"Evicting Einstein"

<blockquote id="quote"><font size="2" face="Verdana, Arial, Helvetica" id="quote">quote:<hr height="1" noshade id="quote"><i>Originally posted by Thomas</i>
<br />[Re: Phys.Rev.Lett. 24, 1373-1376 (1970)]:<blockquote id="quote"><font size="2" face="Verdana, Arial, Helvetica" id="quote">quote:<hr height="1" noshade id="quote">[tvf]: Suppressing outliers can bias the solution. And the mean error of a solution parameter is supposed to be much less than the mean error of the contributing observations.<hr height="1" noshade id="quote"></blockquote id="quote"></font id="quote">Using all the data points shown ..., a slope of 1.51 &plusmn; 0.32 is obtained.<hr height="1" noshade id="quote"></blockquote id="quote"></font id="quote">That value is not significantly different than the one in the article, and does not make your case for a Newtonian value.

<blockquote id="quote"><font size="2" face="Verdana, Arial, Helvetica" id="quote">quote:<hr height="1" noshade id="quote">the error is substantially larger than the 0.19 quoted in the paper<hr height="1" noshade id="quote"></blockquote id="quote"></font id="quote">I take this as an indication that you failed to reproduce the authors' solution, not an indication that the relevant error should be larger. Did they specify "mean error" or "probable error"? Did they apply weights to the individual data points? Did they fit a different functional form?

<blockquote id="quote"><font size="2" face="Verdana, Arial, Helvetica" id="quote">quote:<hr height="1" noshade id="quote">However, even without the least squares calculation, it is obvious that there are unknown systematic errors which render the data analysis useless: about 1/3 of the data points have a distance from the line of regression of 2 or more standard deviations of the individual errors. This should be next to impossible if the data spread is of a purely statistical nature.<hr height="1" noshade id="quote"></blockquote id="quote"></font id="quote">It is impossible, and the data is “of a purely statistical nature”. I’m guessing you are using the wrong mean error with the regression diagram, because I see no inconsistency between the diagrams and the solution. You do realize that the 1.7” offset occurs only at the limb, and that most stars have an offset less than the mean error of that offset, don’t you?

<blockquote id="quote"><font size="2" face="Verdana, Arial, Helvetica" id="quote">quote:<hr height="1" noshade id="quote">On the other hand, if these offsets are caused by other influences not taken into consideration in the analysis (e.g. plasma effects in the solar corona)<hr height="1" noshade id="quote"></blockquote id="quote"></font id="quote">We know enough about the density and content of the solar corona to be sure that is not a possible explanation.

<blockquote id="quote"><font size="2" face="Verdana, Arial, Helvetica" id="quote">quote:<hr height="1" noshade id="quote">I think that this data analysis does not even have the standard of a student exercise, let alone a ground breaking scientific publication.<hr height="1" noshade id="quote"></blockquote id="quote"></font id="quote">You have not justified such a conclusion, and saying it on the basis of what you have done indicates bias. If you want to get a truth, you need to learn to suppress your own biases. E.g., have an objective third party whose competence you trust redo the analysis without knowing either the authors' result or yours, and without knowing of any "preferred" result. Or contact the authors with a short list of questions. Or, if you are super-confident that you now know data analysis better than published authors, write your analysis up and submit it to the same journal to correct the errors in the published paper.

OTOH, if you care more about vindication than truth, good luck, but you will have no further use for my opinions.

[Re: D.S. Robertson et al., Nature 349 (1991) 768]:<blockquote id="quote"><font size="2" face="Verdana, Arial, Helvetica" id="quote">quote:<hr height="1" noshade id="quote"><blockquote id="quote"><font size="2" face="Verdana, Arial, Helvetica" id="quote">quote:<hr height="1" noshade id="quote">[tvf]: The plot shows the data distribution, not the data itself, as the caption says.<hr height="1" noshade id="quote"></blockquote id="quote"></font id="quote">I can only see one vertical scale in [the plot] and since the smooth curve is the theoretical deflection radio sources near the sun, the histogram should display a deflection as well.<hr height="1" noshade id="quote"></blockquote id="quote"></font id="quote">No, it should not. You have misunderstood the plot and its explanation. The basic plot is of number of observations vs. solar angular distance. Superimposed on that plot is a plot of predicted deflection angle vs. solar angular distance.

Such misunderstandings should at the very least give you pause about being so confident in your "refutations" of published experts. -|Tom|-

Since the data is clear the light is bent by gravity would it be beyond the limits to suggest that light has mass?

<blockquote id="quote"><font size="2" face="Verdana, Arial, Helvetica" id="quote">quote:<hr height="1" noshade id="quote"><i>Originally posted by Jim</i>
<br />Since the data is clear the light is bent by gravity would it be beyond the limits to suggest that light has mass?<hr height="1" noshade id="quote"></blockquote id="quote"></font id="quote">You can suggest that if you wish, but it does not follow from light bending. Gravity affects all objects of whatever mass, from zero to infinity, exactly the same. -|Tom|-

I see how gravity effects everything between zero and infinity but not zero and infinity. How can gravity have an effect on zero?

<blockquote id="quote"><font size="2" face="Verdana, Arial, Helvetica" id="quote">quote:<hr height="1" noshade id="quote"><i>Originally posted by Jim</i>
<br />I see how gravity effects everything between zero and infinity but not zero and infinity. How can gravity have an effect on zero?<hr height="1" noshade id="quote"></blockquote id="quote"></font id="quote">A literally zero mass would not exist. However, it is customary to adopt "test masses" for various purposes, where these hypothetical objects have zero mass by definition. Then we examine the dynamics of such test masses in the limit as their mass approaches zero.

This is the same limit process used in differential calculus. If it is unfamiliar, read the introduction of a basic calculus book.

As this applies to the present discussion, the acceleration of a target body in the gravitational field of a source mass is independent of the target body's own mass. "Independent" means the same acceleration occurs for any mass, where "any" includes "zero" or even "negative", whatever that might mean. -|Tom|-

There is no need for the licture in dynamics to determine weather or not light has mass. Either light has mass or it has none. If light has mass then the gravity of the sun will cause the beam of light to bend-right? That is all I was asking so lets keep this simple.