Hmmm anyone want to talk about this one??
- Thread starter mojoreef
- Start date
Help Support Reef Aquarium & Tank Building Forum:
Jiddy
Well-known member
No thanks Mojo
Scooterman
Well-known member
Fine sand seem to house less phosphates, the end results looks like a close race to the finish. 108 days isn't very long testing & the BB was not even tested.
I can't wait to see the next part!
I can't wait to see the next part!
NaH2O
Well-known member
- Joined
- Jan 25, 2004
- Messages
- 8,568
mojoreef said:
ROFL! I'm glad I'm not the only one that noticed that! Holy cow....72F - 86F.
How can you measure and put emphasis on mortality rates with these temperature fluctuations. Why not put some controls on these variables....how 'bout a chiller, or something to maintain a static temperature? I'm sure I would see some death in my tank if I subjected it to these kinds of swings.
I really like this...
Ha! Glad to know my tank is a death trap :lol: - at least is has some beautiful equipment to look at. CLEARLY the data shows that a bare bottom tank would be horrible. I wonder how urchin larvae would have fared
........
wrightme43
Well-known member
Cant stay long, must get back to shoveling sand from the field back into my tank. Do you think the crickets and hay will hurt anything?
Oh well, They have got to be better than That stupid bare bottom, all this time I just thought my fish were sleeping....
Oh well, They have got to be better than That stupid bare bottom, all this time I just thought my fish were sleeping....
Scooterman
Well-known member
I don't understand how they can make a statement like that about BB tanks. Did you notice the disclaimer at the end LOL!
wrightme43
Well-known member
Ok I went and got some more sand. I have had to dig out little depressions for the fish to lay in. I now have a 20" sand bed. It really pretty cool. It looks like sand art. LOL
Beeba
Well-known member
wrightme43 ... you crack me up
wow...I don't know where to start with this one...
I read through this article...I'm speechless, all I can say is it is laughable...a joke.
I'm about to go on a long tirade... Before I start, Id like to qualify some things...In my years in college, I took numerous classes on statistics and research...and in my job I deal almost daily with statistics and statistical models in our materials lab...I'm no chemist or biologist, but on the subject of statistics, I'd consider myself "above average"
Ok...this "study" of theirs is laughable,from a statistical viewpoint it is so flawed, biased, and improperly conducted as to be a joke. Lord, I hardly know where to start with this one. From this study, one can draw almost no absolutley statistically significant conclusions whatsoever.
I'll start with their substrate selection, using the coarse substrate over the plenum as a specific example. They said it has a "mean" value of 2mm. "mean" value is the total sum of the indivdual observations divided by the total number of observations....the more familiar term for this is "average". The mean value of the particle size tells you very little about the substrate as a whole. Mean values are useless as a descriptor of the entirety of the substrate unless the other 4 main statistical values are described as well...median, mode, range and standard deviation.
Say you analized each individual grain in a sample of sand and you made a list of all of them according to diameter. The "mean" value would be the average diameter of the sand. the "mode" would be the most commoly occuring value in the list "median" would be the value that occupies the center of the list, at equal distance from the extremes of the range. The "range" is the difference from the smallest diameter grain to the largest one. The "standard deviation" would be the average amount to which each individual grain differs from the mean value. Once all these values are accounted for, one can get a much better idea of the actual gradation of the substrate as a whole.
Ok..they had a coarse substate with a "mean" value of 2mm...well, there are a lot of different gradations of substrates that can have a mean value of 2mm....did the substrate have a range of 0.1-4mm with a median and mode of 2mm? or did the substrate have a range of .01mm-10mm with a median and mode of 1mm? both subsrates are very different in overall paritcle size, but could easily have a mean value of 2mm....and since the differences in the two substrates can mean large differences in the sand vs. water ratio in the overall volume of the substrate, consequental observations can be greatly affected...
