# Percentage change of dependent variable (MLR)

#### bellcross

##### New Member
Greetings,

I went through SPSS regression analysis of a certain dataset, tested assumptions, and finally I got a multiple regression model: y=292.953+9.548x_1+14.982x_3+8.907x_5+16.808x_7

Now suppose that next year x_1 is expected to increase by 50% (other assumed to stay the same). I am asked to estimate an effect on y. What is the best way to tackle this problem?

#### hlsmith

##### Omega Contributor
Do you know how to get the predicted value of y given the model you presented. Well you could times the value of X1 by 1.5. Do you want to tell us more about the model variables?

#### bellcross

##### New Member
Do you know how to get the predicted value of y given the model you presented. Well you could times the value of X1 by 1.5. Do you want to tell us more about the model variables?
Sure I did it, but then I decided to approach the task in the following way. I randomly assign values within a 95% CI of the mean and subsequently multiply x1 by 1.5 and compare both results. Y increases by 23-25% on average. Should be valid approach. What do you think?

#### GretaGarbo

##### Human
Sure I did it, but then I decided to approach the task in the following way. I randomly assign values within a 95% CI of the mean and subsequently multiply x1 by 1.5 and compare both results. Y increases by 23-25% on average. Should be valid approach. What do you think?
I think it is wrong. Why should it be "valid"?