1) what is the difference between a GLM and a GLMM?
GLM: only has fixed effects
GLMM: has fixed and random effects
both have: a set distribution and associated link function, response variable not assumed to be linear or normal, response variable should be continuous, explanatory variables are categorical and can have continuous covariates
2)What is the difference between fixed and random effects?
Fixed: this is the treatment in your experiment, should have an effect on your response (slope changes)
Ex: soil characteristics are the fixed effect when testing the influence of soil on plant growth of different species
Random: intercept changes but should not have an effect on your response (slope does not change)
Ex: species is a random effect when testing the influence of soil characteristics on plant growth of different species
3) Checklist of steps for running a GLM/GLMM (Box 4 Bolker)
- specify model: define – fixed effects, random effects, covariates, response variable, a priori decide on the number of factors to include in the model
- choose distribution and appropriate link function
- check assumptions: homogeniety of variances, outliers, does your distribution match the assumed distribution? (goodness of fit statistics). If variances are not homogeneous or the distribution does not match then change your model, adding effects or covariates, interactions etc
- recheck assumptions of final model (see previous step)
- check overdistribution (especially for poisson distributions)
- can choose/compare models using AIC values (lower AIC = better)
4) Common errors that are made when performing a GLM/GLMM
Selecting the wrong distribution and/or the wrong link function
not checking the fit of the model to the distribution
Systematic review paper/workflow review /50
Title: informative, clear, useful. /2
Abstract: summarizes enture paper succinctly including implications /5
Introduction: Clear, appropriate length, introduces topic, states purpose and objective /10
Methods: Replicable, transparent /5
Results: Written description can stand alone /3
Discussion: Clear, does not repeat results, highlights novelty, relates to previous research, states implications /10
Lit cited & lit reviewed: appropriate and extensive (at least 40 citations and sufficient body reviewed) /5
Figures & tables: sufficient, clear, stand alone capacity, summarizes salient findings /10
Open-science products /25
datasets on figshare
presentation on slideshare
figures/infographics on figshare
interactive figure on plot.ly
video on camtasia
Information content sufficient to communicate main finding and include salient evidence/main findings /5
Visually captivating and attractive to appeal to a wide audience /5
Text associated with product well written and appropriately conveys details needed to understand product (i.e. meta-data, tags, text on slides, text on infographic, etc). /5
Science behind product sound, evidence-based, not too simplistic, and can be assessed by a scientific non-expert to your field /5
Related evidence also referenced (other workflows, reviews, statistical tests, products etc) very important to show how your work connects to other /5
(1) Thank you for the presentations last week (past). Super useful to get a grasp of the scope of the research and analytics for your fields.
(2) Reminder (future): beginning next week, you guys are up to bat for 3 weeks of advanced statistical presentations. We have the first two weeks covered by students. The last week, I can do as best I can.
Oct 23rd: GLM/GLMMS
Nov 6th: CART & Ordination
Nov 13th: SEMs and Advanced regressions.
hi team, just a quick update.
1. I deleted all commenting on blog. Spam is so annoying. Those bots!
2. Remember, you have to publish the data you collect for your review on figshare as part of your grade. It must include meta-data etc.
3. I have had a request for guidelines. If you are not afraid of getting scooped, I strongly recommend you publish paper on peerj as a pre-print.
Copy for formatting style of this paper.
Summary – you must publish data, include a PRISMA report, have at least two figures, and write a paper like the peerj example.
We can discuss this all next week, Oct9th, in class. However, in the interim keep cranking on those papers, get the two-three figures ready, the PRISMA report, and maybe the methods & results too.
The more you have prepped for next week, the more I can help.
Hey Guys! The overall objective of my research project is to create a lab-on-a-chip device platform for drug/chemical assays using fruit fly embryos. The first part of my project is to study Drosophila oviposition (specifically, its response to physical properties such as textures). From this, I plan to micro-fabricate a patterned device to collect embryos and to pass chemicals through them for developmental imaging. So, for this course, I think I’ll be looking at how statistics is used in presenting oviposition data and errors.
For my research I am focusing on how plant facilitation effects associated animal species. For my systematic review I will look at the statistical tests used in studies that examine plant (specifically shrubs) facilitation that also incorporate another trophic level in some way. My search terms are facilitat* AND shrub.
My research topic is focusing on transcription factor EB (TFEB), known as the master regulator of lysosomal biogenesis, and how its activity and downstream targets are affected by exercise in skeletal muscle.
For my review I will look at the statistical tests used in my area, I could search: Exercise AND skeletal muscle* AND autophagy OR lysosome* OR TFEB
My research project will be looking at restoration ecology and the stats tests used within this field. More specifically, I will be looking at studies investigating methods of restoring invaded ecosystems.
Hey guys! My systematic review will centre around the statistical tests have been used in studies relating to pollinator facilitation and/or the co-evolution between plants and their pollinators.
Possible search terms: plant OR shrub OR cushion AND pollinat* OR arthropod OR insect AND facilitation OR evolution OR co-evolution
and here is the camtasia app that Ally mentioned instead of youtube: http://www.techsmith.com/camtasia.html
Google drive for collaboration within course: http://gizmodo.com/5904653/what-is-google-drive