Trimethylpurine wrote:
...

You cant make observations without a hypothesis, and you cant make a hypothesis without a model. *see below for the difference.

....

A car mechanic does science all the time.

Scenario: Turn key, [observe that the] engine turns over, [observe that the] engine wont start.

Two obvious models: 1) No fuel getting into the engine
2) No spark.
...

[bold italics] added by me. If the mechanic doesn't notice (observe) that the engine turns over and that it won't start, unless he has these observations, either directly himself, or reported to him as observed results from turning the key by someone else, he has no basis to form the hypotheses.

Whether this is a scientific method or not is irrelevant to appreciate the results.

Some may feel that this study is not valid, and yet some will make conclusions about which coffee grinder is the "best" based on a grainy picture of ground coffee that somebody said is very "fluffy" and produces a "sweeter" espresso.

Bret wrote:[bold italics] added by me. If the mechanic doesn't notice (observe) that the engine turns over........

We're talking in circles now. Folks, let's keep it to something specifically about grinders or coffee please.

Bret wrote:Observations are the FIRST step in the scientific process, so making observations from which to form a hypothesis to test is absolutely a part of the scientific process.

+1! In medicine, which like coffee, involves science as well as art, careful observation has been, historically, the beginning of knowledge.

The initial observations of grind size by a single grinder of many different makes has failed to show a clear, easy to identify pathway to universally good coffee. It's more complicated than "a unimodal distribution of particles in X range gives better espresso than bimodal or broader particles".

One reason to post the data in its fetal (it's not even developed enough to be infancy) stages, is to get thoughtful people to look at the data and provide meaningful insight into what questions to ask next that might bring some breakthrough conclusions.

I think this is an initial attempt to identify and quantify all of the variables that go into an individual's decisions as to what is considered an excellent grinder.

Great thread!

Bret wrote:They posted the re-test of the Ek43 with the same coffee used to test the Breville, et al.





that a narrow particle distribution is better than a wider one, then my Breville Smart Grinder Pro is seeming to be a great bargain, especially as compared against the results in the original post.

Baratza has the particle distribution graphs online now. They have them free for download so I believe we are allowed to repost them, but this is found http://www.baratza.com/wp-content/uploa ... -Graph.pdf

It looks nothing like the graph Socratic came up with, really bringing into question the results:

day wrote:Baratza has the particle distribution graphs online now... It looks nothing like the graph Socratic came up with, really bringing into question the results:

i am not sure which grinder setting Socratic used, but the two graphs aren't too dissimilar. The socratic plots are done using sieving where you can't capture the size distribution on the finer end. Also, the socratic plot uses a linear scale for the particle size while the Baratza plot uses a logarithmic scale.

And it wasn't stated clearly whether the y-axis on Baratza's is volume %, particle count or surface area. The choice of representation will change the shape significantly, and the interpretation too as a result. Also, there are more than one source(rep from Bunn & Mahlkonig, just from my memory) that say you can't compare the results from different methods directly, as explained by David above (each method has its own bias).

I hear what your saying. As far as I can tell they measured it in volume of coffee, where 1q is .25mm. While they may have used different methods that dont directly translate, it seems to my uneducated eye that the Baratza graph clearly shows the volume of coffee falling in a much wider distribution the finer you grind, and a much narrower distribution the coarser you grind. Because the second peak in the 150um is at 400um and then trails off at 800um, it would seem that they would have to have ground quite a bit finer to get the second peak to 300um. This would not only go outside of the stated specs for the Preciso but it would also imply a much less unimodal grind. That would, in turn, make the clearly defined peak presented by Socratic a bit confounding, even when considering the apparent differences in methodology, would it not?