Vous êtes sur la page 1sur 42

Brady Campaign State Rankings and the Firearms-related Death Rates Einsteins Work Function Reappears

14.0 12.0

Firearms-Deaths Rate, y

10.0 8.0 6.0 4.0 2.0 0.0 0

VA

y = -0.096x + 12.09 Mathematical Relation Line joining NJ-VA pairs

NJ A

10

20

30

40

50

60

70

80

90

100

Brady Campaign Score for the State, x


Figure 1: The Brady Campaign 2011 Score, x (0 to 100, click here), for a state is plotted on the horizontal axis. The Overall firearms-related death rate (per 100,000), y, is plotted on the vertical axis. The (x, y) pairs for New Jersey (72, 5.19) and Virginia (12, 10.94) are joined by a straight line with the equation y = hx + c = - 0.096x + 12.09. This is NOT a statistical correlation since there is ABSOLUTELY NO UNCERTAINTY in the numerical values of the constants h and c in this mathematical equation relating x and y (see America Under the Gun, by Arkadi Gerney, Chelsea Parson, and Charles Posner, hereafter GPP study, click here).
Page | 1

Table of Contents
No. 1. 2. 3. 4. 5. 6. 7. 8. Page No. Abstract 2 Pre-Introduction 3 Introduction 8 Extension of Einsteins idea of a work function 10 Non-Statistical Analysis of Gun Violence Data 17 Least Squares Analysis of Gun Violence Data 21 Appendix 1: Work function and Millikans Experiments 26 Appendix 2: Michigan counties firearms homicides and suicides 30 Topic

1. Abstract
Three of the ten gun violence indicators (Overall firearms-related death rates in 2010, Firearms-related homicides death rates in 2010, and Overall firearms related death rate for 2001-2010) in the recent GPP study (America Under the Gun, by Arkadi Gerney, Chelsea Parsons, and Charles Posner) are compared with the Brady Campaign scores, for the effectiveness of the gun laws in various states. It is seen that the states with the higher Brady scores also, in general, have lower death rates, supporting the viewpoint of reduced gun violence being linked to gun laws effectiveness, as also noted in the GPP study. We also avoid the pitfall of using statistical correlations since it is now universally believed that statistical correlations do not imply any causation and therefore do not advance any viewpoint in favor of or against anything. We will therefore use the concept known as the work function, first conceived by Einstein, in 1905, to explain the puzzling observations on photoelectricity, for which he received the Nobel Prize (not for E = mc2, or his revolutionary theory of relativity). Finally, a brief consideration of the firearms homicides and suicides data for Michigan (1999-2003), at the county levels, reveals a nearly PEREFCT correlation between firearms-related suicides and the county population, implying a PERFECT homogeneity of contributing factors to (firearm) suicides across population sizes and therefore the need to more effectively address this tragic and preventable cause of self-inflicted violence and deaths.
Page | 2

2. Pre-Introduction
I just came up with this unconventional subsection heading, after completing this article. It would be too much of an effort to rewrite it and incorporate the short discussion here which is based on a conversation that I had last night with a couple of young friends. I have also shared these ideas, and many other professionals who seem to be in agreement.

The crescent moon with Jupiter, appearing like a bright star in the sky, in the north, north-west direction in the early hours of the night, around 10 PM, on May 12, 2013 (in Detroit, Michigan). Taken with my Nokia cellphone camera. Heres how I tried to explain last night the basic problem with the current methodology, which uses ratio analysis, as we were driving back home with a beautiful crescent moon in the sky and with Jupiter appearing as a bright star, close to it. Let consider the firearms-related homicides data for California (CA) and Alabama (AL). In 2010, there were a total of 1342 homicides in CA and 282 in AL. The homicide rate per 100,000 population was 3.63 and 5.92, respectively. These figures are obtained from Table 4 of America Under the Gun, see Ref. [1], a recent comprehensive study of the gun violence problem in the US.
Page | 3

Firearms-related Homicides, y

1600 1400 1200 1000 800 600 400 200 0 0 50 100 150 200 250 300 350 400 450 500

CA

AL

State population, x [in 100,000s]


Figure A: Graphical illustration of the meaning of the various death rates used in gun violence studies. Here we consider the firearms-related homicides for two states California (CA) and Alabama (AL). In 2010, there were 1342 homicides in CA and 283 in AL. The homicides death rates were, respectively, 3.6 per 100,000 and 5.92 per 100,000. Alabama ranks No. 1 in this gun violence indicator, according to the GPP study. California received 4-stars (the highest) and 81 out of 100 points for the effectiveness of its gun laws (the Brady Campaign scorecard for 2011). AL, with a Brady score of 14, received a 1-star. The data can be represented graphically as illustrated in Figure A. California, with the higher population x has the higher firearms-related homicides. We do not really need an explanation for this elementary observation. The higher the population, the more the likelihood of having criminal elements in the population who will use firearms and commit violent acts such as a homicide. So far so good. The problem arises when we try to create a y/x ratio, such as the firearms homicide rate per 100,000 of population, to analyze this most
Page | 4

fundamental empirical observation. The death rate is obviously lower for CA than it is for AL. What else did we expect? After all, in the case of CA, the numerator is being divided by a large number. Can we now simply conclude that Alabama is the worst state for this gun violence indicator? Look at the graph again. The slope of the two dashed straight lines, which connect the (x, y) pairs for each state to the origin equals y/x = m, the death rate for each state. We will refer to these lines as rays passing through the origin. What we are really saying is that the (x, y) pair for AL should fall on the same line as the California. Alabama is perceived as being worse because, after duly accounting for its lower population, it has the higher homicides. This also means that if we breakdown CA into subunits with smaller populations (such as counties), all of the data would line up nicely along the California ray, the line joining the points (0,0) and (3.6, 1283) on our graph. With a moments reflection, it is clear that such a scenario is highly unlikely. One can, of course, test the argument being made here by obtaining data for individual CA counties. (I have done this for earlier traffic-related fatalities, a related but different topic altogether. Gun death rates are often now being compared to traffic fatality rates. There are as many people dying in traffic accidents as firearms-related deaths (true!), but I dont see anyone trying to take away my car, goes one line of reasoning found on an internet blogs! see also brief analysis for Michigan the counties in Appendix 2.) Likewise, it seems premature to claim that CA has a lower firearms-related homicides death rate and all of this must, somehow be related to the various gun laws that have been enacted by California. So, there it is. I have created a straw man as they say and now everyone on both sides of this debate can take their stab as they wish and argue till the cows come home as some like to say recalling ancient and more peaceful pastoral time when no one killed anyone. (Wait, we cannot forget the Abel and Cain story from the Old Testament. The first child born to Adam and Eve was killed by the younger brother! Homicides literally go back to Biblical times.) Anyway, what next? How do we proceed from here?
Page | 5

