|
|
Post by fullthrottlehunter on Jul 23, 2016 16:27:43 GMT -5
FTH, will you at least agree that all of the MAR's surveys are invalid because they are far from random and unbiased? I say this because the prerequisite of receiving one is that you had to return your Annual Deer Harvest Survey. The MDNR-WD proved that hunters who belong to hunting organizations are far more apt to return surveys and support MAR's regulations when they do. (2001 Hunter attitudes on QDM by Peter Bull and Dr. Ben Peyton.) No, I don't agree with that at all. You cannot predict what a hunter's opinion is based on the fact that he returned a survey. The only thing you can predict is that they will be more likely to also return the next survey. That's it. Any other assumption is pure conjecture. I'd also like to make another point as to my statistical example in a previous post. You'll notice that the greater the number of returns, the more accurate. That was needed due to the very small sample size of 10. It is a proven fact that when sampling/surveying a low number of people, you need a higher percentage of that sample size to be accurate. As the sample size increases, the percentage of the sample size needed to be accurate decreases. |
|
|
|
Post by Dale Malusi on Jul 23, 2016 22:28:09 GMT -5
We have already covered what was changed. "Estimates of harvest, hunting effort, and hunter numbers were reduced by 6.4%, 4.7%, and 1.9%, respectively...." Please, explain why this matters. I've gone one step further and showed what a 4.7% reduction means in regards to hunter effort (days). With nearly 2 decades of statistical sampling to compare to, I'm confident they understand why and how the change needed to be made. The problem here is a few guys that don't understand statistics cry fowl and proclaim manipulation. It's a perception problem, not a manipulation one. Please explain why it doesn't matter. It seems the shoe is on the other foot. Why not just publish raw numbers? I guess that if you estimated what was likely to probably occur if the respective estimates of the likelihood of the probable statistical estimation calculations put forth to guess what was or were possible that may or may not have been, you may come up with any perceptional conclusion you like. |
|
|
|
Post by fullthrottlehunter on Jul 24, 2016 5:28:41 GMT -5
We have already covered what was changed. "Estimates of harvest, hunting effort, and hunter numbers were reduced by 6.4%, 4.7%, and 1.9%, respectively...." Please, explain why this matters. I've gone one step further and showed what a 4.7% reduction means in regards to hunter effort (days). With nearly 2 decades of statistical sampling to compare to, I'm confident they understand why and how the change needed to be made. The problem here is a few guys that don't understand statistics cry fowl and proclaim manipulation. It's a perception problem, not a manipulation one. Please explain why it doesn't matter. It seems the shoe is on the other foot. Why not just publish raw numbers? I guess that if you estimated what was likely to probably occur if the respective estimates of the likelihood of the probable statistical estimation calculations put forth to guess what was or were possible that may or may not have been, you may come up with any perceptional conclusion you like. Shoe is on the other foot? Yea, dont think so. What are the raw numbers going to tell you? Specifically, how are the raw numbers going to tell you total harvest, total hunter numbers, and total hunter effort for an entire population of hunters based solely on the raw numbers from ~28,000 of them? I'll help you. It wont. Simply reporting those raw numbers alone will not do a thing for you besides tell you what those who returned a survey did, and that's it. What does that help you or anyone? That's where math and statistics help you predict or estimate what the entire deer hunting population did for that year based upon those who returned their survey. It's not a simple guess. It's not corrupt manipulation or a conspiracy. It's called science using time proven statistical survey methods and sampling techniques. The same methods and techniques used by our DNR for decades, and other scientists for far longer than that. |
|
|
|
Post by Dale Malusi on Jul 24, 2016 12:56:00 GMT -5
Prediction, estimate, likely, probably, are all guesses.
"That's where math and statistics help you predict or estimate what the entire deer hunting population did for that year based upon those who returned their survey. It's not a simple guess."
They should just base the guess on the replies they received. It's that simple.
Instead of digging a bigger hole every year, just be honest. It's that simple.
