What methods are used to count insects

Googled something ancient and interesting https://studfiles.net/preview/2427515/page:56/.

Surely everything is progressing now? There are also computers)

There are no such computers that would catch and count insects themselves yet. In general, everything depends on the specific group: from the traps of Malez and Mekrike to the " number of butterflies per kilometer of the route "(adjusted for the visual acuity of the accountant).

I was referring to what is now called "data analysis".

Then You will need here.

methods ALWAYS depend on goals and objects. eggs are hundreds of times more than adults, for example, usually.
relative estimates (the percentage of populated plants, for example, determining the economic threshold of harmfulness for carrying out treatments or not) seem more meaningful than futile attempts in a non-uniform space to get absolute values that will be different tomorrow. Or dynamics, for comparative mining methods-a trap in one place or parallel for comparison, for example.

This post was edited by CosMosk-06.07.2019 01: 56

I will bet: just with" bare " percentages in statistical analysis, there is a lot of hemorrhoids. For example, while walking through biotope A, a researcher noticed two plants of a certain type, one of them was inhabited by a phytophage, and the second was not-the population is 50%. And in biotope B-three plants, one inhabited and two-not-comes out 33.3 (3)%. Can we say that in biotope A the population is higher than in biotope B? No. But that's because we're borrowing absolute numbers. And if you had only percentages in the table in front of you, and without confidence intervals, as is often the case in entomology publications, then you would only have to throw up your hands. Alas, we must admit that in Russian entomology, including in the editorial boards of journals, with knowledge of statistics, mostly sad, because they print whatever they want. But, for example, doctors have strict control, "naked" percentages will not be missed. Here, more reading material.

ENO, can you pre-estimate the sample size required for a given confidence level? (namely, that it is most often not indicated (and therefore " our " ecol.works in internat.they are practically not found in journals and are not quoted), and any figures (for example, without reliability, very little means-about scientific populism and "scientists have proved") Here, apparently, the reliability is not great, and this is normal if it is specified. for some purposes, it will do-the presence/absence at all, as for quarantine or particularly dangerous pathogens.
I agree with you, but I'm not describing the techniques here) In any case , it all starts from the goal, if the goal is not to ride-to justify the purchased equipment as an end in itself.
Thanks for the article on the topic-it looks very healing, but alas, I'm not a tech guy-not a smart programmer, so I'll apply it instead of a compress.
And, anyone can call themselves a researcher, a boneless tongue)


This post was edited by CosMosk-06.07.2019 02: 10

I don't really understand what you mean. There is no such thing as a "confidence level" at all, there is a level of significance. When the achieved significance level is less than a given critical point, they often talk about the reliability of an alternative hypothesis, but this, by and large, is also wrong, since reliability is a busy term denoting a probability equal to 1. It is more correct to say "significance".

But I have a hunch. what did you mean by that power. In order to calculate it, you must first have such a wonderful indicator as the strength (magnitude) of the effect. Where it is taken in practice - this mystery is great, apparently, the chosen ones from above are let down, and the rest are sucked out of their fingers. Doctors have the easiest way with this case, they rely on"clinical significance". Well, for example, if a new drug compared to the old one will give an increase in patient survival by 1%, then for such a result, especially in bourgeoisie, the researcher may not be praised, but quite the opposite, for embezzlement. Here, for example, 10% is another matter, and this value is set when calculating the required sample size (while the power is usually set to 0.8, this is their standard). Entomologists do not have any clinical significance, the strength of the effect is much less interesting than the fact of its presence. Therefore, I recommend that you skip the power analysis for real-world research in this area, and choose the largest sample size that you feel the strength to process. And yes, I haven't seen any works on entomology with power analysis in bourgeois journals, and the criteria for selecting sample sizes are usually kept silent. In general, planning experiments on mice/humans and environmental studies are two big differences, since in the latter case we are dealing not with an experiment, but with observation. Here one particularly advanced "dataminer", I remember, generally claimed that mathematical statistics are not applicable to field research, but IMHO these are his personal cockroaches, mathematics is universal. You just need to take what nature graciously gives you, and not show off with attempts to make preliminary calculations of volumes, capacities, etc.But this does not negate a competent description. For example, instead of the percentage of population, it is much more informative to give the number of inhabited plants and the number of uninhabited ones (or the total number of studied ones). Let it be two columns in the table instead of one, but the reader will always be able to calculate them himself in the way he likes best.

Yes, thank you all for developing the topic. Interesting. And for interesting links.

ENO, and in what topic did dataminer write?


It may also be useful to verify the methods. For example, using analysis, we obtained tree data. And then they cut it down, roughly speaking, and each copy was directly counted and described. And we checked the method. If it came together, then we will continue to use boldly, without sawing trees.

Yes figs you will remember. It is clear that in the section "Biophysics and mathematical methods in biology".

I didn't understand about the tree. What kind of magic analysis will help you count insects on a tree without actually counting them?

I mean samples again and confidence intervals. We know the behavior of insects: for example, they like the sun, something else, seasonality. And one branch was counted, the second, and estimated that probably more likely on other branches it will be about the same.

And by the way, there is also a kind of method when we mark instances to understand how often they come across again, etc.