Heres my take. It is clear that it would be too hasty to vote against AL or start cheering CA. Now, take a look at the graph again. Do you see that the (x, y) pairs for CA and AL fall on a line which does NOT pass through the origin? For that matter, any two points on a x-y graph can always be joined by a straight line. The general equation of this straight line is y = hx + c, not y = mx. Hence, the ratio y/x = m = h + (c/x). The ratio y/x = m = h if and only if the intercept c = 0. The y/x ratios are different for CA and AL because of this nonzero intercept c that we failed to account when we decided to use the ratio y/x = m to make our comparison. What does this mean and what is the significance of the nonzero intercept c? This is the logical question that we must now answer. The answer lies in looking at least one more state to see where the data falls. If we were to take a state with a population between the AL and CA populations what would be find. This is exactly analogous to the suggestion made earlier that we should investigate the CA homicides data on a county-by-county basis.
1600

Firearms-related Homicides, y

1400

1200
1000 800 600 400 200 0 0 50 100 150 200 250 300 350 400 450 500

NJ NY MI

State population, x [in 100,000s]


Figure B: The data for three more states Michigan (MI), New York (NY) and New Jersey (NJ) are added here to our x-y diagram of Figure A.
Page | 6

The solid line joins the CA and AL data points and has the equation y = hx + c = 3.26x + 127.08. The slop h = 3.26 is higher than the AL slope but lower than the CA slope. Amazingly, we find the (x, y) pair for MI falls practically on this CA-AL line. The number of firearm homicides predicted for MI is 449 and the actual observed value if 440. The (x, y) pair for NY falls below the CA-AL line but the number of homicide deaths 537 is intermediate between the CA and AL numbers, as we would expect. fall in between as expected. The predicted value for NY was 759 versus the actual observed value of 537. The NJ data points fall between the AL line (taken in isolation) and the CA-AL line. NJ with a higher population has fewer homicides than AL, after accounting for this size difference. It appears therefore that more general law y = hx + c, as opposed to y = mx, or y/x = m, merits our attention. Instead of comparing states using various y/x ratios, we might gain some new insights by considering the alternative viewpoint, or the more general law y = hx + c. Often, as we just found out, when we analyze our empirical (x, y) observations, we find that the data lies up nicely along a straight that does not pass through the origin. This type of a ratio analysis is now being widely used to study various complex problems. Two important examples of recent interest are the Debt/GDP ratio problem, which has exploded into the public blogosphere, after some coding Microsoft Excel coding errors were discovered by a graduate student, in a paper published by two Harvard economists, Reinhart and Rogoff (or RR). They studied all of the available Debt/GDP ratios for 44 countries, over a period of some 200 years in some cases (USA, for example) and compared it to the average GDP growth rate to draw some controversial conclusions that call for austerity programs. Another example is the Airline Quality Rating developed by Professors Brent Bowen of the College of Technology, Purdue University and Professor Dean Headley of the Barton School of Business, Wichita State University. As we see from the pre-introduction here, these y/x ratios may be biasing our understanding of the problem. Depending on the problem being analyzed, the ratio analysis sometimes favors the larger size entity (such as states with
Page | 7

bigger population, here CA, in the gun violence study), or the smaller entity (such smaller airlines which operate fewer flights and therefore fewer opportunities for delayed flights, missed baggages, and so on).

3. Introduction
What can we learn from the x-y graph in Figure 1 presented on page 1 here? A correlation does not necessarily imply causation, as noted by Gerney, Parsons, and Posner, in the concluding paragraphs (see page 38, Ref. [2]). This cautionary note is now usually attached to all statistical analysis of large volumes of data. There are three kinds of lies: lies, damned lies, and statistics, goes a famous saying which was made even more famous by the American humorist Mark Twain. The origins of this saying are shrouded in the ancient mystery of at least two plus centuries. Hence, we will carefully avoid the use of statistical arguments here in our discussion of gun violence data. Instead, we will examine a new approach based on the idea of a work function. Figure 1 here is a simple x-y graph. The Brady Campaign for Prevention of Gun Violence publishes its annual scorecard (click here) rating the effectiveness of the gun laws enacted by various states, see Ref. [1]. On the horizontal axis we plot the Brady Campaign Score for each state, see Table 1. On the vertical axis we plot the overall firearms related death rate, which was obtained from Table 2 of the recent study entitled America Under the Gun, by Arkadi Gerney, Chelsea Parsons, and Charles Posner, Refs. [2-4]. A similar ranking of states is also provided by the Law Center to Prevent Gun Violence, see Ref. [5]. In the Brady methodology, each state is given a score, from 0 to 100, with a score of 81, the highest, earning California a four-star rating in 2011. There are states which received no stars at all, 31 of 50, and there are also states (three of them) which received a ZERO Brady score! The Law Center methodology awards a maximum of 210 points (based on 29 different criteria) but a ranking of states, similar to the Brady scorecard is not readily available, see Ref. [3].
Page | 8

What can we learn from Figure 1? The state of New Jersey with a Brady score of 72 (earning 3-stars) had an overall firearms death rate of 5.19 per 100,000 . Virginia, on the other hand, with a Brady score of 12 (earning 1-star) had an overall firearms death rate of 10.94, more than double the NJ death rate. Any two points on a x-y graph can always be joined by a straight line, with the general equation y = hx + c.

