Research Methods in Plant and Soil Sciences
LIVE SITE: Atmospheric CO2
Last update: August 16, 2017
Research Methods 2017 (on line, see below)
“Enter as a Guest”, enter your name, click “Enter Room”, Can enter questions/comments in the chat window.
Routine, Monthly Activity, By Day Records,
Experiment 502 data
Experiment 502 plot plan
Wheat Growth Stages
FINAL 2017 Spring
Friday May 5, 7:30am Room 463
OFF Campus: You will need to go to
it.okstate.edu. You can download SAS 9.4 from the software
distribution center. Use your OKSTATE email and password. Depending on
your operating system (32 or 64 bit) download the appropriate version.
|SED = square root (2*MSE/reps) = square root (2*s2/reps)||Suggestions. fill our work order on http://pss.okstate.edu/forms/pass-it-work-order You can also download and install the software yourself from the Software Distribution Center website at https://app.it.okstate.edu/sdc/. You will also get new license files from Software Distribution Center.|
Tuesday/Thursday 8:00 am, Ag Hall 463
|"Good Researchers aren't characterized by being smart, but by those who ask the right questions"|
|Class Goal: Identify problem, formulate hypothesis, implement treatment structure/experiment and generate observations that prove or disprove hypothesis.||
When is Research in a specific
field justified? The applied versus basic quagmire.
Basic Research (defined)
Applied Research (defined) (Borlaug)
Duncans Table 5%
Duncans Table 1%
Checklist for Troubleshooting Non-estimatable Contrasts
Assumptions of analysis of
Experimental error is random independent and normally distributed about a
zero mean with a
|Ian: Your scientists were so preoccupied with whether or not they could, they didn’t stop to think if they should. Jurassic Park||THE WALL, Dr. Borlaug|
|Mean Separation: Calculation of the SED||
Experimental Method should
attempt to remove all sources of variability other than the effects being
evaluated within the treatment structure.
(Everything other than the variables being evaluated should be held constant)
Swallow, Willima H. 1984. Those overworked and oft-misused mean separation procedures - Duncan's, LSD, etc. American Phytopathological Society. October, 919-921.
Relationship between Mean Square Errors and Wheat Grain Yields in Long-term Experiments. J. Plant Nutr.
Washington Post: Too much of too little (November 2013)
Frozen versus Non-Frozen Sample Preparation for Plant Tissue Phosphorus Analysis. Commun. Soil Sci. Plant Anal. 20:197-211.
Emergence of Corn as Affected by Source and Rate of Solution Fertilizers Applied with the Seed. J. Fert. Issues. 3:18-24. (PDF)
Economic and Agronomic Impacts of Varied Philosophies of Soil Testing. 1982. Agron J. 74:492-499.
a. compute the F-Test for combining any two years (not covered in class), need SAS printout. Run glm analysis by year and over years (combined). 3 sets of treatment means (year 1, year 2, combined)
b. treatment means (by year and combined) Does this change your conclusions (use the treatment structure) given for 502
From experiment 502, generate a linear model of NDVI versus grain yield for 3 different year periods. The SAS printout will do. Provide a written interpretation of what you found, and what the R2 were for each relationship. What percent of the variability in y (yield) was explained by x (NDVI)?
Using the SAS program provided in class, generate a functional surface response model using two random sources of data plotted against a dependent variable that you measured. Example, NDVI, at Feekes 4 and NDVI at Feekes 5, versus grain yield or "buac" in Experiment 502.
