Ichy's Moonlight Effect Experiment Horribly Flawed

Preliminary Post: This topic will be both technical and long. Please be patient. Here's the outline:
0) Review of Ichy's Experimental Design, referencing appropriate posts.
1) Analysis of Group 1
2) Group 2
3) Group 3
4) Group 4
5) Group 5
6) Conclusion: Experiment design fails to block unintended effects or to isolated effect to be studied
7) Review of statistical analysis
Conclusion: Error in considering the degrees of freedom invalidates all statistical conclusions.
9) Advice: My Approach to Designing the Experiment.

Rules of Engagement:
A) I'm not going to entertain semanitic issues or obvious trolling. EG and PP are warned that I will quickly ignore them should they chose to ignore this point.
B) I'm not going to educate you. Read about the science on your own. gotham and BF are warned.

0) Review of Ichy's Experimental Design, referencing appropriate posts.

I'm going to concentrate on one particular paragraph from the FEB forum. I'm not permitted to quote, but try this link: http://www.theflatearthsociety.org/forum/index.php?topic=36906.0. The post is timestamped: March 03, 2010, 01:03:59 PM.

The experiment took a total of 7 days and nights. It was taken in a climate controlled green house. Temperature data should verify little change at all in environment. As it was inclosed, the space was also from of variance in stress due to wind. (No fans present in building either). Plants were divided into 5 groups of which each had 12 plants. The first was a control group left untouched other than watering and nutrition. the 2nd was subjected to only sunlight by being covered by a pot every night. The 3rd was subject to only moonlight by being covered by a pot every day. The 4th and 5th were isolated in a corner and covered broadly (from a distance though blocking all window views) by a thick screen and only subject to artificial light of a lamp. The 4th was covered with a pot during the day, the 5th during the night. The sunlight only produced the number of rows of collenchyma as the control but were reduced in thickness. The moonlight only had many more rows and the 4th and 5th had similar results to the sun only group however, the thinning was much less noticeable. The plants were developed to have functional emergent cotyledons as photosynthetic functioning rose to optimum levels prior to experimentation (grown on same timetable).

1) Analysis of Group 1

Group 1 received both sunlight and moonlight. It was treated as the control group. It was not covered at any time. Since we don't know the dates of the experiment or the latitude of the experiment or how open the building was to the sky or the cloudiness of the weather during the experiment, we cannot make any conclusions about the amount of moonlight received by this group.

Note: we already see a woeful lack of recording.

2) Group 2

This group was to receive only sunlight, according to the author; however, he erred. The group was covered during only the night. The author forgets that the Moon shines during the day. Furthermore, the author has not blocked the effect of covering of the subject plants. The covering may afford this group an unintended advantage such as protection from temperature variances, loss of moisture, and adverse nocturnal pests.

Note: Clearly we're seeing a very poorly developed design.

3) Group 3

This group was subject to being covered during the day. Clearly this error screams for attention. If the experiment is about the harm of moonlight, then having a group denied sunlight makes no sense. Clearly our knowledge of photosynthesis provides sufficient understanding that such a covering is harmful in and of itself. There is no need to look to other factors.

Note: Now we know this experiment's design is so bad that its results are worthless. But let's press on for form alone.

4) Group 4

This group received no light, not sunlight, not moonlight, and not artificial light. It was covered during the daylight, though due to poor reporting, we don't know how many hours this was.

5) Group 5

This group complements Group 4. It was covered at night, though again we don't know for how long. It received no moonlight at all.

6) Conclusion: Experiment design fails to block unintended effects or to isolated effect to be studied

So which groups received the hypothesized harmful moonlight?

Group 1: All available
Group 2: That available during the day
Group 3: That available during the night
Group 4: None
Group 5: None

So which groups were harmed by reducing available sunlight/artificial light?

Group 1: No
Group 2: No
Group 3: Yes
Group 4: No
Group 5: Yes

So what groups should be compared to test the hypothesis? The only choice is 1 (control) and 2 (some moonlight). We must include the control. We can’t use 3, as it was harmed by other effects. We can’t use 4 or 5, as they didn’t receive moonlight.
So let’s compare the results of 1 and 2:

The difference in the median values between the two groups is not great enough to exclude the possibility that the difference is due to random sampling variability; there is not a statistically significant difference  (P = 0.687)

There you have it. No result justifies the conclusion.

