cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
Register for our Discovery Summit 2024 conference, Oct. 21-24, where you’ll learn, connect, and be inspired.
Choose Language Hide Translation Bar

How to compare two depend variables with several independent variables

Hello, JMP community,
 
There are two variables (serum cytokine and chemokine) and three independent variables [Mice (1: male, 2: female), Experimenters(1: man, 2: woman), and Treatment (1: saline, 2: LPS)].
I want to determine the relationships (statistical significance, such as P value) between two variables and each independent variable and their fixed factors, such as Mice✕Experimenter✕Treatment".

 

Should I use the Fit Model? 

 

Thanks so much.

 

ZY

 

 

1 ACCEPTED SOLUTION

Accepted Solutions
Victor_G
Super User

Re: How to compare two depend variables with several independent variables

Hi @ScatterMarten13,

 

Welcome in the Community !

I'm not sure to understand if you already have a dataset ready for the analysis, or if you want to plan your experiments ?

 

  • In the first case, depending on the data quality and representativeness, you may or may not be able to identify some relationships. Analyzing the quality of your data collection before any modeling (evaluate sample size & power, correlations among variables, multicollinearity, measurement repeatability & reproducibility, etc...) will be key to understand the outputs from your model and determine how much trust you can put in your model. Then, using Fit Model could help analyze the results and evaluate the relevance of the model, both statistically and with domain expertise.
  • In the second case, you can create a design according to your needs, to investigate all main effects and the 2- and 3-factors interactions. You can create the design using the platform MSA Designs (since Mice and Experimenters could be interpreted respectively as "Gauge" and "Operator" factors) or more generally using the Custom Designs platform. The analysis is then a lot easier as the data collection is linked to your objective and the assumed model you have specified.

In each case, doing validation experiments to confirm your interpretations and conclusions will help confirming your hypothesis.

I hope this first discussion starter may help you, 

Victor GUILLER
Data & Analytics

View solution in original post

2 REPLIES 2
Victor_G
Super User

Re: How to compare two depend variables with several independent variables

Hi @ScatterMarten13,

 

Welcome in the Community !

I'm not sure to understand if you already have a dataset ready for the analysis, or if you want to plan your experiments ?

 

  • In the first case, depending on the data quality and representativeness, you may or may not be able to identify some relationships. Analyzing the quality of your data collection before any modeling (evaluate sample size & power, correlations among variables, multicollinearity, measurement repeatability & reproducibility, etc...) will be key to understand the outputs from your model and determine how much trust you can put in your model. Then, using Fit Model could help analyze the results and evaluate the relevance of the model, both statistically and with domain expertise.
  • In the second case, you can create a design according to your needs, to investigate all main effects and the 2- and 3-factors interactions. You can create the design using the platform MSA Designs (since Mice and Experimenters could be interpreted respectively as "Gauge" and "Operator" factors) or more generally using the Custom Designs platform. The analysis is then a lot easier as the data collection is linked to your objective and the assumed model you have specified.

In each case, doing validation experiments to confirm your interpretations and conclusions will help confirming your hypothesis.

I hope this first discussion starter may help you, 

Victor GUILLER
Data & Analytics

Re: How to compare two depend variables with several independent variables

Dear Victor_G,

 

Thank you for your advice.

I have data and will start with the Fit Model.

 

Regards.

 

ZY