Question

Topic: Student Questions

Scientific Study, Low-rsq , Low P, Contradictory?

Posted by Anonymous on 25 Points
Hello, I am currently working on a scientific study where I am comparing the correlation between two independent variables, temperature and particulate matter concentrations. All variables are collected at the same time and same place. Conducting a correlation analysis on Minitab resulting in a p-value of less than 0.01, stating there is significance in the data? I performed a regression analysis and yielded an R-squared of less than 0.10. There are 770 observations.

So, with the p-value stating that the data is significant, does the R-squared state the value of the significance? Only 10%? I guess I am asking how does this relatively low R-sq value in scientific standards describe that the data? Can the R-sq value be an indicator of strength? 1 being strong and in this case .10 being weak? So in this example there is significance between temp and particulate matter conc. but its a weak significance? And if so can that level of strength be considered a percentage, for example in my case only a 10% correlation?

Thanks!
To continue reading this question and the solution, sign up ... it's free!

RESPONSES

  • Posted on Author
    JV thanks for the heads up. I wasn't really focusing on that aspect when I was reviewing other questions. Would the statistics part of my question vary so much from science to marketing? If yes then thanks and I'll just close out the question.

    Thanks.

Post a Comment