7) Review of statistical analysis

Ichy runs many tests, but does so independently. He does not decrease the degrees of freedom with each test, a common mistake among novices.

He also did not establish his p-value BEFORE running the tests, as far as we know. I would expect that an alpha of 0.01 would be appropriate for someone hoping to demonstrate a public health hazard. He choses alpha of 0.05.

Conclusion: Error in considering the degrees of freedom invalidates all statistical conclusions.

Yes, he screwed up the statistics too.

Basically he did all that work, bragged about it and the FE-supporting results, and it's all horribly bogus. Oddly enough FEer commended his work, James and LW in particular. What happened to FEers' belief that only the most vetted experiments were acceptable? How could they turn around and accept such a horrible work as Ichy's?

I just want to say that ichi did his tests in a greenhouse so temperature, moisture, and pests were taken out of the variable equation.

9) Advice: My Approach to Designing the Experiment.

-Placeholder-

May I entreat everyone to review this critique before I write up my design in case Ichy would like to try again? Thanks.

I just want to say that ichi did his tests in a greenhouse so temperature, moisture, and pests were taken out of the variable equation.

I do  appreciate that they are lessened but don't agree that they're eliminated. I've encountered many pests in my parents' greenhouses. I've also seen many cases of microclimates in greenhouses, so moisture loss could well be greater for uncovered plants. Thanks for the input though.

This group was to receive only moonlight, according to the author; however, he erred. The group was covered during only the night.

Did you mean sunlight here?

Anyway, this is a very good critique. Perhaps you should also add a section about the conclusions he drew. Apparently his standard of "harm" was excess collenchyma growth, which he states can only possibly come from mechanical stress, which was impossible in those circumstances. Even if performed correctly, his experiment merely showed that collenchyma growth can be caused by another source.

9) Advice: My Approach to Designing the Experiment.

-Placeholder-

May I entreat everyone to review this critique before I write up my design in case Ichy would like to try again? Thanks.

My approach would simply have 2 groups.  One control group that is uncovered at all times and another group that is covered only at night.  I would also wait and run this experiment so that the dates are centered around a full moon as close to the winter solstice as possible.  Then, of course, document as many of the details as possible.

6) Conclusion: Experiment design fails to block unintended effects or to isolated effect to be studied

So which groups received the hypothesized harmful moonlight?

Group 1: All available
Group 2: That available during the day
Group 3: That available during the night
Group 4: None
Group 5: None

So which groups were harmed by reducing available sunlight/artificial light?

Group 1: No
Group 2: No
Group 3: Yes
Group 4: No
Group 5: Yes

So what groups should be compared to test the hypothesis? The only choice is 1 (control) and 2 (some moonlight). We must include the control. We can’t use 3, as it was harmed by other effects. We can’t use 4 or 5, as they didn’t receive moonlight.
So let’s compare the results of 1 and 2:

Quote from: Ichy

The difference in the median values between the two groups is not great enough to exclude the possibility that the difference is due to random sampling variability; there is not a statistically significant difference  (P = 0.687)

There you have it. No result justifies the conclusion.

Sorry but you cannot make up a conclusion about two variable interactions based on an experiment set up testing one.

Quote from: Lorddave on September 04, 2010, 03:01:32 AM

I just want to say that ichi did his tests in a greenhouse so temperature, moisture, and pests were taken out of the variable equation.

I do  appreciate that they are lessened but don't agree that they're eliminated. I've encountered many pests in my parents' greenhouses. I've also seen many cases of microclimates in greenhouses, so moisture loss could well be greater for uncovered plants. Thanks for the input though.

Temperature was closely and statistically monitored.

7) Review of statistical analysis

Ichy runs many tests, but does so independently. He does not decrease the degrees of freedom with each test, a common mistake among novices.

He also did not establish his p-value BEFORE running the tests, as far as we know. I would expect that an alpha of 0.01 would be appropriate for someone hoping to demonstrate a public health hazard. He choses alpha of 0.05.

Science, especially biological science calls for an alpha value of 0.05. That is elementary high school knowledge Clocktower.

