In the absence of any context, it is nearly impossible to offer useful advice. E.g.: "I have to produce..": Why? At whose behest? In some institutional context, perhaps? "intermediate Excel ... users": high school? college? graduate school? psychologists? accountants? sci-fi authors? middle management? And "intermediate" with respect to what?
Further comments embedded below... On Fri, 27 Feb 2004, Jacob Thomas wrote (edited): > I have to produce a simple handout for intermediate Excel Chart/Graphs > users, which unfortunately for me has to include Scatter Graphs. I can > follow the logic behind the trend line but am not sure how to > interpret the trend line equation. For instance on the chart I have > produced the trend line equation is y = 1.4909x + 3.7636. What variable is "y"? What is "x"? In what units are they measured? Note, please, that "y = f(x)" does not really (as it may seem to) say that the observed values of y are equal to this function of x; "y" in this equation should be interpreted as "the predicted value of y" (using the function of x that in some sense best predicts it). Commonly in textbooks one imposes a circumflex (^) over the "y", and reads the result "y-hat", as a way of distinguishing between some observed value y and the corresponding predicted value y-hat. (If this sounds to you like kindergarten, so much the better; if it distresses you to be addressed in kindergarten language, recall that you provided no information about your own level of understanding.) All that one can say, from the information available, is that "y" can be predicted from "x" by starting with 3.76 and adding 1.49 times the value of x. This will be defensible only over a limited range of values of x, but you have supplied no information as to what that range might be. You also have failed to supply information (presumably available from the "trend line" routine; I am unfamiliar with Excel and would never use it for this purpose) that would help one to tell whether either of the constants reported (3.76 or 1.49) differ significantly from zero (or any other value, for that matter). Whether it is useful, or reasonable, or good, or bad, that the fitted function predicting y increases by 1.5 units every time x increases by 1 unit, I cannot tell. > Is this good or bad, ... Impossible to tell, out of context. In fact, without context, one cannot even tell what "good" and "bad" might mean. > and at what point does it become of no use? On what continuum are you perceiving "points" (in which this question makes sense)? "Of no use" with respect to what intended utility? As it stands, I find the question meaningless, therefore unanswerable. (This is not to imply that your idea(s) be meaningless, only that I do not perceive them clearly enough to make sense of them.) > I am looking for a simple explanation/definition for non > mathmaticians, and in less than 30 words!! Thanks . . There may be such an explanation suitable for your context; but if there is, I rather suspect you must generate it yourself (one cannot explain what one does not understand), and it will take far more than 30 words for anyone to help you come to an understanding adequate to the task. More might be written, but perhaps this is enough for a start. ------------------------------------------------------------ Donald F. Burrill [EMAIL PROTECTED] 56 Sebbins Pond Drive, Bedford, NH 03110 (603) 626-0816 . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