Ok, let's assume they have a nice smooth sand gradation on a standard bell curve...the range is 0.1-4 mm, the median is 2mm and the mode is around 2mm give or take...If you read Jaubert, Goemans, Gamble and other plenum advocates, they all stress the importance of using a substrate no finer than 2-4mm, it's critical to plenum performance. This substrate range, projected onto a similar bell curve would have a median and mode of around 3mm give or take. So you can see, there is substantial difference in the substrate...the one they used has a mean value that is only the same as the beginning of the actual range of the recommended substrate over a plenum...bottom line, they didn't even set up their plenum as recommended, thier most coarse substrate over it is still too fine. This in and of itself is enough to invalidte any of their conclusions reguarding the plenum system. (but it still goes MUCH deeper than this... ) In statistics we have a term called "researcher bias", where the researcher, either knowingly or subconsiously, manipulates conditions in the experiment to skew the results in one direction or another....could be some of that going here? Didn't even set the plenum up as recommended... :doubt:
Ok...on to the livestock an other things....aside from the fact that their sampling was pathetically small (I know, the logistics of carrying out a larger experiment are huge), it appeared to lack any control at all. Were the livestock used in the study randomly selected from a larger population gathered in the same manor and QT'd to ensure basically uniform health? Same with the rock, was it selected randomly from a larger, similarly handled and QT'd popultion? It doesn't appear so...so couple this with the small sampling size, and you quickly have a large amount of mathematical bias....which will totally skew the outcome of the experiment.
Another HUGE flaw in the methodology here was the total lack of a control, or "placebo" group. A placebo group is an equal random draw from the same population that is NOT exposesd to the stimuli being tested, as a basis of comparison. And it's absolutely critical to have a placebo group if one wishes to prove statistical significance of one's findings. They simply didn't have a control group...you all who have pointed out that there was no BB tanks in the study are right on the money, that would have been the placebo group.
Next was their paragraph on the statistical analysis....I read, re-read and re-read again that nasty train wreck...here is what I gathered from it...they are using various statistical models to determine statistical significance and demonstrate the validity of their actual observations. Basically, what this means is that the differences they are observing in these tanks are not due to a randomness in testing proceedures and methods to any statistically significant degree. (I'm going to run this article by my old stats professor at the college and see if he agrees with my intrupretaion of thier statistical analysis.) What can one infer from this? Their methodology of testing the various parameters are reliable, and the results are due to condtions in the individual tanks and not randomness in the tests themselves, that's about it.
Bottom line...this experiment is so fundamentally flawed that one really can't draw any kind of valid conclusion from it, other than the test results of their observations on the first place, and then only as they apply to that very tank they were testing... Does it prove that one substrate is better than another in reguards to mortality? ABSOLUTELY NOT!!!!
MikeS
I read through this article...I'm speechless, all I can say is it is laughable...a joke.
I'm about to go on a long tirade...
Before I start, Id like to qualify some things...In my years in college, I took numerous classes on statistics and research...and in my job I deal almost daily with statistics and statistical models in our materials lab...I'm no chemist or biologist, but on the subject of statistics, I'd consider myself "above average" Ok...this "study" of theirs is laughable,from a statistical viewpoint it is so flawed, biased, and improperly conducted as to be a joke. Lord, I hardly know where to start with this one. From this study, one can draw almost no absolutley statistically significant conclusions whatsoever.
I'll start with their substrate selection, using the coarse substrate over the plenum as a specific example. They said it has a "mean" value of 2mm. "mean" value is the total sum of the indivdual observations divided by the total number of observations....the more familiar term for this is "average". The mean value of the particle size tells you very little about the substrate as a whole. Mean values are useless as a descriptor of the entirety of the substrate unless the other 4 main statistical values are described as well...median, mode, range and standard deviation.
Say you analized each individual grain in a sample of sand and you made a list of all of them according to diameter. The "mean" value would be the average diameter of the sand. the "mode" would be the most commoly occuring value in the list "median" would be the value that occupies the center of the list, at equal distance from the extremes of the range. The "range" is the difference from the smallest diameter grain to the largest one. The "standard deviation" would be the average amount to which each individual grain differs from the mean value. Once all these values are accounted for, one can get a much better idea of the actual gradation of the substrate as a whole.
Ok..they had a coarse substate with a "mean" value of 2mm...well, there are a lot of different gradations of substrates that can have a mean value of 2mm....did the substrate have a range of 0.1-4mm with a median and mode of 2mm? or did the substrate have a range of .01mm-10mm with a median and mode of 1mm? both subsrates are very different in overall paritcle size, but could easily have a mean value of 2mm....and since the differences in the two substrates can mean large differences in the sand vs. water ratio in the overall volume of the substrate, consequental observations can be greatly affected...
