In all the example of linear regression model I have studied X is taken as fixed value ( midpoint of interval). Can we have a regression model in which X is in interval form and how this will affect regression analysis. And how this will affect condition mean.

  • $\begingroup$ Yes, such models have been formulated. What are you willing to assume about $X$? What does an interval form for it mean, exactly? Will $X$ also be a random quantity or not? If so, what will you assume about its distribution? $\endgroup$
    – whuber
    Feb 15, 2018 at 20:26
  • $\begingroup$ I am asking whether x values can be continuous i.e in intervals like 100-120, 120-140.... or it takes only discrete values like 115, 130 ... $\endgroup$ Feb 16, 2018 at 7:25
  • $\begingroup$ Could you clarify what you mean by "interval form"? Are you referring to values that are actually ranges of numbers or do you mean the so-called "interval" data type (a continuous variable whose numerical differences are meaningful)? You seem to be using both meanings in your question! $\endgroup$
    – whuber
    Feb 16, 2018 at 14:51
  • $\begingroup$ Interval i.e. continuous data type. $\endgroup$ Feb 17, 2018 at 4:13


Browse other questions tagged or ask your own question.