Linear regression

[laurenhosking]

I’m doing a linear regression between race time (y-axis) and average training speed in the month prior to the race( xaxis) which would be my dependent variable and which my continuous. I have time as my dependant is this right or the other way round 

[PaigeMiller]

---

@laurenhosking wrote:

I’m doing a linear regression between race time (y-axis) and average training speed in the month prior to the race( xaxis) which would be my dependent variable and which my continuous. I have time as my dependant is this right or the other way round 

---

Well, normally this is something you have to decide, which variable is independent and which variable is dependent. You decide based upon your knowledge of the subject matter. If changes in one variable are thought to possibly be the cause of the change of the other variable, then the cause is the independent variable and the result in the dependent variable. (Example: adding more soap to water increases the amount of suds) If they are both correlated, but one is not thought to be the cause of the other, then linear regression doesn't particularly make sense (example: you have measured height and weight of individuals).

So which case is your case? That's really up to you, but I lean toward the latter.

PS: your usage of "xaxis" and "y-axis" is not helpful, and you can't designate variables thus until you decide which is the independent variable, that one (according to the almost universal practice) is plotted on the x-axis.

[PGStats · Accepted solution]

Adding to @PaigeMiller comments...

Since training precedes the race, and a mechanism can be hypothesed to link training to race performance, I would consider training as an independent variable and race performance as a dependent variable. I.e. "Your race performance depends, to some extent, on your training", but not much the other way around.

[PaigeMiller]

Yes, @PGStats, I can see how you could argue this one either way. And this particular issue may not be relevant for a beginner anyway.