What is pseudoscience?

[Joe_Felsenstein]

@Dan_Eastwood

as uniformed about science as you.

Thanks. I salute you in gratitude. And maybe someone will tell me what the correct uniform is so I can get one.

[T_aquaticus]

@Biostats

Speaking as a statistician, I will say that when I am performing an analysis on a dataset to assess a specific hypothesis, I am indeed quite disappointed when there is little evidence for a preconceived idea. That doesn’t mean null results can’t be interesting or elucidating, but I intrinsically want my idea (or a colleague’s idea) to show merit. Try as I might to avoid this, this is often what I feel.

I think most scientists believe the Journal of Negative Results should be more popular than it is. At the same time, scientists don’t want to spend time and effort writing a paper on negative results.

@Biostats

My attitude is a bit different when performing an analysis that is more exploratory in nature to aid in hypothesis generation. Here, I am often more pleased to be shown to be “wrong” about preconceived notions because these results can be interesting, and I have no real horse in the race yet per se.

There is danger in forming hypotheses after data analysis. Experiments looking at gene expression data are can be prone to this danger. This is why any candidate genes found in the initial data are tested further in other assays to elucidate the mechanisms involved and to confirm that the relationships are real. Pseudoscience often looks at just one angle, or just one comparison. Good science tries to test an idea from multiple independent angles.

[Dan_Eastwood]

@T_aquaticus

Pseudoscience often looks at just one angle, or just one comparison.

Or thousands, and rotten cherry picking.

[Mercer (John Mercer)]

@T_aquaticus

There is danger in forming hypotheses after data analysis.

Very much. I’ve seen too many scientists who should know better pick one item out of a large set and present statistics as though they had started with the chosen item.

[Biostats (Scott)]

@T_aquaticus

There is danger in forming hypotheses after data analysis.

I agree wholeheartedly. This is why exploratory data analysis should often be focused on descriptive statistics and data visualization as opposed to modeling and inference. We must also be careful of data-mining and looking for something until we find something that “works.”

[Dan_Eastwood]

@Biostats

We must also be careful of data-mining and looking for something until we find something that “works.”

Also known as, “Torturing the data until it confesses.” There have been a few times in my career where I’ve had to say “No”. Telling them I don’t want my name on their manuscript is also very effective.

[AllenWitmerMiller (Semi-Expired Professor & Minister.)]

As to the OP, “What is pseudoscience?”, a telltale clue is anything with the word “discovery” in the name. Examples:

The Discovery Channel
The Discovery Institute

@Joe_Felsenstein

I salute you in gratitude. And maybe someone will tell me what the correct uniform is so I can get one.

It is the white lab coat. In photos it helps if you can have beakers and test tubes with colored liquids surrounding you. A stock photo background will do in a pinch. (See any “golden era” YEC or ID tutorial on Youtube. Also, see The Discovery Institute.)

@Mercer

To Axe I would add another marker: he (under supervision) did an earlier study that showed that barnase is resilient in the face of many mutations:

Oh! Oh! Pick me! Even I know this one: Barnase is an enzyme which breaks down barns.

@Dan_Eastwood

I have encountered homeopaths trying to do real science.

Me too. But I don’t think one’s personal orientation should matter.

[Dan_Eastwood]

@AllenWitmerMiller

Me too. But I don’t think one’s personal orientation should matter.

You’re kilin’ me here!