Ok, let's assume they have a nice smooth sand gradation on a standard bell curve...the range is 0.1-4 mm, the median is 2mm and the mode is around 2mm give or take...If you read Jaubert, Goemans, Gamble and other plenum advocates, they all stress the importance of using a substrate no finer than 2-4mm, it's critical to plenum performance. This substrate range, projected onto a similar bell curve would have a median and mode of around 3mm give or take. So you can see, there is substantial difference in the substrate...the one they used has a mean value that is only the same as the beginning of the actual range of the recommended substrate over a plenum...bottom line, they didn't even set up their plenum as recommended, thier most coarse substrate over it is still too fine. This in and of itself is enough to invalidte any of their conclusions reguarding the plenum system. (but it still goes MUCH deeper than this...
) In statistics we have a term called "researcher bias", where the researcher, either knowingly or subconsiously, manipulates conditions in the experiment to skew the results in one direction or another....could be some of that going here? Didn't even set the plenum up as recommended... :doubt: Ok...on to the livestock an other things....aside from the fact that their sampling was pathetically small (I know, the logistics of carrying out a larger experiment are huge), it appeared to lack any control at all. Were the livestock used in the study randomly selected from a larger population gathered in the same manor and QT'd to ensure basically uniform health? Same with the rock, was it selected randomly from a larger, similarly handled and QT'd popultion? It doesn't appear so...so couple this with the small sampling size, and you quickly have a large amount of mathematical bias....which will totally skew the outcome of the experiment.
Another HUGE flaw in the methodology here was the total lack of a control, or "placebo" group. A placebo group is an equal random draw from the same population that is NOT exposesd to the stimuli being tested, as a basis of comparison. And it's absolutely critical to have a placebo group if one wishes to prove statistical significance of one's findings. They simply didn't have a control group...you all who have pointed out that there was no BB tanks in the study are right on the money, that would have been the placebo group.
Next was their paragraph on the statistical analysis....I read, re-read and re-read again that nasty train wreck...here is what I gathered from it...they are using various statistical models to determine statistical significance and demonstrate the validity of their actual observations. Basically, what this means is that the differences they are observing in these tanks are not due to a randomness in testing proceedures and methods to any statistically significant degree. (I'm going to run this article by my old stats professor at the college and see if he agrees with my intrupretaion of thier statistical analysis.) What can one infer from this? Their methodology of testing the various parameters are reliable, and the results are due to condtions in the individual tanks and not randomness in the tests themselves, that's about it.
Bottom line...this experiment is so fundamentally flawed that one really can't draw any kind of valid conclusion from it, other than the test results of their observations on the first place, and then only as they apply to that very tank they were testing...
Does it prove that one substrate is better than another in reguards to mortality? ABSOLUTELY NOT!!!! MikeS
chris&barb
Reef-Geeks
Brain freeze
wrightme43
Well-known member
They proved one thing to me. Dont put puffers in tanks that have huge varibility in temprature. LOL
It just seems the level of viable information being put out these days is brutal. Imagine new folks loking at a study like that from a man with a BA?? You start putting all these articles together and it really makes you shake your head as to what some folks are trying to prove.
MIke
MIke
mojoreef said:
Exacly Mike...from this study one can actually reliably conlude very little, even though it has all kinds of neat statistical data...sure, I have no doubt that the results they saw from tank to tank were statistically significant based on models, but to draw a conclusion and support a thesis from these observations is difficult due to the almost total lack of experimental control, the high degree of mathematicial bias, ect...this is the danger of statistics...
MikeS
chris&barb
Reef-Geeks
so is everyone here BB or do you all just think this study is crap?im not being smart at all i just want to know.i understand the principal of BB but i just dont like the looks of it.isn't there a certain amount of maintainance in both?like sucking detritus off of glass or from sand.if you never did either would you have the same result?
clayswim
Well-known member
Not all of us are BB; and yes, we seem to agree the study is crap.
Personally I use a shallow layer of sand because I like the looks of it and it gives something for my bottom critters to pick through. If not for those reasons I would go BB just to keep things detritus-free.
As to the study, my brain hurts even thinking of where to begin; so I'm not going to. That's one of the worst thought out studies I think I've ever read. So many variables were ignored that it makes you wonder if there was any planning at all. I do have to give them credit for one thing though; they made an excellent presentation of worthless figures and statistics.
Clayton
Personally I use a shallow layer of sand because I like the looks of it and it gives something for my bottom critters to pick through. If not for those reasons I would go BB just to keep things detritus-free.
As to the study, my brain hurts even thinking of where to begin; so I'm not going to.
That's one of the worst thought out studies I think I've ever read. So many variables were ignored that it makes you wonder if there was any planning at all. I do have to give them credit for one thing though; they made an excellent presentation of worthless figures and statistics.Clayton
