Final Review

• Purpose of an Abstract


  
  

  • helps you conduct, present, read, write, and condense a study
• What is an abstract useful for?


  • can be useful for conference programs, book chapters, collections
• Process of an abstract


  
  

  • process includes: topic, research question, methods, results and conclusion
• When would you perform an SNK analysis? Why would you perform this analysis?


  • if anova shows there is a significance
  • to tell apart groups
• why would you want to perform an SNK analysis?


  • to tell apart groups
• WHat does the SNK analysis control? 


  • controls alpha at 0.05 and type I error rate
• Similarities between F-test and Anova that allows F-distribution to be used for both


  • they both compare variances
  • theyre reported as Fn,d=Fcrit, p<or> 0.05
• What is a problem with multiple pairwise comparisions?

  • increased type I error rate
• ANOVA vs T-test


  • ANOVA can compare more than two groups simultaneously
• 2- Factor Anova


  
  

  • asseses interactive effects between factors
• Simple F-Test


  
  

  • tests for homogeneity of variances
• SNK


  • performs multiple pairwise comparisions while controlling alpha
• 1-Factor Anova


  • determines differences in 5 levels within a single treatment group
• T-test, two sample T-test, ANOVA


  • assesses whether the means of two groups differ
• Why is psuedoreplication problematic in experimental designs?

  • artifically inflates sample size
• Populations


  • measured by parameters
• How can Experimental artifacts be minimized? 


  • effective controls
• Summarizing the frequency distribution


  • Central Tendency: mean, median and mode
  • Dispersion: variance, standard deviation, sum of squares
• Features of a manipulative experiment and how to incorporate them in study


  • control
  • precision
  • accuracy
  • consistent methodologies
• Type II Error


  • accepts H0 when it is false
  • accepts Ho when Ha is true
• Type I error 


  • Rejects H0 when it is true
  • accepts Ha when Ho is true
• Not an assumption of parametric statistics


  • outliers
  • heterscedasticity
• Sample


  • measured by statistics
• What can create experimental artifacts?

  • low degree of precision
  • systematic error
• Assumptions of parametric statistics


  • normality of data
  • data are independent
  • homoscedacity
• WHat is psuedoreplication? 


  • replicates that arent independent
• what are issues with Psuedoreplication


  • artificially inflate sample size
• how to avoid Psuedoreplication


  • ensure independece of data at the experimental design stage