Simple Definition of honest
: not hiding the truth about someone or something : not meant to deceive someone
Simple definition of hide
: to screen from or as if from view : obscure
Simple Definition of obscure
: difficult or impossible to know completely and with certainty
Yeah, the other foot.
|
|
|
|
Post by fullthrottlehunter on Jul 24, 2016 16:04:57 GMT -5
1. It's not a guess.
2. They do exactly that. Estimate based upon responses that they receive.
I'm sorry Dale, but you're way off base on this. It's called science. That's what science is. The fact that some don't understand it, in no way shows there is anything manipulative going on. As I said before, the Mother Mary could publish this survey and a few of you still wouldn't believe it.
It's not a statistical problem or manipulation, it's a perception problem due to lack of understanding of how statistical sampling works.
|
|
|
|
Post by Dale Malusi on Jul 24, 2016 22:56:06 GMT -5
No lack of understanding here.
It's for the readers.
You are off base.
You must listen to social science.
|
|
|
|
Post by fullthrottlehunter on Jul 25, 2016 1:09:19 GMT -5
No lack of understanding here. Obviously there is, or you wouldn't have asked why they don't just publish the raw numbers. PS....The 2015 Michigan Deer Harvest Survey is now out. Enjoy. |
|
|
|
Post by ridge on Jul 25, 2016 16:55:55 GMT -5
I have been saying this for several years concerning the MARs surveys and I get back is that a small sample is adequate for a good confidence level. Either it is or it isn't. One can not have it both ways.
quote from fullthrottlehunter:
I'd also like to make another point as to my statistical example in a previous post. You'll notice that the greater the number of returns, the more accurate. That was needed due to the very small sample size of 10. It is a proven fact that when sampling/surveying a low number of people, you need a higher percentage of that sample size to be accurate. As the sample size increases, the percentage of the sample size needed to be accurate decreases.
The last two sentences are contradictory and IMO are not good statistical practice. I do understand that is a common practice for some organizations but it is not for all. All stratas or types of participants are required to be accurate according to MIT and Harvard, who did the groundwork for social or population statistical analysis. The larger the sample the greater the probability becomes that all of the stratas have been included unless the survey/instrument has been distributed in such a manner that it is not probable despite having a large sample. The confidence level can become an illusion if all groups within the target population are not included. Cost and time can be factors in the distribution of a survey but they can also greatly mitigate the value of a survey.
Incomplete data in = false conclusions with little to no value
|
|
|
|
Post by fullthrottlehunter on Jul 25, 2016 18:51:20 GMT -5
No Ridge, they are not contradictory. That is how statistical sampling and surveys work in order to achieve the desired confidence levels or margin of error.
In my example of 10 hunters, a higher percentage was needed to maintain that confidence level. If I were to survey 100,000 people, I would not need to survey that high of a percentage to maintain the same confidence level. If you survey 1,000,000 people, the percentage drops even lower.
I'm sorry, but that is a fact. Look at political polling for instance. They're polling 4-5 thousand people out of millions and can maintain a margin of error of +/- 5%.
Regarding various strata. They are included in the survey. There is no way of thinking otherwise. Those receiving a Harvest Survey are selected randomly. Furthermore, they are also careful in selecting responses from those who voluntarily reported online due to the possibility of bias. Online surveys are not random, but are selected randomly to prevent that bias, although it still may exist to a point.
Multiple Wildlife Agencies have said that Michigan's data collection is far superior to many others. Again, this isn't a manipulation problem. It's a perception problem brought on by some that don't understand how statistical sampling works.
|
|
|
|
Post by ridge on Jul 26, 2016 12:59:50 GMT -5
The harvest surveys that are sent out are random as far as I know. I will agree to that. It is not their random element that I question but the possibility that the trending percentages of the estimates might have changed. The greater the survey bias, IMO, (by having substantially fewer responses), the greater the probability exists that there might be an error in the estimated results. This is a possibility even though the years of past surveys may be long. Notable changes do happen. To disregard them can have a marked effect on the validity of the estimated results.
The percentages can drop because the actual number surveyed increases. 10% of 100,000=10,000 5% of 1,000,000=50,000
In that example which I did not check for any confidence level, the actual number increased by 40,000 or by 5X while the percentage decreased by 5%
If all of the strata are included and the random element is there, the confidence level and validity is normally high.