Table 1: Brady Scores and Firearms-Related Death Rates


Overall Overall Firearms State Brady Firearms Firearms Homicide Score death rate death rate rates (2010) (2010) (2001-2010) Per 100,000 Per 100,000 Per 100,000 GPP Table 4 GPP Table 2 GPP Table 3 California (CA) 81 7.88 9.1 3.6 New Jersey (NJ) 72 5.19 5.04 2.96 Massachusetts (MA) 66 4.12 3.38 1.92 New York (NY) 62 5.22 5.2 2.72 Connecticut (CT) 58 5.85 5 2.74 Hawaii (HI) 50 3.31 3.02 n/a Maryland (MD) 45 9.32 11.32 5.3 Rhode Island (RI) 44 4.66 4.23 1.62 Illinois (IL) 35 8.29 8.68 4.5 Pennsylvania (PA) 26 10.29 10.36 3.94 Michigan (MI) 25 10.89 10.83 4.45 North Carolina (NC) 16 11.78 12.54 3.94 Colorado (CO) 15 11 11.19 1.93 Oregon (OR) 15 11.95 10.8 1.57 Washington (WA) 15 9.06 8.98 1.7 Alabama (AL) 14 16.36 16.62 5.92 Minnesota (MN) 14 6.88 6.67 1.23 Delaware (DL) 13 9.8 9.2 5.01 Virginia (VA) 12 10.94 10.86 3.39 Data Sources: America Under the Gun Report (April 2013) from Center for American Progress and the Brady Campaign for Prevention of Gun Violence (2011 Scorecard)
Page | 9

The numerical value of the slope h is given by equation 1 which then yields the numerical value of the nonzero intercept c. Since the straight line passes through both the points (x1 , y1) and (x2, y2), we can use either equality given as equation 2 to determine c once h is determined using equation 1. Slope h = (y2 y1)/(x2 x1) Intercept c = (y1 hx1) = (y2 hx2) ..(1) ..(2)

There are absolutely no unknowns in this way of looking at the gun violence data. The only uncertainty, if anything, comes from the estimates of the population of the state which is used to determine the death rate per 100,000. There is no uncertainty in the Brady score. There can be no uncertainty either in the number of firearms related deaths for a state, because these are crime data and must be properly recorded and we live in a society of laws where such data are indeed duly recorded. Hence, we can say with a great deal of confidence that the overall firearms-related death rate decreases as the Brady score increases. There is no uncertainty here since we are not relying on any statistical arguments. So, we do not have to issue the correlation is not causation disclaimer. This is not meant to be rhetorical. This is simply a FACT that is being observed and analyzed here using well understood mathematical principles that even a high school or middle school student can understand and relate to. Now, take a look again at Figure 1. Besides the NJ and VA data, we find the data for three other states, Connecticut (CT), Illinois (IL), and Colorado (CO) that fall along this NJ-VA line, either slightly above or below it. Our conclusion that a higher Brady score indicates reduced guns-related death rates is not changed. This covers 5 of the 19 states that earned 1-stars to 4-stars in the 2011 Brady Campaign scorecard (click here).

4. Extension of Einsteins idea of a work function


With this background, let us now consider the data for two more states, Massachusetts (3-stars, Brady score 65) and Washington (1-star, score 15).

Page | 10

The (x, y) values are (15, 9.06) for WA and (65, 4.12) for MA. Hence, the slope h = -0.0988 and c = 10.542 for the straight line joining the WA and MA data. The slope h differs only in the third decimal place. In fact, if we change the MA score to 66 (I had mistyped as 65 as a 66 originally), the slope h = 0.0969 and is virtually identical and equal to the NJ-VA line. The only difference is the numerical value of c which is now 10.42 instead of 12.09 for the NJ-VA line. This means that the WA-MA data fall on a parallel to the NJ-VA line. This is illustrated in Figure 2.
14.00

Firearms-Deaths Rate, y

12.00 10.00 8.00 6.00 4.00

WA A y = -0.0988x + 10.54 A Perfect Mathematical Relation with ZERO Uncertainty


0 10 20 30 40 50

2.00
0.00

MA A
60 70 80 90 100

Brady Campaign Score for the State, x


Figure 2: The overall firearms-related death rate data for Massachusetts (MA) and the state of Washington (WA) has been added here and falls on a parallel to the NJ-VA line deduced earlier. The only difference is the numerical value of the constant c which is lower for the MA-WA pair.

Page | 11

Now, we also see that the data for the CT and IL is neatly sandwiched between these two parallels. If we wish, we could add more parallels to pass through each point that is not exactly on the NJ-VA line or the MA-WA line. Hereon, we will use the term work function to describe the nonzero c in the linear law y = hx + c. This idea of a work function can be understood using the analogy of baseball batting statistics, or the work function conceived by Einstein, in 1905, to explain photoelectricity (see Appendix 1).
14.00

OR MD

Firearms-Deaths Rate, y

12.00

10.00
8.00 6.00 4.00 2.00 0.00

y = -0.088x + 13.26 A Perfect Mathematical Relation with ZERO Uncertainty


0 10 20 30 40 50 60 70 80 90 100

Brady Campaign Score for the State, x


Figure 3: A third nearly parallel line to the data set can be added if we consider the Oregon (OR)-Maryland (MD) data or the North Carolina (NC)- Maryland (MD) data pairs. The slope h = 0.088 for the OR-MD pair is very nearly the same as for the NJ-VA and MA-WA pairs. This gives rise to the idea of a work function akin to that observed in baseball batting stats or the work function conceived by Einstein for the photoelectricity problem. If we analyze the game-by-game batting statistics of a baseball player, we will find that the data also falls on a family of parallels. Einsteins photoelectric law
Page | 12

also yields a family of parallels (for the energy of an electron produced when light shines on different metals). This has already been discussed in the two earlier articles on this problem, Refs.[6,7], see also the discussion section later in this article. This idea of a movement of the data along nearly perfect parallels is again supported by the Oregon (OR)-Maryland (MD) pair, with the equation y = -0.088x + 13.625, which falls above the NJ-VA line.
14.00

