https://communities.sas.com/t5/SAS-Data-Science/linear-regression/m-p/326034#M4894 <P>That depends on what question you are trying to answer. I'm assuming that not all of the variables are actual covariates that "need" to be in the model. Going through each of the 200 variables is pretty much a "digging" process, and eventually the likelihood is that you'll end up with one that is significant, even if it doesn't make sense to have it in the regression model.</P> Thu, 19 Jan 2017 16:26:08 GMT jo_rod 2017-01-19T16:26:08Z https://communities.sas.com/t5/SAS-Data-Science/linear-regression/m-p/325965#M4893 <P>friends, suppose i have 200 variables and i have to make a linear regression model on that , then what is the best &nbsp;way of understanding each variable. or i should first reduce the variables by using factor analysis or using automatic linear regression to finalise the significance variables?</P> Thu, 19 Jan 2017 13:50:55 GMT https://communities.sas.com/t5/SAS-Data-Science/linear-regression/m-p/325965#M4893 Bipin_gaud 2017-01-19T13:50:55Z https://communities.sas.com/t5/SAS-Data-Science/linear-regression/m-p/326034#M4894 <P>That depends on what question you are trying to answer. I'm assuming that not all of the variables are actual covariates that "need" to be in the model. Going through each of the 200 variables is pretty much a "digging" process, and eventually the likelihood is that you'll end up with one that is significant, even if it doesn't make sense to have it in the regression model.</P> Thu, 19 Jan 2017 16:26:08 GMT https://communities.sas.com/t5/SAS-Data-Science/linear-regression/m-p/326034#M4894 jo_rod 2017-01-19T16:26:08Z