datesoftgo.blogg.se

• REGRESSION LINE EXCEL 2008 FOR MAC CODE

I gathered data for this part inaccurately which affected my regression line. We are then tasked to gather data and asked to estimate where an unknown would lie. We are given that the relationship is linear. Thanks for the help and I hope that this thread (though I jacked it) will be helpful to a great many more people.

REGRESSION LINE EXCEL 2008 FOR MAC CODE

Linr wrote:Okay, I didn't realize that there were different types of references and I was still getting it wrong until I looked carefully at your code and the link you provided. Then it's time for more questions: Did we pick the wrong model? Were our measurements faulty? Is there some unexpected bias in the experiment? etc. It's certainly possible to find that the data doesn't support the model very well. If this is for a class, and you're unsure about why the model is Y=M⋅X instead of Y=M⋅X+B, talk with your instructor. If you don't know what the model is, you have no justification for using a trend line. Choosing the appropriate model depends on knowledge, say from physics, about how the dependent and independent variables should be related. For example, it is simply wrong to use a linear trend line (Y=0 or not) if the underlying model is Y=B⋅Xᴹ. The trend line equation type depends on the model, not on the data. For Y=M⋅X, to get the best estimate for the value of M that the data supports, we must use a linear trend line and set the Y intercept to zero because that's the model we're using. See Help, Index, Trend lines in charts for more details. The model determines which type of Calc trend line is appropriate for the situation (linear, logarithmic, exponential, or power) and we can then estimate, based on the data, the model parameters (M and B for the ones shown). That is, is it expected to be Y=M⋅X, Y=M⋅X+B, Y=Xᴹ, Y=B⋅Xᴹ, Y=Mˣ, etc? The first of these corresponds to a linear model where the Y intercept is zero. I'm getting a little off topic because this is a Calc forum and not a statistics forum, but before you use a trend line you need to have some idea of the underlying "model" for the data. See Help, Index, Filling cells automatically for an explanation of the "fill" concept. See Absolute, relative, and mixed references for an explanation for why the first two parameters aren't changed the same way. Double click it and the TREND formula will be filled to the remaining rows with the third parameter changing from A2 to A3, A4, etc. You'll see a small black square the the lower right corner of C2. For five data points put the X values in A2:A6 and the Y values in B2:B6. Row 1 is for labels, as shown on my example. You only need to enter the TREND() formula once, even if there are 1000 rows.

Linr wrote:I have to manually use"TREND" each time for the number of values?No, though if you're new to spreadsheets you might not realize that.