Firearms-Deaths Rate, y

12.00 10.00 8.00 6.00 4.00 2.00 0.00 0 10 20 30 40 50 60 70 80 90 100

Brady Campaign Score for the State, x


Figure 4: A third PERFECT parallel, passing through the MD data has been added, see dashed line. The data for Oregon (OR, 15, 11.95) and North Carolina (NC, 16, 11.78) falls slightly below this PERFECT parallel. A PERFECT parallel to the NJ-VA line, passing through the MD data point is illustrated in Figure 4. Exactly similar conclusions follow if we look at the firearms homicides rates, as opposed to the overall firearms death rates. As noted in the earlier discussion in Ref. [7], and very sadly so, suicides are a big part of the overall firearms related deaths. For example, the state of Massachusetts and
Page | 13

Washington have very nearly the same population but very different overall firearms deaths, as seen from the breakdown of the firearms-related deaths given below. It is very difficult to recover from a gunshot wound to ones own head in a suicide attempt, as opposed other methods that have been used for this tragic violence on the self. Breakdown of Overall Firearms-related deaths for MA and WA State
Overall Homicides Suicides Others

Massachusetts (2010)
270 126 138 6

Washington (2010)
609 114 404 91

Table 2: Firearm Homicides vs. Overall Firearms Deaths


State
California (CA) New Jersey (NJ) Massachusetts (MA) New York (NY) Connecticut (CT) Hawaii (HI) Maryland (MD) Rhode Island (RI) Illinois (IL) Pennsylvania (PA) Michigan (MI) North Carolina (NC) Colorado (CO) Oregon (OR) Washington (WA) Alabama (AL) Minnesota (MN) Delaware (DL) Virginia (VA)

Overall Brady Firearms Score deaths(2010) GPP Table 2


81 72 66 62 58 50 45 44 35 26 25 16 15 15 15 14 14 13 12 2935 456 270 1011 209 45 538 49 1064 1307 1076 1123 555 458 609 782 365 88 875

Firearms Homicides (2010) GPP Table 4


1342 260 126 527 98 n/a 306 17 577 501 440 376 97 60 114 283 65 45 271

Homicides % of Overall firearms deaths


45.7 57.0 46.7 52.1 46.9 n/a 56.9 34.7 54.2 38.3 40.9 33.5 17.5 13.1 18.7 36.2 17.8 51.1 31.0

Data Sources: America Under the Gun Report (April 2013) from Center for American Progress and the Brady Campaign for Prevention of Gun Violence (2011 Scorecard)
Page | 14

Opponents of gun control measures have, however, criticized this as an attempt by gun control advocates to conflate the firearms-related death by including irrelevant data. The criticism is a valid one if we use y/x ratios, the death rate per 100,000 population. This objection can, however, be overcome by considering the relative changes in the homicide rates for different states. Take California, which received 4-stars and the highest Brady score of 81 out of a possible 100, and Alabama, considered to have one of the worst gun violence records in the country (see Alabama Gun Violence Fact Sheet, page 53 of the GPP report). The aggregate GPP state rank for Alabama, after taking into account all ten gun violence indicators, was the third worst 8.6, with only Alaska (5.2) and Louisiana (5) have lower aggregate ranks (see Table 1 of GPP on page 11). These (x, y) pairs for these two states, based on their Brady scores, can again joined by a straight line with a negative slope, see Figure 5.
20.00

Firearms Homicides Rate, y

18.00 16.00 14.00 12.00 10.00 8.00 6.00 4.00 2.00 0.00 0 10

AL

y = -0.112x + 18.19 Mathematical Relation Line joining CA-AL

MI

CA

NY

20

30

40

50

60

70

80

90

100

Brady Campaign Score for the State, x


Figure 5: The Brady Campaign 2011 Score, x, for a state is plotted on the horizontal axis. The firearms-related homicide rates (per 100,000), y, is plotted on
Page | 15

the vertical axis. The (x, y) pairs for California (81, 9.1) and Alabama (14, 16.62) are joined by a straight line with the equation y = hx + c = - 0.112x + 18.19. Note again that the trend line superimposed on the x-y graph here is a purely mathematical result. This is NOT a statistical correlation. There is ABSOLUTELY NO UNCERTAINTY in the numerical values of the constants h and c in this mathematical equation relating x and y. Unlike the situation with the overall firearms-related deaths, there are no other data points that fall on this line that can be considered to be a standard of reference for this important gun violence indicator. However, the data for five other states with Brady scores spanning the range of 0 to 100 fall below this line. These are Michigan (25, 10.83) Illinois (35, 8.68), Hawaii (50, 3.02), New York (62, 5.2) and New Jersey (72, 5.04). The data reveals a great deal of scatter, and therefore a weak correlation coefficient or coefficient of determination if we use the tools of traditional statistical analysis. The downward trend in homicide rates with increasing Brady scores is, however, unmistakable. Now, take a look again at Figure 5 and consider the (x, y) pairs for MI and NJ. These two points fall on a line which is very nearly parallel to the CA-AL line. The slope h = -0.123 is quite close to h = - 0.112 for the CA-AL line but the intercept c = 13.91 is lower. Consider also the IL and NY data points. These too are seen to fall on a line that is roughly parallel to the CA-AL with a slightly different value for the nonzero intercept c. The slope h = -0.129 and c = 13.19. A family of parallels can therefore be envisioned to provide a good description of the data for this important gun violence indicator. The conclusions are again inescapable. A higher Brady score, which means more effective state gun laws, is also associated with a lower firearms-related homicide rate for the states. There is no strong correlation if we invoke the conventional methods of statistical analysis. Nonetheless, the viewpoint presented here leads to a conclusive understanding of the effectiveness of various state gun laws and their beneficial effects on reduction of the firearms-related death rates. Correlation is not causation. Can we still continue to say that?
Page | 16

