- Thread starter statfran23
- Start date

I'm not saying it's wrong, but I'm asking why are you using that formula?

Why not (.2)*(40 + 38 +37) + (.4)*(30) ?

(Trying to encourage some learning, metacognition.)

If that is the right formula, do you think the median would be more appropriate here? What about the mode?

Why not (.2)*(40 + 38 +37) + (.4)*(30) ?

(Trying to encourage some learning, metacognition.)

If that is the right formula, do you think the median would be more appropriate here? What about the mode?

1) the formula I gave is just another way of writing your formula.

2) Where are you getting the formula that you're suggesting?

The point is to encourage some independent search and discovery. Did the formula come from a website, a book? Trying to use the teach a man to fish principle.

So yes, the arithmetic average/mean is the sum of the values divided by the number of values. In general for follow-up, it might be a better idea to look at median follow up rather than mean since it's not uncommon to have huge outliers in followup, or for shorter periods, the distribution is naturally right against the theoretical lower bound of 0 follow up. If there are few values and you're interested in the most common follow up length then mode might be more appropriate. Usage of statistics in medicine requires more thought than is currently encouraged (from an insider perspective), so I'm trying to encourage a bit more prodding and inquiry in to being able to answer "Is this the right formula or approach?"

1) the formula I gave is just another way of writing your formula.

2) Where are you getting the formula that you're suggesting?

The point is to encourage some independent search and discovery. Did the formula come from a website, a book? Trying to use the teach a man to fish principle.

So yes, the arithmetic average/mean is the sum of the values divided by the number of values. In general for follow-up, it might be a better idea to look at median follow up rather than mean since it's not uncommon to have huge outliers in followup, or for shorter periods, the distribution is naturally right against the theoretical lower bound of 0 follow up. If there are few values and you're interested in the most common follow up length then mode might be more appropriate. Usage of statistics in medicine requires more thought than is currently encouraged (from an insider perspective), so I'm trying to encourage a bit more prodding and inquiry in to being able to answer "Is this the right formula or approach?"