It was a long time ago.. The girl called me to count foragers formik, and there, according to the method, the interval should be between calculations, well, we filled it out "with what we could" and neither bumps under the back, nor horseflies, nor foragers could prevent us - so much for the error and the influence of factors, remembering

I remembered the Gaza Strip - "mosquitoes".

AGG, what's the problem?

 

OK, "significance". How likely is the result to be reliable? I'm not a programmer, so I probably can't master statistics, no matter what, alas, alas.
Digging into words instead of direct answers immediately tires me out. Either it is clear what I wanted to say and the speech is in essence, or "around-yes-about" - as women "just do not understand and do not recognize anything."
Although I agree that research should also begin with the definition of ambiguous concepts used (total fail of philosophers and other "neuro-conitivists").

"a term denoting a probability equal to 1. It is more correct to say "significance". " - there is an opinion that a vanishingly rare event, incredible, but still already happened, has a probability of 1, like our existence))

Your second paragraph is pretty strong noodles, sorry..
Cool book imho, the second half of which I do not mind at all. But where the idea is developed in words, without numbers-it is clear and practical:
http://avtor-kmk.ru/pages/showitem.php?id=608
(sorry for the freaking image size)


This post was edited by CosMosk - 07.07.2019 11: 16

I mean samples again and confidence intervals. We know the behavior of insects: for example, they like the sun, something else, seasonality. And one branch was counted, the second, and estimated that probably more likely on other branches it will be about the same.

And by the way, there is also a kind of method when we mark instances to understand how often we come across them again, etc.

For each species, it is exhausting to calculate their own "distribution" coefficients, since they "aggregate" differently and in general different species are different species, just like the whale and hippopotamus are artiodactyls. Well, the landscape is extremely heterogeneous, In any day you can collect two or dozens of copies. one type, and NEVER again, and what should I do with it mathematically?
Therefore, I consider rare and special species to emphasize the features of the "ecosystem", but first of all, dominants and subdominants should be studied-in terms of ecological leadership. Then the first ones, the rarities, also take on some kind of non-integral context.
Or some special features, again in terms of goals and objectives, as one quarantine bug should be enough for complete quarantine and genocide around.

This post was edited by CosMosk - 07.07.2019 11: 27

So you don't have to be a programmer to perform statistical processing - there are a lot of push-button programs. Although non-button ones, of course, are much cooler in terms of the arsenal of methods and the ability to design your own for a specific task.

That's exactly what you need to define the terms correctly. If you, for example, called spiders insects, and I corrected them, would you also consider it a quibble and immediately "get tired"? And since there is still no normal formulation of questions, what "direct answers" can be given?

I didn't understand the last post at all. I can only say that depending on the purpose of the study, it may focus on dominants and recedents. or the entire community. In general, synecological research in terms of statistical processing is the most complex and diverse in biology, moreover, it is one of the growth points of modern applied statistics. There is a lot of literature on this subject, mainly in Bourgeois languages. In Russian, I recommend Rosenberg and his friends.

I first thanked you for the link to the book, I didn't read it, but when I clicked on it, I was blown away - they demand a dime. Was that an advertisement?

Well, here found the right link to Kozlov this, for those who have an extra 180 rubles. no. I scrolled through it diagonally. First impressions are not very good. There are a lot of words, but there is little information that is useful for real research. Unless, the recommendation to bend the words " reliability "in favor of the phrase" statistical significance", which I immediately said. Otherwise, it is a hodgepodge of recipes set out in a form that practically guarantees their thoughtless use by novice researchers, which is fraught. Some of them IMHO definitely harmful. Tables and graphs that would fit in the previous century, when there was such a thing as "computer time". And there is a lot of water in the form of an unnecessarily detailed description of special cases, namely those aspects of them that are not directly related to the topic of statistical analysis. And the interpretation of the concept of "experiment" as any data collection process is very unusual, at least for the Russian school.

This post was edited by ENO-08.07.2019 01: 46

Thank you.

Interesting. I don't know how cool the book is, but the first impression seems to be exactly what I wanted.

"Computer time" is now also a concept, when we calculate or model something, it is introduced as a performance indicator in fact.

Where is it entered? I haven't seen any biology publications that mention this. Now all the houses consider, for the most part, pirated "Statistics" and "SPSS". And earlier they were recorded in the queue for the intitut computer. But they were not biologists, biologists tried to count something on their knees sometimes. And as a reassurance, we still write manuals on checking the normality of chi-squared and other wildness. Of course, even now specialists in some branches of knowledge are being lined up to join supercomputers and clusters of them, but counting insects is not the case here.

I do not agree, although the author Kozlov threw us 200-300doll.for analyzing his material on flies for 2-3months and creating a table..
(someone else wants to fish for the grant for their own pleasure, and get the summary data ready for processing? - a fucking native will earn more money as a cab driver in 10 days than a "scientific subcontractor")
Although for the book I again respected him a little, spending another 180r. in favor of KMK.

This post was edited by CosMosk - 11.07.2019 03: 16

I have made several critical theses, and what specific ones do you disagree with? I suggest leaving the author's identity aside. It happens that mediocre researchers write well and vice versa.

I mean "computer time" is a term used when modeling on a computer, which reflects how long the processor worked.