4. A Non-Statistical Analysis of the Gun Violence Data


This article is actually a continuation of the discussion on gun violence data initiated in Refs. [6,7] with the attempt now being made to relate the gun violence data to the Brady Campaign scores. As we see here, the idea of a family parallels sweeping through the data set emerges quite naturally if we consider the firearms related death rate data for individual states and compare them with the Brady scores using the x-y diagrams. There is no need to involve any statistical arguments. Only very simple and basic mathematical concepts have been used here. NO STATISTICS! ABSOLUTELY NO STATISTICS!! In this context, it is worth noting that an exactly similar procedure was used by Robert Millikan to analyze the data in his Nobel Prize winning experiments that led to a confirmation of Einsteins ideas on the quantum nature of light, Refs. [8-10], see also Appendix 1. First off, Millikan was the first American Nobel laureate who was educated entirely in the US. All earlier American Nobel laureates had emigrated from Europe and so had their early education in a very different cultural environment. Second, Millikan was also an old man, relatively speaking, when he initiated the work that got him the Nobel Prize. Anyway, this is what Millikan did. Just like we have done here, he avoided statistics and still managed to determine two fundamental constants of nature, viz. i) the absolute magnitude of the electrical charge on a single electron q, and ii) the Planck constant h using arguments exactly similar to those used here. To determine the absolute electrical charge q on a single electron, Millikan did exactly what we have done here with the gun violence data. The measurements made by Millikan, in this part of his experimental program, are a little difficult to follow without an extensive background in physics. But, it is sufficient to note that no statistical arguments were used. Millikan uses only the differences in x and y values and the various slopes, as done here, to arrive at the value of q. The logic that Millikan uses to determine q is truly
Page | 17

remarkable. Anybody who wants to be called scientist must study Millikans original papers and understand how he determined q. Even economists and social scientists and medical researchers (where statistical arguments are now being widely used) can benefit from a study of Millikans (oil drop) experiments that led to the value of q. What he did next, to determine the universal constant called the Planck constant h, is very easy to follow.

Millikan graph for lithium with only two data points (see arrows) in the first of two papers on the photoelectric determination of the Planck constant h. Millikan published two papers on the latter topic, both in 1916. In the first paper, see Ref. [8] (click here), he describes his initial experiments with the alkali metal lithium (the same metal that we now find in the ubiquitous lithium-ion batteries today) aimed at testing Einsteins photoelectric equation which is quoted in the very first sentence of this paper. Millikan notes that the Planck constant can be determined from the slope of the K-f graph since, according to Einstein, K = W = hf W. He then determines h from just two (K, f) values, h = (K2 K1)/(f2 f1), just like we did to find the slope of the NJ-VA line in Figure 1, see equation 1.
Page | 18

The revolutionary idea of elementary energy quanta was conceived by Planck, in December 1900, when he derived the equation, y = mxn [e-ax /(1 + be-ax)] which is a re-statement of the famous blackbody radiation law using generalized mathematical symbols. The x-y graph reveals a maximum point. This equation can also be applied to many other problems outside physics, see, for example, Ref. [7], where a simplified version was used to discuss gun violence data. Einstein uses the version with b = 0, y = mxne-ax (with b = 0), to develop the photoelectric law which is a linear law of the type y = hx + c. In his second paper of 1916, Ref. [9] (click here), Millikan describes more detail experiments with two metals, lithium and sodium (the metal that is present in common salt and many other foods). Five different frequencies of light were used in the experiments with lithium and six different frequencies in the experiments with sodium. Tables of the raw data are still available for study and analysis in the 1916 papers, which lead to the following results, if we use linear regression analysis. For lithium, V0 = 0.4126f 3.593 in 1st 1916 paper (2 points only). For lithium, V0 = 0.4223f 3.922 in 2nd 9196 paper (5 points). For sodium, V0 = 0.4069f 4.288 in 2nd 1916 paper (6 points). Here V0 is the stopping potential determined by Millikan, see also Millikan own graph in Appendix 1. The maximum kinetic energy of the electron K = V0q where q is the absolute magnitude of the electrical charge on a single electron. This has already been determined by Millikan, in independent experiments (the oil drop experiments). Millikan, however, avoids even this level of elementary statistical analysis and finds the slopes between various points on his graph and takes an average value to deduce the best value for the universal Planck constant h. This is discussed nicely in his Nobel lecture, Ref. [10], where we also find the graph for sodium. Interestingly, Millikan never seems to have prepared a composite x-y graph with the data for both lithium and sodium.

Page | 19

Was Millikan trying to avoid leaving physicists with the impression that the lines are NOT PERFECTLY parallel, while presenting the measurements of a UNIVERSAL constant h, the Planck constant? One can only wonder. The main points here are: a) the method used to determine h from the slope of the V0-f graphs for lithium and sodium. We only find the graph for sodium Millikans 1916 paper and also in his Nobel lecture, and b) the slight difference in the slopes for sodium and lithium. Thus, we can take the trends that we now see in the gun violence data as a true movement along essentially parallel lines. The nonzero intercept c in Millikans experiments is the work function for the metal (the minimum energy or the work that is needed to bring the electron out of the metal, as Millikan states in his Nobel lecture). Likewise, a work function c appears in the number of firearms-related deaths in different states, each operating with a complex web of gun laws, and other social and cultural factors that affect gun violence and the resulting tragic deaths (suicides and homicides and the increasing number of mass murders). For those who find this discussion of physics too much to handle in this context, I suggest the discussion of baseball statistics, see Refs. [11-13]. And, for those who do not understand, or care for, American sport of baseball - I suggest Millikans experiments, again! Or just study the Debt/GDP ratios and the articles on the errors made by Harvard economists, Carmen Reinhart and Kenneth Rogoff, since the coding error made in their Microsoft Excel programs was discovered by a graduate student, Thomas Herndon, see Refs. [14-21]. Herndon also using the baseball batting average analogy [16]. Now, let us look at three of the ten gun violence indicators from the Gerney, Parsons, and Posner study using conventional least squares analysis. These results are presented in Figures 6 to 8.

