolutes. But the problem is that it is very difficult for humans to operationalize it that way. Take for example eggs and cholesterol. I remember in the late 80s, my father (also a scientist/biologist) stopped eating eggs because apparently Science said eggs are bad for you (due to some papers)... later in the mid/late 90s Science said that eggs are actually not bad, but good for you (because, even if they where high on cholesterol, it was the good cholesterol).
So, for people (even other scientists!) that are not experts in the subject, it becomes a matter of belief... believe in the papers some random team published, because it was published in Nature.
> Nothing is ever proven, but it contains a method to constantly find the theory most likely to deserve your faith.
I love this quote!
Comparing this to faith where from the start you cannot check anything doesn't make any sense (if you know that something is true or false it isn't faith anymore, that's the point).
Otherwise, you could believe any PDF you can find on Google and know it to be true and representative. What if someone wrote a net to generate 10,000,000 studies on the same set of topics and scattered them throughout the Internet? Given a random study, you wouldn't know whether it's real or generated, without the "authority" aspect of a journal.
Now, whether or not the journals actually do a good job at authenticating the studies is another question. But, the principle stands that they are what we trust as consumers of science (a role which scientific researchers themselves play as well).
I think this total lack of understanding of how the scientific process works is one of the biggest problems facing America today.
False. If you did not do the study/experiment yourself, you are putting your faith in scientists that did. This is philosophically equivalent to someone putting their faith in, say a monk, or a pastor.
If you go one level deeper, the monk might say, 'do X penance for Y years to verify Z claim', then it's up to you whether you want to follow that route or not. Until then, his claim is not unfalsifiable, like many skeptics claim.
Even for mundane day-to-day claims, 'science', as commonly understood, falls short.
e.g. Science cannot prove to me that a mango tastes sweet, without putting the condition that I must taste it. The only 'proof' it can provide is 'Taste it and see for yourself'. If I say, 'I will only taste it AFTER you prove it is sweet', then nothing will happen. Because Taste is subjective. Yet, everyone, 'miraculously' is able to come to a consensus.
There are truths that are individually/subjectively verifiable, but collectively/objectively unverifiable.
The denigration of the former type of truths is something armchair scientists must avoid. Real scientists never denigrate them.
Science can define what a sweet taste is, by assigning it to a set of measurements that define the boundaries of sweetness, as a technical term.
Is that sweetness to you? Nobody cares. It's a technical term that you need to accept to participate in the conversation productively.
"need to accept" is a "subjective" consensus - meaning it is useless if the terminology is not accepted.
Acceptance is a subjective decision, at which point you're simply going by majority vote. And majority is not a barometer for truth.
Science absolutely cares about subjective acceptance, and faith in experts.
These sorts of model, which are presented as capital S Science, are at best guesstimates, that don't receive nearly the amount of popular coverage when they shown to be wrong in the fullness of time (or indeed need to be reinforced by changing data). If a model is no better than a coin toss that's faith based to me.
A recent example; WHO "walks backs" statement made a day before about data showing that asymptomatic carriers infecting others is very rare. At the same time saying between 6% and 41% of the population may be asymptomatic with a 16% error margin. https://www.youtube.com/watch?v=Im0G7jb78jc
That doesn't mean I preach to someone that they shouldn't be drinking wine because beer is superior.
I doubt the rigor of MOST studies.