However elements such as the strata are not always correct which cause results such as in political surveys to be highly variant.
I do believe that the Harvest Surveys when there are adequate responses have a high probability of possessing good confidence levels. The problems enter in when the response levels, despite the math trending, are very low such as the last two years. I believe the DNR should have qualified that and made it more clear than they did. There may be real important reasons why the response level is low besides apathy. If it was just apathy, I would not be so concerned but that is really not known but circumstances can often provide reasons for the low response level. I suppose one could label those as educated guesses. Those reasons not due to apathy could be important to program direction and content.
When discussing surveys in Michigan, a problem often occurs because one person is discussing the Annual Harvest Survey and the other person is discussing some particular MARs survey. The methods for choosing participants are different and the target population is different.
IMO, the check station data would be more effective if it were done as a separate report rather than as an element of a MARs trial period or of the Annual Harvest Survey unless collections practices become more standard geographically and unless trial vs. not trial data is compared. It also should not include exterior results such as that collected from various meetings and banquets. That speaks to a question created in consistency. Check station data are collected in an even more voluntary manner (than the Annual Harvest or MARs surveys) and thus are not strictly surveys but they can show trends if they are not unduly influenced and they are consistent in collection practices. Organizational data such as Boone and Crocket or MBH or that of hunting contests can be illustrated, published but it should be done in a separate manner as individual categories. That way the bias of individual collection methods and participants can be noted.
Cost and time are important factors in establishing collection and data processing in a survey but they can also have very negative effects on those instruments and that should not be denied but objectively noted as part of the report.
|
|
|
|
Post by ridge on Jul 26, 2016 15:00:41 GMT -5
A check between The Annual Harvest Survey Report and the Check Station Data Report this year illustrates some interesting differences.
The Check Station Data said that the LP harvest was up 17%. The Check Station Data said that the UP Harvest was down by 20%.
The Annual Harvest Report said that the LP harvest was up 5.2%. The Annual Harvest Report said that the UP harvest was down by 49%.
The MARs trial information most often uses the check station data. So which one presents the best picture? Which estimate should be trusted or should either one be trusted? We are told that Michigan issues some of the best results in the country. Which results are they describing? As far as number harvested estimates, are the check station data that accurate? Should those number harvested estimates by the check station data even be used? My personal confidence level is not very high.
|
|
|
|
Post by fullthrottlehunter on Jul 26, 2016 18:41:32 GMT -5
I've grown tired of trying to explain how this survey works to the readers here. You've just spent nearly a full paragraph explaining the same thing that I did, proving my point, but don't even realize it. Think what you wish.
|
|
|
|
Post by ridge on Jul 28, 2016 12:09:31 GMT -5
fullthrottlehunter, I was not disagreeing with your explanation of the system so much as I was saying that the DNR makes the assumption that the rate of trending always stays the same. It can but that does mean it does. That is why I think the raw data should be shared separately. I was not saying that they should not provide their estimates as a report. As to manipulations, in one of my previous posts, I stated that I did not know whether that happened or not.
Part of my discussion above was about the MARs report in contrast to the Harvest Survey (and one was in contrast with the Check Station data). If I confused the subject or thread, please forgive me but I thought that it was important. My point was that the attempt was made in the Harvest Survey by the large number of surveys sent out that all of the strata could be included. We don't know that by the low number of surveys returned in the Harvest Survey in the last two years. My point was also that the MARs surveys lacked the attempt to provide the random factor or to provide any attempt to include all of the strata in the target group. I also stated my belief as to why the differences could have happened, time, cost, and possibly bias. My first attempt to contrast the Harvest Survey with the MARs Surveys confused the matter, I think.
Yes, we will most likely disagree for the present time. You believe, I think, that the DNR is always 100% accurate and without bias in their surveys and survey process while I believe that there are factors which may cause accuracy problems and bias. Also it was and is my contention that manipulation could happen while you believe that there is no possibility of that ever happening.
