Dear experts,
Kindly, I want to know, how can we interpret the following XmR Trend chart data?
Fit (R^2) < 0.80
Ryx = -0.501
Slope = -0.311
Sigma = 2.909
Probability = 0.532

Many thanks in advance.
Omar

Dear experts,
Kindly, I want to know, how can we interpret the following XmR Trend chart data?
Fit (R^2) < 0.80
Ryx = -0.501
Slope = -0.311
Sigma = 2.909
Probability = 0.532

I am not sure if this is a XmR chart with autocorrelation (because you put Fit(R^2)<0.80), or is an XmR chart with trend (you named it XmR Trend), can you explain?

If is XmR with autocorrelation, the (R^2)<0.8 means that there is little autocorrelation in your data, so you can use a normal XmR without autocorrelation with almost the same result. The autocorrelation coefficient is documented to have an efect when R^2>0.8 (by Donald Wheeler)

Ryx =-0.501, is a low correlation coeficient (not so low for real life data but for a controled experiment is medium low), so... you can't be sure that there is a relationship between your variables, the low slope (almost 0 = -.311) show the same (altho is not a rule, because it depend on the magnitude order of your data, but if you manage numbers of 1rst order of magnitude, this slope is nill).

Sigma, could be the XmR sigma, but I can't be sure without the data.:cfingers:

I am not sure if this is a XmR chart with autocorrelation (because you put Fit(R^2)<0.80), or is an XmR chart with trend (you named it XmR Trend), can you explain?

If is XmR with autocorrelation, the (R^2)<0.8 means that there is little autocorrelation in your data, so you can use a normal XmR without autocorrelation with almost the same result. The autocorrelation coefficient is documented to have an efect when R^2>0.8 (by Donald Wheeler)

Ryx =-0.501, is a low correlation coeficient (not so low for real life data but for a controled experiment is medium low), so... you can't be sure that there is a relationship between your variables, the low slope (almost 0 = -.311) show the same (altho is not a rule, because it depend on the magnitude order of your data, but if you manage numbers of 1rst order of magnitude, this slope is nill).

Sigma, could be the XmR sigma, but I can't be sure without the data.:cfingers:

Dear Darius,
it is an XmR chart with trend.
thanks.

So the interpretation still works, but as I said to interpret the importance of a slope, one has to check against the actual range of the data.

ie. if the data range is between (1 and 3) the slope of -.3 could be significative, I always use the minimum "X" and the maximum "X" when I obtain an equation and check the difference (of both estimates) as the contribution against the actual range of the data.

As a thumb rule, as I said R, less that 0.5 is low (is no relation), between 0.5 and 0.7 is medium low (you can't say that there is a relation) for non controled experiment, 0.7 to 0.9 is good correlation (there is a relation between the two variables) and 0.9 to almost 1 (there is no doudt in the relation). But first check for outliers, one could make your correlation too good or too bad.