6. Examples of Experimental Designs
From Andy Law
views
comments
Related Media
Introduction to Experimental Design
Training session with Dr Helen Brown, Senior Statistician, at The Roslin Institute, January 2016.
***********************************************
These training sessions were given to staff and research students at the Roslin Institute. The material is also used for the Animal Biosciences MSc course taught at the Institute.
***********************************************
*Recommended YouTube playback settings for the best viewing experience: 1080p HD
************************************************
Content:
‘One factor’ design
-Compare 2 or more groups
---Eg different genotypes, different interventions
-Eg Test effect of treatment on red blood cell count in mice:
‘One factor’ design example
-‘One factor’ design example:Test effect of treatment on red blood cell count in mice
-Objective: Test whether red blood cell count differs between control mice and mice injected with Chloram
-Sampling unit: Mice
-Reducing variability:
-Inclusion criteria:
---mice within specified weight range
---pre-treatment RBC within specified range
-Avoiding bias:
---randomise mice to treatments
Does Chloram affect RBC in mice?
-One factor design
--- Factor is treatment with 2 levels
-Randomise mice to treatments
Randomisation using Excel
Does Chloram affect RBC in mice?
-Mice randomised to treatment group
-Statistical analysis – t-test
-How many mice? - Covered in Part 2
Another ‘One factor’ design
-Another ‘One factor’ design: Does Chloram or Treatment X affect RBC in mice?
-Still a ‘one factor’ design - factor is treatment
-Mice randomised to 3 treatment groups
‘Two factor’ design
-Compare 2 things at one
---Eg different genotypes and different interventions
-Make comparisons within more than one group
---Compare interventions within each genotype
-Test interaction: eg does effect of intervention differ between genotypes?
‘Two factor’ design example
-‘Two factor’ design example: Test whether treatment AND strain affect RBC in mice
-Objectives:
---Does Chloram affect RBC in mice?
---Does RBC differ between strains of mice?
---Does Chloram affect RBC in both C3H and CD-1 strains?
---Does the effect of Chloram differ between C3H and CD-1 strains?
-Experimental unit: Mice
-Inclusion/Exclusion criteria:
---weight within specified range
---pre-treatment RBC within specified range
-Measures to avoid bias:
---randomise mice to treatments, within strains
Test effects of treatment AND strain on RBC in mice
-Two factor design
---Factors are treatments and strain
-RANDOMISE each strain of mice to treatments…
Randomisation using Excel
---Sort the treatment column using the random number column
---Add a column of sequential mouse numbers
Test effects of treatment and strain on RBC in mice
-Two factor design:
---Factors are treatments and strain
---Each strain of mouse randomised to treatments…
-Statistical analysis: Two-way ANOVA / GLM and t-tests
Designs with more factors
-Can tests 3 or more factors in a design
-Eg treatments, diets, strains, and their interactions
-However, interpretation becomes more difficult,
-May be preferable to carry out separate experiments
Training session with Dr Helen Brown, Senior Statistician, at The Roslin Institute, January 2016.
***********************************************
These training sessions were given to staff and research students at the Roslin Institute. The material is also used for the Animal Biosciences MSc course taught at the Institute.
***********************************************
*Recommended YouTube playback settings for the best viewing experience: 1080p HD
************************************************
Content:
‘One factor’ design
-Compare 2 or more groups
---Eg different genotypes, different interventions
-Eg Test effect of treatment on red blood cell count in mice:
‘One factor’ design example
-‘One factor’ design example:Test effect of treatment on red blood cell count in mice
-Objective: Test whether red blood cell count differs between control mice and mice injected with Chloram
-Sampling unit: Mice
-Reducing variability:
-Inclusion criteria:
---mice within specified weight range
---pre-treatment RBC within specified range
-Avoiding bias:
---randomise mice to treatments
Does Chloram affect RBC in mice?
-One factor design
--- Factor is treatment with 2 levels
-Randomise mice to treatments
Randomisation using Excel
Does Chloram affect RBC in mice?
-Mice randomised to treatment group
-Statistical analysis – t-test
-How many mice? - Covered in Part 2
Another ‘One factor’ design
-Another ‘One factor’ design: Does Chloram or Treatment X affect RBC in mice?
-Still a ‘one factor’ design - factor is treatment
-Mice randomised to 3 treatment groups
‘Two factor’ design
-Compare 2 things at one
---Eg different genotypes and different interventions
-Make comparisons within more than one group
---Compare interventions within each genotype
-Test interaction: eg does effect of intervention differ between genotypes?
‘Two factor’ design example
-‘Two factor’ design example: Test whether treatment AND strain affect RBC in mice
-Objectives:
---Does Chloram affect RBC in mice?
---Does RBC differ between strains of mice?
---Does Chloram affect RBC in both C3H and CD-1 strains?
---Does the effect of Chloram differ between C3H and CD-1 strains?
-Experimental unit: Mice
-Inclusion/Exclusion criteria:
---weight within specified range
---pre-treatment RBC within specified range
-Measures to avoid bias:
---randomise mice to treatments, within strains
Test effects of treatment AND strain on RBC in mice
-Two factor design
---Factors are treatments and strain
-RANDOMISE each strain of mice to treatments…
Randomisation using Excel
---Sort the treatment column using the random number column
---Add a column of sequential mouse numbers
Test effects of treatment and strain on RBC in mice
-Two factor design:
---Factors are treatments and strain
---Each strain of mouse randomised to treatments…
-Statistical analysis: Two-way ANOVA / GLM and t-tests
Designs with more factors
-Can tests 3 or more factors in a design
-Eg treatments, diets, strains, and their interactions
-However, interpretation becomes more difficult,
-May be preferable to carry out separate experiments
- Tags
-