Time might illustrate who is correct in this matter. But our government can do many things to protect the illusion of transparency. It is my contention that the citizens always need to work to see that honesty is essential in the behavior of government.
|
|
|
|
Post by Dale Malusi on Jul 28, 2016 22:14:53 GMT -5
I've grown tired of trying to explain how this survey works to the readers here. You've just spent nearly a full paragraph explaining the same thing that I did, proving my point, but don't even realize it. Think what you wish. You are off base.Your vieled insults are getting old. And you call me ignorant. |
|
|
|
Post by fullthrottlehunter on Jul 29, 2016 3:19:59 GMT -5
fullthrottlehunter, I was not disagreeing with your explanation of the system so much as I was saying that the DNR makes the assumption that the rate of trending always stays the same. It can but that does mean it does. That is why I think the raw data should be shared separately. I was not saying that they should not provide their estimates as a report. As to manipulations, in one of my previous posts, I stated that I did not know whether that happened or not. No worries Ridge. We can disagree and remain civil. With that said, the raw data isn't going to help you discern anything. The only thing it will help you with is possibly a comparison of those who returned a survey from year to year. It's not going to help you to determine total hunter numbers, deer harvest, or hunter effort for all deer hunters as they do now. The latter is the entire goal of the annual Harvest Survey. Further, it has been proposed on here that what they do is just a guess. It's far from it. Part of my discussion above was about the MARs report in contrast to the Harvest Survey (and one was in contrast with the Check Station data). If I confused the subject or thread, please forgive me but I thought that it was important. My point was that the attempt was made in the Harvest Survey by the large number of surveys sent out that all of the strata could be included. We don't know that by the low number of surveys returned in the Harvest Survey in the last two years. My point was also that the MARs surveys lacked the attempt to provide the random factor or to provide any attempt to include all of the strata in the target group. I also stated my belief as to why the differences could have happened, time, cost, and possibly bias. My first attempt to contrast the Harvest Survey with the MARs Surveys confused the matter, I think. It's a fact that there is no way of knowing how anyone is going to "vote" based on the single fact that they returned a harvest survey. Harvest Surveys are completely random and include all strata. MARS surveys are sent to those who were chosen randomly, not only once but now twice, and who have stated they hunted in the proposed area. It's important to survey those effected by the regulation. Without knowing where they hunted, as stated when returning their random Harvest Report, it is impossible to ascertain where they primarily hunt. That becomes important because it's a fact that people can/do answer the survey questions differently when it effects them personally, as opposed to a general question. It's a statistically solid and mathematically accurate form of survey. I have no problem questioning the survey, even if the outcome is not in my favor. I'm well aware that they could repeat the survey over and over and end up with similar results. Yes, we will most likely disagree for the present time. You believe, I think, that the DNR is always 100% accurate and without bias in their surveys and survey process while I believe that there are factors which may cause accuracy problems and bias. Also it was and is my contention that manipulation could happen while you believe that there is no possibility of that ever happening. Time might illustrate who is correct in this matter. But our government can do many things to protect the illusion of transparency. It is my contention that the citizens always need to work to see that honesty is essential in the behavior of government. Yes, we will continue to disagree. The DNR has absolutely nothing to gain by intentionally manipulating surveys or the Deer Harvest Report. They also have nothing to gain by manipulating MARS surveys or the results obtained from them. Nothing. |
|
|
|
Post by ridge on Jul 29, 2016 13:03:39 GMT -5
Yes, we will continue to disagree. The DNR has absolutely nothing to gain by intentionally manipulating surveys or the Deer Harvest Report. They also have nothing to gain by manipulating MARS surveys or the results obtained from them. Nothing. (quote by fullthrottlehunter)
The above quote is only true if there is no one in leadership or if there is no one with access to the survey and its results that has bias or a previous agenda, who influences the survey or its results. Our group and myself believe that is a huge if and perhaps an impossible one. Words and actions provide more evidence than a statement issued by the Department.
