No, the following does not belong into the humor section, because I know of people who made a career with the method described below. This is serious! This is another article in my series on the usual cheating in science.

POP-science culture (POP = publish or perish = popular) ensures that only publications count in academia. Successful grant applications also count, but the grant you get only if you have many publications. And “friends” count, which you get with coauthoring and publications. And all that impacts science – no conspiracy theory necessary here. This is science today:

(Oh - BTW, to write this article in this way was inspired by Computer Science Research Generator)

Straw-man: “I want to publish more papers. Give me some inspiration, will you?”

Easy. Just take some law of chemistry or material-science relation. Any property will do.

Straw-man: “What, like melting temperature and how it depends on several other parameters, say crystal structure or whatever?”

Perfect! Now look for any parameter P in there that may depend on the size, or radius R of the system.

Straw-man: “They all could.”

Yes, but you need to find or put a radius R in there somehow.

Straw-man: “Hmm, I guess this wavelength lambda here, if I imagine that it would not fit into something that is smaller than half the wavelength, then I can put it equal to 2R.”

You are good! We are almost done. Now express parameter P as it depends on R, that is P(R), in form of a ratio P(R)/P. Remember, P is just the value for large R relative to the nanometer scale, i.e. the usual value. Look up some tabulated value for P, say the known parameter for gold. Gold is often used in nanotechnology; you will find thousands of useful references.

Straw-man: “What now?”

This is your first graph: P(R)/P versus R. See how profound it is? See how the graph goes to unity P(R)/P = 1 at large R? See the novel prediction, how it goes to P(R)/P = 0 at a mysterious critical cut-off value Rc at small R? Or maybe to a certain finite value at R = 0?

Straw-man: “Hmm, not really. I mean, there is a value at R = 0, but it is still almost equal to unity.”

The small difference problem, I see. Try again for 1/P, or (1-P) or anything that can blow it up.

Straw-man: “Wait, this, no, what about ……. Wow, beautiful! P(R)/P = 0 at Rc of about 5 nanometers.”

Perfect – a prediction. You got yourself a medium impact factor journal article.

Straw-man: “Ah, that was fun, but lets aim for a higher impact factor.”

Well, high IF needs comparison to experimental data. Pick some data. Anything goes!

Straw-man: “But there are no data on P(R).”

I said: Anything goes!

Straw-man: “Like what? My colleague’s pile of papers on nano tubes? There are plenty of data.”

Perfect. Pick a parameter.

Straw-man: “Well, there is a paper on a so called pseudo-magnetic transition temperature T in Gold/Silver mixed nanotubes. Would that work?”

Everything works!

Straw-man: “But what is R now, the radius of the circumference or half the length of the tube?”

Try first one, leave the other for the next paper. Now argue that one could reasonably expect P(R)/P to behave like T(R)/T.

Straw-man: “But there is no connection!”

Come on, google references, argue thermodynamics, relaxation, a little bit of dimensional analysis. Don’t you see it! Smoke some!

Straw-man: Cough, Cough. “Well, I could kind of see that maybe that crazy quote over here could be misunderstood to say that the square root of P(R)/P perhaps behaves similar to magnetic hardening since both are unrelated to second order phase transitions, at least to first order in the Taylor expansion, kind off, shit, I was holding this graph upside down just now.”

Aha – there you go, so square root of P(R)/P is the new T(R)/T. See that is why we use ratios, they have no units, so you can set them equal without any worry. This is your semi-empirical, analytically exact model. Smooth going there, Straw-man. Just write something that sounds as if you really know your stuff and fill in at least three or more somewhat not too un-related references whenever it doesn’t sound quite right; must be enough for reviewers to be too busy to look into. Now use the known T for gold to plot your novel model for T(R)/T.

Straw-man: “Well, the data fit on the right, but they must fit there, because it is just the ratio going to T/T = 1.”

That’s the beauty of it, my friend; they fit over most of the plot already before we even get started. How could they not.

Straw-man: “But they don’t fit at all below 70 nanometers. And by the way, there are no data below 30nm.”

Yes, yes, that’s good, no data where the predicted Rc for this transition is. That is your insurance, my friend. And about the rest in the middle, what if you use the average between the value of T for Silver and that of Gold?

“Even worse! Doesn’t fit at all.”

Make Rc a free parameter and argue it to be, well, hmm, it is a “pseudo-magnetic” transition you say? Well argue it to be equal to the smallest radius necessary for the particles to become fully metallic or have a nice plasmon transition or whatever. Can you fit the data?

Straw-man: “No! Before, the model was too high, now it is too low.”

What about the ratio of silver and gold in the nano tubes?

Straw-man: “On average there is about three quarters gold by weight. But it’s a broad distribution; some tubes in the sample are pure gold.”

Large uncertainty eh, good! Three quarters you say, so use Rc = (3/4) Rc (gold) + (1/4) Rc (silver).

Straw-man: “I see; good idea. Fits better, too! But it is not a good fit.”

Well, try Rc = [(3/4) R3c (gold) + (1/4) R3c (silver)]1/3.

Straw-man: “Wait. Why? There is no reason to use some power.”

Come on, the transition is obviously sensitive to the volume of the material, and radius cubed is volume.

Straw-man: “Jesus, you are jaded! But the fit went from too low to a little too high again.”

Try Rc squared instead of cubed.

Straw-man: “What the hell?”

Just write that in the nano size realm the surface to volume ratio is very large and that the nano tubes have almost all atoms being surface atoms. Want me to give you some references? You will be fine. See, surface is more important than volume.

Straw-man: “Gosh, you are not only jaded! You are evil! You are a scientist, aren’t you?”

The plot, Straw-man, how does the plot look like?

Straw-man: “Damn! Close!”

Did you not say the nano tube paper had large uncertainties? Never forget the magic of error calculation! Add error bars.

Straw-man: “Holy cow. Look man.”


Congratulations! You got yourself a paper! It is going to be rejected about two times, but on average, with the third journal, it is going through because the reviewers have no time to read and check the major revision they may likely request. You see, they are busy writing papers! Good luck.