Page | 20

5. Least Squares Analysis of the Gun Violence Data


In their study, GPP used ten indicators of gun violence (e.g., overall firearms death in 2010, overall firearms deaths from 2001 through 2010, firearms homicides in 2010, firearms suicides in 2010, and so on; three of the ten are listed for the selected states in Table 1) to create an overall ranking for the level of gun violence in each of the 50 US states. This is given in Table 1 of Ref. [2]. For example, the state of Louisiana (LA) received the ranks of 2, 1, 1, 18, 1, 2, 2, 4, 15, and 4, respectively, for the ten indicators. Adding these rank numbers gives total ranking points of 50 and therefore an average ranking point of 5. This makes LA the state with the worst gun violence record in the nation. After arriving at this average gun violence rank, GPP compared to the ranking of the gun law effectiveness, developed by the Law Center to Prevent Gun Violence, see Ref. [5]. GPP have indicated a preference for the Law Center ranking methodology since it is more comprehensive and consider several factors affecting gun laws. However, after a careful review of both Refs. [2] and [5], I was unable to find a simple table with gun law effectiveness ranks for each state, like the Brady Campaign 2011 scorecard. Hence, I sent an email enquiry and followed it by posting the comments in their blogs, given under Ref. [3]. In the interim, with a little internet searching, I was able to obtain quite easily the 2011 Brady Campaign Scorecard for each state [1], see the links given with Figure 5 caption. The Brady scorecard is an attempt to do exactly the same thing that is attempted by the Law Center rank different states according to the effectiveness of their gun laws. Hence, I have used the Brady Campaign scores in this article to compare the effectiveness of state gun laws rather than the Law Center ranking of the states gun laws. In Figure 6, we consider the overall firearms-related death rates data (for the year 2010) for all the states (19 total) that received 1-star to 4-stars in the 2011 Brady scorecard. Ten of these 19 states have already been considered in the analysis presented thus far where the idea of a work function was developed to explain the scatter in the observations amid the general downward trend in the gun violence death rates with increasing Brady scores.
Page | 21

The cluster of data points (for 8 states) at the lowest Brady scores, with the higher death rates is, perhaps, the clearest indication of the effectiveness, or otherwise, of gun laws. The eight are the 1-star states (NC, CO, OR, WA, AL, MN, DL, and VA) with Brady scores of 12 to 16. A best-fit line, determined using the method of least squares, Refs. [22,23] has been included. Notice that the slope h = -0.102 is very close to the slope h = -0.096 for the NJ-VA line. What is the significance of the low values of the death rates observed in some states with the low Brady scores, e.g., MN and WA which fall below the best-fit line in Figure 6? These low death rates clearly mean strong gun laws are NOT the only reason for these favorable outcomes. We cannot overlook the huge jump in the death rate for Alabama, a state that ranked among the top three worst states in the GPP study. Alabama was the third worst state in the union with an overall rank of 8.6 in the GPP study (with only Alaska, 4.2 and Louisiana, 5 being worse than Alabama). Arizona (AZ), the fourth worst state with an overall GPP rank of 10, received a Brady score of ZERO and so is not considered in the x-y graph of Figure 5. AZ ad an overall firearms death rate (2010) of 14.57 per 100,000 which put it well above the best-fit line (intercept equals 12.2 and for AZ x = 0). Thus the reason(s) for the downward trend in Figure 6 is unmistakable. The best-fit line, as Legendre explains very nicely in the introduction to his famous 1805 paper, is simply a convenient method of fitting a straight line through all these (x, y) pairs. We only need two (x, y) values to fit a straight line and deduce the trend, as was done earlier in Figures 1 to 4, without statistical arguments. Mathematically speaking, the data set here represents an over-determined system. The method of least squares is a convenient method to understand the behavior of this system. If correlation do NOT imply any type of causation, one would be hard pressed to explain why the asteroid Ceres, the largest of the asteroids, which was discovered on New Years Day 1801, was found again after it was lost, see Ref. [25,26]. The astronomer Piazzi, who discovered it (then even called a minor planet because of its size), fell ill after just 41 days of observations of its orbit. Some felt that Ceres must be lost behind the glare of the sun and will
Page | 22

be found after it re-emerges. The German mathematician, Carl Gauss, applied the method of least squares to determine the future locations of Ceres. A complex problem of celestial mechanics was reduced to a simple curve fitting problem to predict the future trajectory of Ceres. Needless to say, Ceres was found and Gauss analysis and least squares correlation played a central role!

18.00

Firearms-Deaths Rate, y

16.00
14.00 12.00 10.00 8.00 6.00 4.00 2.00 0.00 0 10

AL

Overall Firearms Death Rate (in 2010), per 100,000

MD MN Best-fit line r2 = 0.501 y = -0.102x + 12.2


20 30 40 50

CA

HI

60

70

80

90

100

Brady Campaign Score for the State, x


Figure 6: The Brady Campaign 2011 Score (maximum 100, click here) for the state is plotted on the horizontal axis. California received the highest score of 81 points with a four-star rating (the highest). The Overall Firearms-related death rate (per 100,000 population) is plotted on the vertical axis. The higher the Brady score, the lower the death rate. The data for all the 19 states in the 2011 Brady Scorecard, which received 1-star to 4-star ratings, is plotted here. Even without the aid of the best-fit line superimposed on to the data, it is obvious that the data for VA, CO, CT and NJ. Without fear of being politically INCORRECT (depending which side of the fence one is), the negative statistical correlation revealed here can be interpreted as an outcome of the effectiveness of various
Page | 23

gun laws enacted by these states. The Overall Firearms related death rate was obtained from Table 2 of America Under the Gun, by Arkadi Gerney, Chelsea Parsons, and Charles Posner of Center for American Progress (click here). Given this impressive history, it seems downright unscientific to disclaim any causation due to statistical correlations. Why study such correlations is there is no desire to predict future outcomes of complex processes and phenomena? The current gun violence study is a perfect example where one must be willing to risk inferring a causation based on a well-founded statistical correlation.