However I will acknowledge that there is always a very small possibility that the quote is true. There is, too, that possibility that the mechanics of a survey can be flawed such as with language or distribution. Population statistical surveys are a complex operation if done correctly therefore any number of factors, intentionally or not, can provide a flaw that affects the value of the operation. That is why basing policy on a social survey may not provide a direction that is scientifically based nor which is a positive direction for the deer herd or for deer hunting.
|
|
|
|
Post by ridge on Jul 29, 2016 13:27:59 GMT -5
I've grown tired of trying to explain how this survey works to the readers here. You've just spent nearly a full paragraph explaining the same thing that I did, proving my point, but don't even realize it. Think what you wish. You are off base.Your vieled insults are getting old. And you call me ignorant. fullthrottlehunter's post was directed primarily at my post. He did not call me ignorant nor did he call anyone else ignorant. Doehead, you are a very intelligent person and a passionate hunter/person. Please do not take the content of a post personally unless your name is used and the content actually is a slur against you (or someone else). Posts like your reply above will be removed in the future if it is not appropriate. We can debate content forever without demeaning a person. Believe me, if someone is demeaning, I will remove the post or I will ban the person if the post is particularly crude or cruelly offensive. Sometimes a past interaction or feeling can cause a person to react negatively or aggressively when that name is heard or written (or when the person is present). I have been guilty of that. Progress can only be made if we move on/away from the past and if we concentrate on the issues or on the dialog of the present. That is not easy for me but it is necessary if I am going to contribute in a positive manner. |
|
|
|
Post by fullthrottlehunter on Jul 29, 2016 17:47:07 GMT -5
Yes, we will continue to disagree. The DNR has absolutely nothing to gain by intentionally manipulating surveys or the Deer Harvest Report. They also have nothing to gain by manipulating MARS surveys or the results obtained from them. Nothing. (quote by fullthrottlehunter) The above quote is only true if there is no one in leadership or if there is no one with access to the survey and its results that has bias or a previous agenda, who influences the survey or its results. Our group and myself believe that is a huge if and perhaps an impossible one. Words and actions provide more evidence than a statement issued by the Department. However I will acknowledge that there is always a very small possibility that the quote is true. There is, too, that possibility that the mechanics of a survey can be flawed such as with language or distribution. Population statistical surveys are a complex operation if done correctly therefore any number of factors, intentionally or not, can provide a flaw that affects the value of the operation. That is why basing policy on a social survey may not provide a direction that is scientifically based nor which is a positive direction for the deer herd or for deer hunting. You could repeat the survey and end up with basically the same results 95 out of 100 times. I'm not concerned in the least with any manipulation. These are professional people. They aren't going to risk their reputation, the Department's reputation, or their careers by manipulating a survey. They have nothing to gain but everything to lose. |
|
|
|
Post by fullthrottlehunter on Jul 29, 2016 17:56:15 GMT -5
fullthrottlehunter's post was directed primarily at my post. He did not call me ignorant nor did he call anyone else ignorant. Correct. It wasn't an attack or an insult. You and I were having a discussion. Nothing more. |
|
|
|
Post by Dale Malusi on Jul 29, 2016 21:04:36 GMT -5
You are off base.Your vieled insults are getting old. And you call me ignorant. fullthrottlehunter's post was directed primarily at my post. He did not call me ignorant nor did he call anyone else ignorant. Doehead, you are a very intelligent person and a passionate hunter/person. Please do not take the content of a post personally unless your name is used and the content actually is a slur against you (or someone else). Posts like your reply above will be removed in the future if it is not appropriate. We can debate content forever without demeaning a person. Believe me, if someone is demeaning, I will remove the post or I will ban the person if the post is particularly crude or cruelly offensive. Sometimes a past interaction or feeling can cause a person to react negatively or aggressively when that name is heard or written (or when the person is present). I have been guilty of that. Progress can only be made if we move on/away from the past and if we concentrate on the issues or on the dialog of the present. That is not easy for me but it is necessary if I am going to contribute in a positive manner. I'm sorry Dennis , but I stand by my post. Read what I quoted. He said "readers', not Ridge. I, as a reader, took umbrage at his rudeness. You guys should really start a new thread about statistical analysis. |
|
