- Видео 50
- Просмотров 262 955
Biometry Online Lessons
Добавлен 15 авг 2013
This channel contains a series of mini-lectures to accompany Biology 302/593H (Biometry) in the Department of Biology, West Virginia University. The mini-lectures are public, but are intended for students enrolled in Biometry, using Sokal and Rohlf's Biometry as accompanying text, and using SAS JMP for in-class problem-solving exercises. The goal of the course is to guide a student with a one-semester introductory statistics course background to the level of comfort and confidence in performing 80% of the statistics encountered in biological research problems using SAS JMP.
Lesson50 Loglikelihood
Nominal X, Nominal Y. Chi-square, right? Nah. Loglikelihood, aka G-test, is the better approach. In this mini-lesson, I show you the nitty gritty of an extremely simple G-test.
Просмотров: 7 645
Видео
Lesson49 Logistic Regression
Просмотров 4,2 тыс.10 лет назад
In biology, a lot of interesting dependent variables are nominal (e.g., survived/died, got cancer/didn't get cancer, reproduced/didn't reproduce, became diseased/did not become diseased, etc). To regress such a variable on a continuous X variable, we use logistic regression. In this lesson, I show you how logistic regression is done, and how to obtain the function relation prob(outcome a) to X ...
Lesson47 Akaike Information Criterion
Просмотров 90 тыс.10 лет назад
With the possibilities opened up by linear and multiple forms of nonlinear regression, not to mention multiple regression, etc, how is the wise researcher supposed to choose between statistical models? It would be most helpful to have an objective criterion, wrote Hirogutu Akaike, back in ca 1974 in a paper entitled "A new look at the statistical model identification". Now cited more than 20,00...
Lesson46 Nonlinear regression in biology
Просмотров 1,3 тыс.10 лет назад
Lots of variables are not linear functions of one another; in this case we need to find a nonlinear fit. This can be done by least squares fitting of other functions, such as polynomials, or by transforming one or both variables and trying linear fits of the transformed variables, which, upon yield nonlinear relationships between the untransformed variables. I give examples of both approaches a...
Lesson45 Analysis of Covariance
Просмотров 20 тыс.10 лет назад
A logical extension of ANOVA and regression is ANCOVA, or Analysis of Covariance. Two general applications are described, one where we can attain power to detect treatment effects by including a continuous covariate in the model, and one where we are actually interested in the 2nd (or higher) order interaction between a continuous and nominal variable.
Lesson44 More regression and r square
Просмотров 81210 лет назад
In this lesson I describe the assumptions underlying least squares regression, then explain the concept of r-square.
Lesson43 Nitty Gritty of Linear Regression
Просмотров 63410 лет назад
You knew it was coming. I'm not happy that you just understand the purpose of regression, but you also need to calculate a regression coefficient - I'll make it easy with a KISS data set, but I want you to do this at least once yourself with a slightly harder KISS data set so you understand how these slopes are calculated...
Lesson42 The Correlation Coefficient
Просмотров 71210 лет назад
As with our other statistics, knowing what correlation is used for is not enough; we really need to know what calculations underlie our statistics programs to be able to understand them at a deeper level. This mini video shows you the nitty gritty of the calculation of r.
Lesson41 Purpose of Regression
Просмотров 1,2 тыс.10 лет назад
In this lesson I discuss the purpose of regression (contrasted with correlation) in which we postulate Y=f(X) with continuous independent and dependent variables.
Lesson40 Purpose of Correlation
Просмотров 90010 лет назад
Correlation and regression are two tools for examining the relationship between 2 (or more) continuous variables. In this lesson, I explain the difference between the two, and the importance of making the distinction.
Lesson39 Expected Mean Squares
Просмотров 6 тыс.10 лет назад
In designing an experiment, it's important to know that in the end you will be able, with the help of your friendly statistics software, to actually test the effects in your model that you care about. To determine this, you will need to know the 'expected mean squares' associated with each term in your model. This lesson shows how to calculate those with an example taken from the experiment des...
Lesson38 Determining Model Effects
Просмотров 83710 лет назад
The next step in planning an experiment is determining the 'effects' in the model dictated by your experimental design. Here, I review how to determine the list of effects you will add to your model.
Lesson37 Complex Designs Statistical Layout
Просмотров 89310 лет назад
After drawing the physical layout of a complex experiment, use that to depict the orthogonal and nested relationships among independent variables so that it is easy from that point to contruct model effects. An example is given with a reciprocal transplant experiment involving blocking.
Lesson36 Complex Designs Physical Layout
Просмотров 59810 лет назад
A diagram of the physical layout of an experiment can help you visualize the relationships among independent variables, but also you can often see before the experiment is done that a better arrangement might be possible so that you most efficiently test the terms of interest in your model. This is illustrated with the reciprocal transplant experiment described previously.