For the data set included, decipher whether or not the covariate was significant in removing variability in maize grain yield. Output needs to include GLM results, and significant levels of main effects using the appropriate error term to test the effect. Tillage is "split" in this trial, so you need to have 2 error terms. You should also include: 1. treatment means, and standard deviation, 2. least square adjusted means . The Covariate is "preP" or preplant soil test phosphorus.
input rep nrate tillage yield preP;
1 0 0 20 16
1 50 0 25 13
1 100 0 29 13
1 150 0 35 14
1 0 1 35 18
1 50 1 35 16
1 100 1 38 11
1 150 1 39 19
2 0 0 20 22
2 50 0 26 23
2 100 0 30 21
2 150 0 32 25
2 0 1 36 19
2 50 1 36 11
2 100 1 37 18
2 150 1 40 16
3 0 0 17 18
3 50 0 22 19
3 100 0 25 28
3 150 0 29 22
3 0 1 29 13
3 50 1 34 25
3 100 1 38 24
3 150 1 40 18
Please edit-assemble and run a linear-plateau model (non-linear regression) for data of your own, or data from one of your fellow students. Output and data used are required.
SAS Programs (wide range, includes Regression)
TOOL BOX outline
1. Combining sites (SAS 502 DATA, 2012, 2013, 2014, 2015, 2016)
RCBD versus CRD, computing df
2. SAS: 502, 2015 and 2016 data, glm, by site, over sites, means, plots
3. NDVI various Stages versus Yield (2/21 and 2/23, 2017)
4. Exp. 502, by-year, SED example, 2008, 2009, 2010 (kg/ha conv. to bu/ac)
5. Blake Farrow Program (no differences in locations)
6. 502_surface response model (diamond/pyramid shapes)
Orthogonal/non-orthogonal contrasts, over lapping of sums of squares
7. Proc iml and interaction contrasts
8. By-Plant Corn Studies (go to by-plant studies)
9. Covariance, Autocorrelation
10. Contrast EXAMPLE ( Eva ) (PLOT PLANS)
When should I block,
How should I block?
12. Cate-Nelson Procedure, Linear Plateau (middle of entry page)
best example we have is Melissa, NDVI data (combined program) (as txt file)
Linear Plateau (Dennis Wallin)
13. Difference between two independent Regression Equations
14. Stability Analysis
CV corn versus CV wheat
Weather Patterns, 502
CV over time
Combining Experiments Saif
15. Testing for a Normal Distribution
16. Split Plot in Space and Time
Abstract YIELD GOAL
"One shot to the head, or 5 in the chest and bleed to death" Time we all do not have
Accounting for all sources of "error" (rep*trt*year is the error term)
START: Finishing writing, importance, skills
|TEXT from SOIL 5813 Book|
|Spring Semester Academic Calendar|
|Can society justify basic research when 33,000 people die each day due to starvation? Who is "society?"|
SAS 502 DATA, 2012, 2013, 2014, 2015, 2016
Environment (Location, Year, Experiment #, Greenhouse, etc.)
1. Homogeneity of error variance?
2. Did conclusions change after sites were combined? (Could I have homogeneity of error variance, and still see different treatment responses at 2 different sites/years?)
3. How to identify outliers (SD about a mean)
rules associated with removing outliers
Physical Design, Plot Plans
RCBD vs CRD, and Contrasts (link goes to RCBD and CRD examples)
Checklist for Troubleshooting Non-estimatable Contrasts
1. The "Experimental Method" (defined)
2. The "Scientific Method"
3. Formulation of hypotheses
a. 15 Steps to Good Research (Georgetown)
b. Advice on Research and Writing (ASA abstract)
4. By-Plant Corn Studies
Collecting Added Data: When is this useful?
8. Applied Statistical Procedures
SED = square root (2*MSE/reps) = square root (2*s2/reps)
Analysis (Genotype by Environment interactions, plant Breeding)
2. Stability Analysis (excel) (Treatment by environment interactions, N Rate Example)
(example for agronomic experiments, Use of Stability Analysis for Long-Term Soil Fertility Experiments. Agron J. 85:159-167)
3. Magruder excel file
1. Apurba Trial (depth)
2. Sarah, Genotype by Environment Interaction (class sas program)
3. Jeremiah (complete factorial) (class sas program)
Discuss treatment 13 vs rest 'contrast'
4. Lecture, 2/15
5. Lecture, 2/17
Mixed vs GLM (popular press)
Mixed example, Sulu
Mixed EXAMPLE (for 2015 Class)
FINAL 2015 KEY