As for the day and night samples, it was to consider photoperiodicity  and especially phosphoglycolate activity.

Sorry if you weren't able to realize any of this before, but my experiment was not to cater to people unfamiliar with the plant science Clocktower. Hope to hear from you again.

Sorry if you weren't able to realize any of this before, but my experiment was not to cater to people unfamiliar with the plant science

How can yo say that you are familiar with plant science when you take from plants one crucial component needed for their growth and afterward claim that the problems with plant were caused by something else than the lack of this one crucial component.

Quote from: ClockTower on September 04, 2010, 02:26:37 AM

This group was to receive only moonlight, according to the author; however, he erred. The group was covered during only the night.

Did you mean sunlight here?

Anyway, this is a very good critique. Perhaps you should also add a section about the conclusions he drew. Apparently his standard of "harm" was excess collenchyma growth, which he states can only possibly come from mechanical stress, which was impossible in those circumstances. Even if performed correctly, his experiment merely showed that collenchyma growth can be caused by another source.

Good catch. Yes, I meant sunlight there.

I agree with your critique and add it by reference here. Thanks.

Sorry but you cannot make up a conclusion about two variable interactions based on an experiment set up testing one.

I disagree. Data can be reused. Please provide a reference that supports your claim.

Regardless, rejecting a non-conclusive result is not very interesting to most of us. Let's agree that the experiment did not validate the hypothesis.

Temperature was closely and statistically monitored.

Please understand that your protocol did not prevent differences caused by just covering on temperature and humidity. You did not test either under the pots.

As I've made very clear: Your design failed to block other factors.

Science, especially biological science calls for an alpha value of 0.05. That is elementary high school knowledge Clocktower.

Not true. Reference: http://www.jerrydallal.com/LHSP/p05.htm 0.05 is a convention for determining whether to do more research, not to conclude the effect.

As for the day and night samples, it was to consider photoperiodicity  and especially phosphoglycolate activity.

And you erred in how you set that up. Shame really. You wasted a lot of time, resources, and effort when a 30-minute review with a real scientist would have fixed your design errors. Of course, we do appreciate the ridicule that you'd be exposing yourself to.

So you don't have any new arguments? I already shot down everything you have said already. Don't seem too desperate now.

Quote from: Ichimaru Gin :] on September 04, 2010, 08:39:55 AM

Temperature was closely and statistically monitored.

Please understand that your protocol did not prevent differences caused by just covering on temperature and humidity. You did not test either under the pots.

As I've made very clear: Your design failed to block other factors.

Yes, I'm sure in a climate controlled greenhouse, the temperature under the pot had such a staggering difference  How dare I take room temperatures but ignore the temperature of the soil, pot, floor, water I fed to the plants..etc. (It's even worse since they have nothing to do with collenchyma!)

Quote from: Ichimaru Gin :] on September 04, 2010, 08:39:55 AM

Sorry but you cannot make up a conclusion about two variable interactions based on an experiment set up testing one.

I disagree. Data can be reused. Please provide a reference that supports your claim.

Regardless, rejecting a non-conclusive result is not very interesting to most of us. Let's agree that the experiment did not validate the hypothesis.

Sorry but this clearly shows how flawed your own idea of an experiment and interpretting results is.  I plan to stick with scientific conventions but feel free to run free with your own experiments.

Quote from: Ichimaru Gin :] on September 04, 2010, 08:39:55 AM

Science, especially biological science calls for an alpha value of 0.05. That is elementary high school knowledge Clocktower.

Not true. Reference: http://www.jerrydallal.com/LHSP/p05.htm 0.05 is a convention for determining whether to do more research, not to conclude the effect.

Actually it IS used as evidence towards a conclusion.
Please do not talk about what you do not know. May I suggest a basic book on biological studies:
A handbook of Biological Investigation 7th edition by Harrison W. Ambrose III, Katharine Peckham Ambrose, Douglas J. Emlen, and Kerry L. Bright.
This book should be most useful to you since it has been shown you are ignorant of what to do with statistics in biolgy.

So you don't have any new arguments? I already shot down everything you have said already. Don't seem too desperate now.

