Forums
New posts
Search forums
What's new
Articles
Latest reviews
Author list
Classifieds
Log in
Register
What's new
Search
Search
Search titles and first posts only
Search titles only
By:
New posts
Search forums
Menu
Log in
Register
Install the app
Install
Forums
Hunting
Long Range Hunting & Shooting
Lead poisoning from eating game shot with lead core bullets?
JavaScript is disabled. For a better experience, please enable JavaScript in your browser before proceeding.
You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an
alternative browser
.
Reply to thread
Message
<blockquote data-quote="DNADave" data-source="post: 1925969" data-attributes="member: 110467"><p>Nice summary.</p><p></p><p>As a statistician (technically, a statistical geneticist), I think a lot of people do not understand what it is we do. At its heart, statistics is about making decisions and assessing the risk in making a specific decision (at least for inferential statistics, prediction is different). When you set a null hypothesis, you also set an alternative hypothesis. This alternative hypothesis is the one that is aligned with your "preconceived idea", or as we call it in research, your research hypothesis. We always have a theory of what is going on, sometimes precise and sometimes less so.</p><p></p><p>Once you have your null and alternative hypothesis set, you need to design your experiment. Ideally, you create a design that accounts for other factors that could influence your study (in public health, this could be age, weight, gender, smoking status, etc.) and balance those across all of the levels of the groups you want to test (such as new drug or standard of care in a drug study). This design leaves just the effect of interest as the thing to test with your hypothesis. The statistical analysis boils down to assessing how extreme your obtained results are out of random chance. Here's where the making decisions and assessing the risk of that decision comes in: my p-value that I calculate from my test is meant to tell me how likely am I to see a result as extreme as I saw due to random chance. if the null hypothesis is true The reason we use 0.05 as our significance level is because that relates to a 1 in 20 chance of something occurring due to random chance under the assumption of the null hypothesis being true.</p><p></p><p>So, if I perform my experiment and calculate my p-value as 0.04, I do the following since my p-value is less than 0.05. I declare that I have decided to accept the alternative hypothesis (and thus reject the null hypothesis) this is my decision. My risk is that I might observe something as extreme as what I saw 1 out of 25 times if the null hypothesis was actually true, so I feel relatively confident that what I saw is real and that the alternative hypothesis is actually true.</p><p></p><p>Sometimes, we cannot conduct a designed experiment and have to conduct what is called an observational study. This is where we have access to data that is generally already collected. We can reconstruct a design to balance factors by taking selected subsets of this data (this has some degree of bias since you are selecting data to use). Alternatively, we can use all of the data, but this also has bias because certain nuisance factors will not be balanced and have unequal influence on the outcome. Drug studies are an interesting blend of design and observational in that we have a design, but accumulate patients over time to fill in that design.</p><p></p><p>Sorry to geek out on you all, but not all of us scientists lack scruples. The vast majority of us try very hard to remain objective in our research and are constantly trying to improve upon the ways we can maintain objectivity while still making the most out of the data we have since data is expensive.</p></blockquote><p></p>
[QUOTE="DNADave, post: 1925969, member: 110467"] Nice summary. As a statistician (technically, a statistical geneticist), I think a lot of people do not understand what it is we do. At its heart, statistics is about making decisions and assessing the risk in making a specific decision (at least for inferential statistics, prediction is different). When you set a null hypothesis, you also set an alternative hypothesis. This alternative hypothesis is the one that is aligned with your "preconceived idea", or as we call it in research, your research hypothesis. We always have a theory of what is going on, sometimes precise and sometimes less so. Once you have your null and alternative hypothesis set, you need to design your experiment. Ideally, you create a design that accounts for other factors that could influence your study (in public health, this could be age, weight, gender, smoking status, etc.) and balance those across all of the levels of the groups you want to test (such as new drug or standard of care in a drug study). This design leaves just the effect of interest as the thing to test with your hypothesis. The statistical analysis boils down to assessing how extreme your obtained results are out of random chance. Here's where the making decisions and assessing the risk of that decision comes in: my p-value that I calculate from my test is meant to tell me how likely am I to see a result as extreme as I saw due to random chance. if the null hypothesis is true The reason we use 0.05 as our significance level is because that relates to a 1 in 20 chance of something occurring due to random chance under the assumption of the null hypothesis being true. So, if I perform my experiment and calculate my p-value as 0.04, I do the following since my p-value is less than 0.05. I declare that I have decided to accept the alternative hypothesis (and thus reject the null hypothesis) this is my decision. My risk is that I might observe something as extreme as what I saw 1 out of 25 times if the null hypothesis was actually true, so I feel relatively confident that what I saw is real and that the alternative hypothesis is actually true. Sometimes, we cannot conduct a designed experiment and have to conduct what is called an observational study. This is where we have access to data that is generally already collected. We can reconstruct a design to balance factors by taking selected subsets of this data (this has some degree of bias since you are selecting data to use). Alternatively, we can use all of the data, but this also has bias because certain nuisance factors will not be balanced and have unequal influence on the outcome. Drug studies are an interesting blend of design and observational in that we have a design, but accumulate patients over time to fill in that design. Sorry to geek out on you all, but not all of us scientists lack scruples. The vast majority of us try very hard to remain objective in our research and are constantly trying to improve upon the ways we can maintain objectivity while still making the most out of the data we have since data is expensive. [/QUOTE]
Insert quotes…
Verification
Post reply
Forums
Hunting
Long Range Hunting & Shooting
Lead poisoning from eating game shot with lead core bullets?
Top