18.00 16.00

Firearms-Deaths Rate, y

14.00 12.00 10.00 8.00 6.00 4.00 2.00 0.00 0 10 20 30 40 50

Best-fit line r2 = 0.401 y = -0.098x + 12.06

Overall Firearms Death Rate (2001-2010), per 100,000


60 70 80 90 100

Brady Campaign Score for the State, x


Figure 7: The Brady Campaign 2011 Score compared with the overall firearmsrelated death rate for the ten year period 2001-2010. California with the highest score of 81 points and a four-star rating (the highest) has a lower death rate than Virginia with a score of 12 and 1-star rating. Again, the higher the Brady score, the lower the death rate per 100,000.
Page | 24

Exactly similar conclusions are also reached if we consider the data for the two other gun violence indicators listed in Table 1. The overall firearms-related death rates for the ten year period 2001-2010 are not very different from the single year (2010) death rates. Thus, not surprisingly, we see the same trend for this indicator, see Figure 7. The number of firearms related homicides is often found to be a small fraction of the overall firearms deaths, see Table 2. For 2010, the homicides varied from a low of 13.1% (for Oregon) of overall firearms-related deaths to a high of 57% (for New Jersey). However, the same downward trend is again revealed, see Figure 8, when the homicide rate (per 100,000 population) is compared to the Brady score.

Firearms Homicides Deaths Rate, y

7.00 6.00 5.00 4.00 3.00 2.00 1.00 0.00 0 10 20 30 40 50 60 70 80 90 100

Best-fit line r2 = 0.323 y = -0.0299x + 4.93

Brady Campaign Score for the State, x


Figure 8: Another gun violence indicator, the firearms-related homicides death rate (in 2010) is considered here and compared to the Brady score for the state. Again, with increasing Brady scores we find a decline in the death rate. The cluster of four data points (with death rate < 2.00) for the states of CO, OR, WA,
Page | 25

and MN was neglected to determine the best-fit line. Including them reduces the slope and this would imply that gun laws for some states may actually be more effective in reducing deaths due to non-homicide events (such as suicides, for example, or mass murder rampages). In summary, for all the three gun violence indicators examined here, the higher the Brady score (i.e., the more effective the states gun laws according the Brady Campaign scorecard), the lower overall firearms-related death rates observed here. The trends revealed in this section seem to be telling us that enacting strong gun laws, and learning from the experience of such states (as well as the states with the weakest gun laws) will lead to reductions (as much as a factor of 2 in many cases) in the firearms related death rates.

Page | 26

Appendix 1 The Work function and Einsteins Photoelectric law


In his famous 1905 paper on the nature of light radiation, Einstein first showed that visible light can be thought of as a stream of particles (now called photons), each having Plancks elementary energy . Hence, when light shines on the surface of a metal, like lithium (used in the lithium-ion batteries of the modern era) or sodium (common salt is sodium chloride), it produces electrons. Modern photocells work on this principle. Some work must be done to overcome the forces that bind the electron to the metal. Hence, Einstein said that the maximum energy of the electron K = W where W is the work function, the energy that must be given up to produce the electron. Since = hf where h is a universal constant, called the Planck constant, and f is the frequency of light, this means that K = hf W = h(f f0) where f0 = W/h is the minimum frequency, or the cut-off frequency, below which no electrons are produced.

http://www.google.com/imgres?imgurl=http://www.kcvs.ca/site/images/photoelectric.gif&imgrefurl=http: //www.kcvs.ca/site/projects/physics.html&h=143&w=197&sz=6&tbnid=OaTLQpzeZFQLbM:&tbnh=90&tbn w=124&zoom=1&usg=__2Q7bzy57s2Jiin_Jl4DxoEcaiig=&docid=ZuuYk0TGPmMYM&sa=X&ei=UmVqUbC5KtOt0AGZsIH4Cw&ved=0CD4Q9QEwAw&dur=225

Page | 27

Courtesy: http://coraifeartaigh.files.wordpress.com/2010/04/planck_photoelectric_1.gif http://panda.unm.edu/Courses/Fields/Phys491/Notes/PhotoelectricEffect.pdf

This can be rewritten as y = hx + c = h(x x0) where x is the frequency and y the maximum kinetic energy of the electron and x0 = -c/h is the intercept made on the x-axis (horizontal axis). The K-f graph is thus a straight line with a slope h and an intercept f0 on the frequency axis (plotted on the x-axis). Experiments with different metals will yield a series of parallels with a slope h and various values of the intercept f0 = W/h which depend on the work function of the metal, which is a measure of the difficulty of producing the electron, as mentioned by Millikan in his Nobel lecture, Ref. [10]. The photoelectric effect was first discovered, in the form of a curious spark, by Heinrich Hertz, during the course of his experiments aimed at producing radio waves, or electromagnetic waves, in the laboratory using electrical equipment. These investigations engaged the attention of physicists in the late 19th and early 20th centuries. Phillip Lenard, Hertzs assistant, followed up on the curious spark and discovered the cut-off frequency. Lenard received the Nobel Prize in 1905, the same year that Einstein conceived the photoelectric law K = W = hf W = h(f f0). Einsteins explanation for Lenards cut-off frequency f0 = W/h is based on the quantum nature of light.

