Project First Steps and Overview Final projects will be presented by poster or Power Point April 30, 2019, 10:30 AM-12:30 PM, CL 1005. One copy per group will be submitted in a D2L Dropbox. EACH...

Project First Steps and Overview

Final projects will be presented by poster or Power Point April 30, 2019, 10:30 AM-12:30 PM, CL 1005. One copy per group will be submitted in a D2L Dropbox. EACH STUDENT will also submit THEIR OWN written report using MS Word that will be submitted via D2L dropbox April 25, 2019, 11:59 PM.

First Steps

1. 
   

   Establish your research group.

   
   

• 
  

  I would like to know by Feb. 12 who is in your group.

  
  


• 
  

  The groups can be of size 1, 2 or 3.

  
  

2. 
   

   Find your data set. Your data set should include a sample size of at least 100 (the more the merrier!).

   
   

• 
  

  You must identify one quantitative response variable that you wish to explain or predict using the others.

  
  


• 
  

  You need at least 4 predictor variables.

  
  
  
  
  • 
    

    It is not necessary to throw away variables in order to only have 4 predictors. If you have more, it won’t complicate what you’re doing. Chances are, when you find out which are useful, it won’t be very many of them.

    
    
  
  
  
  


• 
  

  Your data set needs at least one categorical variable.

  
  
  
  
  • 
    

    If it your data set does not have a categorical variable to begin with, then simply plan to recode a quantitative variable.

    
    
  
  
  
  


• 
  

  At least two of your predictor variables should be quantitative.

  
  


• 
  

  I need to approve your data set by March 12.

  
  

Overview

1. 
   

   You will do a lot of things to the data that will not ultimately be displayed on your poster or PPT; although, they will be reported in your paper. If you have your data earlier in the semester, you can do various things as we learn them. On March 12, we begin to talk about Variable Selection. At this point, you can start seeing how to select the “best” model or two.

   
   


2. 
   

   Each group will do the following with their data:

   
   

• 
  

  Find two models they deem as “best” and explain why. Using values of the root mean square error and/or Adjusted R-square work well. Also, the significance of the model and the variables therein is important. This is almost certainly part of the poster or PPT.

  
  


• 
  

  Code in at least one interaction term between a categorical variable and a quantitative variable and explore to see if it is significant or not. This won’t make the poster if it doesn’t end up being important, but you’ll write about it in your report.

  
  


• 
  

  Code in a higher order term if anything leads you to think it could be beneficial.

  
  


• 
  

  Create confidence intervals for the coefficients of the best models.

  
  


• 
  

  Use the best models for prediction and confidence intervals.

  
  


• 
  

  Use a residual analysis to assess models for meeting the regression assumptions.

  
  

3. 
   

   Each poster/PPT will be different. I don’t want to give everyone the same instructions and cause some people to display things that are not important for their research.

   
   

• 
  

  Each of you will need a title that is descriptive of your research question:

  
  
  
  
  • 
    

    “Predicting Average Life Expectancy in Years Based on a Country’s Per Capita Gross Domestic Product” OR

    
    
  
  
  • 
    

    “Explaining the Relationship Between Life Expectancy in Years and a Country’s Per Capita Gross Domestic Product”

    
    
  
  
  
  


• 
  

  Make sure your name(s) are beneath the title.

  
  


• 
  

  Each poster or PPT should have an abstract.

  
  


• 
  

  Each will have a conclusion.

  
  


• 
  

  Then there is what lies between.

  
  


• 
  

  Make sure to list references or acknowledgements if credit is due.

  
  

4. 
   

   Each set of presenters will hand in a set of appendices to go with their poster or PPT.

   
   

• 
  

  One should be a table listing all variables (name, description, units).

  
  


• 
  

  Next will be your SAS code.

  
  


• 
  

  Then your SAS printouts, but PLEASE tell me what I’m seeing. It can be multiple pages.

  
  


• 
  

  Title the different sections of the output (either from within SAS or within Word).

  
  

(In the code and the printout, as well as the report, I should see you’ve done all I asked even if some things didn’t make the cut for the presentation.)

5. 
   

   The poster or PPT itself will be displayed/presented in our classroom on final exam day, April 30.

   
   

• 
  

  Each GROUP should send me an electronic copy of their poster or PPT with the appendices.

  
  
  
  
  • 
    

    You need not print the appendices. I will make a “Presentation” Dropbox item to which you can submit it all.

    
    
  
  
  
  


• 
  

  EACH STUDENT should submit their OWN written report about their team’s complete project, but also clearly identifying their part in it.

  
  
  
  
  • 
    

    Reports should have:

    
    
    
    
    • 
      

      Abstract

      
      
    
    
    • 
      

      Introduction

      
      
    
    
    • 
      

      Methods

      
      
    
    
    • 
      

      Results

      
      
    
    
    • 
      

      Conclusion/Discussion

      
      
      
      
      • 
        

        Include any limitations to the research, possible confounding variables, whether the results can be generalized, how the study could be improved or furthered, etc.

        
        
      
      
      
      
    
    
    
    
  
  
  • 
    

    The written report should be submitted in the “Report” Dropbox I’ll create, as well, by 11:59 PM the last day of regular class (April 25).

    
    
  
  
  
  

These instructions are subject to change with notice.