Quote from: ClockTower on September 04, 2010, 01:59:59 PM

Quote from: Ichimaru Gin :] on September 04, 2010, 08:39:55 AM

Temperature was closely and statistically monitored.

Please understand that your protocol did not prevent differences caused by just covering on temperature and humidity. You did not test either under the pots.

As I've made very clear: Your design failed to block other factors.

Yes, I'm sure in a climate controlled greenhouse, the temperature under the pot had such a staggering difference  How dare I take room temperatures but ignore the temperature of the soil, pot, floor, water I fed to the plants..etc. (It's even worse since they have nothing to do with collenchyma!)

Quote from: ClockTower on September 04, 2010, 01:54:15 PM

Quote from: Ichimaru Gin :] on September 04, 2010, 08:39:55 AM

Sorry but you cannot make up a conclusion about two variable interactions based on an experiment set up testing one.

I disagree. Data can be reused. Please provide a reference that supports your claim.

Regardless, rejecting a non-conclusive result is not very interesting to most of us. Let's agree that the experiment did not validate the hypothesis.

Sorry but this clearly shows how flawed your own idea of an experiment and interpretting results is.  I plan to stick with scientific conventions but feel free to run free with your own experiments.

Quote from: ClockTower on September 04, 2010, 02:08:11 PM

Quote from: Ichimaru Gin :] on September 04, 2010, 08:39:55 AM

Science, especially biological science calls for an alpha value of 0.05. That is elementary high school knowledge Clocktower.

Not true. Reference: http://www.jerrydallal.com/LHSP/p05.htm 0.05 is a convention for determining whether to do more research, not to conclude the effect.

Actually it IS used as evidence towards a conclusion.
Please do not talk about what you do not know. May I suggest a basic book on biological studies:
A handbook of Biological Investigation 7th edition by Harrison W. Ambrose III, Katharine Peckham Ambrose, Douglas J. Emlen, and Kerry L. Bright.
This book should be most useful to you since it has been shown you are ignorant of what to do with statistics in biolgy.

And how many scientists have you had peer review your work ichi?  I'm sorry but you can't say your work is scientific until you have other scientists peer review it.  The Flat Earth Society doesn't count.

As I promised this summer, I am trying my best to get my work published in a science article. (Which means I would be open peer review outside of my peers  )
However, it is taking a LONGGG time just for a single study to be included. My dinosaur study alone could take another year.

As I promised this summer, I am trying my best to get my work published in a science article. (Which means I would be open peer review outside of my peers  )
However, it is taking a LONGGG time just for a single study to be included. My dinosaur study alone could take another year.

Excellent!
But... How many people in the biology field have you spoken to about it?

Quote from: Ichimaru Gin :] on September 04, 2010, 02:59:32 PM

As I promised this summer, I am trying my best to get my work published in a science article. (Which means I would be open peer review outside of my peers   )
However, it is taking a LONGGG time just for a single study to be included. My dinosaur study alone could take another year.

Excellent!
But... How many people in the biology field have you spoken to about it?

Many people with PhDs that teach throughout the metropolitan area. 90% of those I've met with teach at small liberal arts colleges.
The big boys at the D1, huge state research schools don't have time for me  

Quote from: Lorddave on September 04, 2010, 03:01:21 PM

Quote from: Ichimaru Gin :] on September 04, 2010, 02:59:32 PM

As I promised this summer, I am trying my best to get my work published in a science article. (Which means I would be open peer review outside of my peers   )
However, it is taking a LONGGG time just for a single study to be included. My dinosaur study alone could take another year.

Excellent!
But... How many people in the biology field have you spoken to about it?

Many people with PhDs that teach throughout the metropolitan area. 90% of those I've met with teach at small liberal arts colleges.
The big boys at the D1, huge state research schools don't have time for me  

Nothing wrong with liberal arts colleges.  Though a PhD in biology is not something I'd expect a lot of professors at small liberal arts colleges to have.  Not biology professors anyway.
Now did they all agree with your conclusions?

None of them replicated it :[. They were more interested in helping me submit my findings correctly to an article than reviewing them to be honest.

None of them replicated it :[. They were more interested in helping me submit my findings correctly to an article than reviewing them to be honest.

So none of them agreed that your conclusion was valid?