Lesson35 Complex Designs Conceptual Layout
Просмотров 69310 лет назад
A picture can be worth a thousand words in helping a reader to understand your experimental design. Many publications and grant proposals do not contain a conceptual layout of the experiment, but would be much clearer if they did. The approach is illustrated in this lesson with the conceptual layout of a reciprocal transplant experiment.
Lesson33 Nitty Gritty of Nested ANOVA
Просмотров 10 тыс.10 лет назад
Lesson33 Nitty Gritty of Nested ANOVA
Lesson31 Depicting Complex Experimental Designs
Просмотров 1 тыс.10 лет назад
Lesson31 Depicting Complex Experimental Designs
Lesson30 Two way Designs Without Replication
Просмотров 93910 лет назад
Lesson30 Two way Designs Without Replication
Lesson29 Confounding and Pseudoreplication
Просмотров 10 тыс.10 лет назад
Lesson29 Confounding and Pseudoreplication
Lesson100 ANOVA and a posteriori demo with JMP
Просмотров 50610 лет назад
Lesson100 ANOVA and a posteriori demo with JMP
Lesson27 Statement of Interaction Results
Просмотров 1,1 тыс.10 лет назад
Lesson27 Statement of Interaction Results
Lesson26 The Meaning of Statistical Interaction
Просмотров 10 тыс.10 лет назад
Lesson26 The Meaning of Statistical Interaction
Lesson25 2 Way 3 Way and N Way Designs
Просмотров 79110 лет назад
Lesson25 2 Way 3 Way and N Way Designs
Lesson24 The Mechanics of Planned Comparisons With a KISS
Просмотров 59010 лет назад
Lesson24 The Mechanics of Planned Comparisons With a KISS
Lesson23 Planned Comparisons and Orthogonality
Просмотров 2,4 тыс.10 лет назад
Lesson23 Planned Comparisons and Orthogonality
Lesson21 Guarding Against Type I and II Errors
Просмотров 55310 лет назад
Lesson21 Guarding Against Type I and II Errors
whatever your name is professor, thank you
Very Well. Precision
All videos by this professor are great. He calmly explains all subtle things and make it feel so interesting, intuitive and most importantly easy to understand and perform. Thank you so much.
Fantastic, thanks for the thorough and thoughtful explanation! I'll be back for more :)
Very clear lecture - I never really got this, but your explanation made it very clear. Thank you so much.
Note: there is a mistake in the calculation of AIC; should be 42.17 for the first model and 37.86 for the second. Sorry!
Great explanation.
Excellent video!!! Thanks!
Excellent presentation.....very simple and understanding.....now I understood what is ANCOVA
I liked it!
This is simply the best. Thanks Prof.
Great information
its easy to understand thank you very much
Thank you, this is really helpful. Well teached and explained
Than you for the video! Very helpful.
Thank you for sharing this beyond your class. It was extremely helpful!
Excelente. Thanks
Thank you so much!
Thanks!
Thanks!!!!
You have the best disposition and voice for instructional videos! Good job explaining ANCOVA - it was very helpful.
THE BEST EXPLANATION I HAVE EVER READ!
wonderfully explained. I always found it difficult to express intraction in a statement. Thanks a ton!
Your lessons are wonderful!
thank you!
Great info, great pace. I would have liked a bit more color on the nested factors with fixed effects. It seems like the example got cutoff without a full explanation of why this can't be.
Thank you!
Great video. Ty.
This is great!! Could you please answer a quation for me? Once you produce a good ANOVA modle with nested factors. can you then add covariates to make an ANCOVA?
+SuQanne Sure. Covariates are not limited to simple factorial designs.
Brilliant lecture. Thanx a lot !!
Thank you for a clear presentation about this topic. It was a useful review that helped me remember what ANCOVA was and how to word it for a proposal's statistical analysis section.
Should the genotype/ecotype be treated as a random effect as they are randomly chosen from the population and does not represent any level?
You have lucidly explained such complex experimental design..
Sir, I shall be obliged to you for posting these videos.
Very useful
very helpful
nice lecture
Thanks a lot! Great tutorial!
Great explanation, thanks!
This was great! It really helped me to put pseudoreplication into a more practical way of looking at it. However I do have a small question. If you are working in the field and you have a treatment that cannot be isolated to an individual (for example the fertilizer experiment you mentioned),in order to avoid pseudoreplication would you be best off increasing your study site area so you can randomly assign treatments? Cheers
Generally, yes. But you do what is practical, remembering that you need replication of plots within treatments (if you can't apply treatments to individuals), and that replication will determine the MS error for your treatment effect; thus the larger the N of plots, the smaller the MS error, and the larger F. In the end, it's a balancing act between practicality of treatment application, how large you want your statistical universe to be, and number of individuals you are willing to measure. Sometimes, for example with growth chambers, the experimental unit is limited by resources available (number of growth chambers). If you can't have replicate growth chambers for practical reasons, there are other strategies that can help; e.g., rotating treatments among growth chambers.
just in time to save the day! Thank you!