finding the impacts of independent variables to dependent variables

#1
Hi everyone. I am new in this forum and also new in statistic. My background is civil engineering.
I've been looking for a solution for my question and found out that this forum is very helpful.
I need to explore how my independent variables affect/impact the dependent variable, which one has significant influence and which one hasn't.
The problem is my data is far from normally distributed, it has almost no distribution and those transformation don't help much.
Is there a way I could find out the impacts ?
Thank you very much guys
 

hlsmith

Not a robit
#2
There are a group of statistics called non-parametric (parameters are things such as means). These test do not have assumption requirements for data distributions. Since there are many types of tests and ways data can be grouped, please tell us about the variable formatting for your data and we can recommend some tests.

For example, compare two group values for a countinuous variable. Would use a Wilcoxon rank sum test.
 
#3
thanks hlsmith. I am exploring how different variables (congestions, weather, passengers,etc.) affect the Average bus travel time variability (dependent variable).
Each DV observation is the SD of average bus travel time on different sites.
My DV has no particular distribution. For some periods if I keep segmenting the data it could be fitted by a log-normal, dagum or log-logistic.
If I do a Wilcoxon signed-rank test and H0 (median difference between the pairs is zero) is rejected. Does it mean my variable is significantly affect my DV ?
 

hlsmith

Not a robit
#4
It would vaguely mean the continuous variable differs based on the iv given you are not missing any confounder. I typically determine the Median of both groups and report them with interquartile ranges and pvalue so reviewers understand where the difference is.
 
#6
I did some segmenting of DV and now my DV is Pearson 5 (Inverse Gamma) distribution. Could any one help me what should I do for the next step. Thanks