Page | 28

On Cathode Rays, by Phillip Lenard, Nobel lecture, http://www.nobelprize.org/nobel_prizes/physics/laureates/1905/lenardlecture.pdf As Millikan notes in his 1916 paper on the determination of the Planck constant h, Einsteins almost reckless hypothesis of light quanta flies in the face of established facts such as interference (of light waves, which produce light and dark interference patterns) and was apparently made solely to furnish an explanation for some remarkable facts brought to light by recent investigations, viz., the energy of an electron thrown out of a metal by ultraviolet light, or X-rays, is independent of the intensity of light while it depends on its frequency. Although the focus has always been on the universal constant, the Planck constant, given by the slope h, the nonzero intercept, or the cut-off frequency discovered by Lenard, is also of great fundamental importance and can lead to an understanding of many problems outside physics. The cut-off frequency f0 = W/h is a manifestation of the property called the work function W of a metal. Each metal offers a different environment for the interaction between the photon (with energy ) and the electron. To bring the electron out of the metal, and overcome the forces binding the latter to the metal, the photon must have the minimum energy (W = 0 = hf0). No electrons can be produced if the photon energy is less than this cut-off or threshold. Likewise, profits are only produced when a threshold or minimum revenue level is exceeded. Billionaires only appear in a population above a certain population level. Unemployed appear in a population only when the labor force exceeds a minimum level. Gun violence deaths are likewise related to a minimum population level, as revealed by the linear law y = hx + c, which has been discussed in detail in Refs. [6,7]. Supply and demands laws, the basic laws of economics, are also often depicted as linear laws with a nonzero intercept c, which can also interpreted as a manifestation of a work function in economics.

Page | 29

Movement Along Parallel Lines


Einsteins photoelectric law is, perhaps, the only natural law observed where one can conclude a segregation of the data along a set of parallel lines. This has been discussed already in Ref. [6] and some further notes are added here in Appendix 1. Einsteins law means that the K-f graph is a series of parallels with the slope h equal to the Planck constant. Experiments with a single metal, such as sodium, see Millikans graph in Appendix 1, should yield a straight line with the slope h. However, if the data for different metals is plotted on the same graph, a series of parallel lines with observed. Millikan confirmed Einsteins law with experiments on two metals, lithium and sodium. Unfortunately, he never prepared a composite graph to illustrate the segregation of the data along two parallel straight lines. An analysis of Millikans original data (given in the form of tables in his 1916 paper) is still possible and lead to the following results for lithium and sodium. For lithium, V0 = 0.4126f 3.593 in 1st 1916 paper (2 points only). For lithium, V0 = 0.4223f 3.922 in 2nd 9196 paper (5 points). For sodium, V0 = 0.4069f 4.288 in 2nd 1916 paper (6 points). Here V0 is the stopping potential determined by Millikan, see also Millikan own graph in Appendix 1. The kinetic energy of the electron K = V0q where q is the absolute magnitude of the electrical charge on a single electron. This has already been determined by Millikan, in independent experiments (known as the oil drop experiments, Millikan gives a nice account of this in his 1924 Nobel lecture). Notice the small differences in the slopes for sodium and lithium and also the limited number of data points. Light of five different frequencies were used with lithium and six different frequencies with sodium. Of course, Millikan was using visible light of different colors. Hence, the frequency range is limited. However, these limited observations, on two metals, were enough to confirm Einsteins photoelectric law and the Planck idea of an elementary energy quantum.

Page | 30

Physicists never looked back to check the validity of the theory or the numerical value of the Planck constant h. Millikan himself wanted to test Einstein equation with a third metal, potassium but these experiments were never pursued at some problems developed with the laboratory apparatus used in the experiments. The results with lithium and sodium were apparently satisfactory enough, even for Millikan, who started out more as a critique of Einstein wanting to prove that Einstein was wrong about his light quanta. Now, if we accept the universal validity of the linear law, y = hx + c, relating various data pertaining to gun violence (see also Refs. [6,7]), then the only inescapable conclusion is that there must be a family of parallel lines sweeping through the x-y diagram. This would explain the apparent scatter seen when we consider the data for all the states in the gun violence problem. Other examples are the Airlines Quality Ratings (AQR) problem, Refs. [36-39]. The AQR scores for different airlines are developed following exactly the methods used to arrive at the aggregate rank for each state after consider ten gun violence indicators.

Appendix 2: Michigan Firearm Homicides and Suicides (Average number for the years 1999-2003)
County/City Homicides Rate Suicides Rate Berrien 6 3.5 8 4.9 Calhoun 4 2.9 8 6.1 Genesee 28 6.4 23 5.3 Ingham 6 2.1 12 4.4 Macomb 13 1.6 40 4.9 Oakland 26 2.2 54 4.5 Saginaw 12 5.6 11 5.4 Wayne (out-county) 42 3.8 60 5.3 Detroit 304 32.5 41 4.3 Total 494 4.9 540 5.4 Data source: Firearms Homicide and Suicides Michigan, February 2006 http://www.michigan.gov/documents/Firearm_Homicide__Suicide_Report_162 746_7.pdf The other counties reported homicides/suicides of 0, 1, 2, and 3 and accounted for the difference.

Page | 31

If we ignore Detroit city (304 out of 494 homicides, about 60%), a general upward trend in homicides is observed with increasing population. Detroit City is no longer an exception when it comes to suicides. A PERFECT linearity is observed for suicides with increasing population, described by the best-fit equation y = 4.669x + 1.028 with r2 = 0.9778, see Figure 9.
70

Firearms-related Suicides, y

60

50

40

30

20

10

y = 4.67x + 1.03 r2 = 0.9778 Michigan counties (1999-2003)


0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0

County population, x [in 100,000s]


Figure 9: Another gun violence indicator, the firearms-related suicides death rate for Michigan counties (1999-20003) is considered here. The average number of suicides in each county and the suicide rate is reported which gives the average population to prepare this x-y plot. The firearms-suicides data for Michigan including Detroit city, reveals a strong correlation with population levels, with a nearly PERFECT value for the correlation coefficient (r), or coefficient of determination (r2). There is considerably more scatter in the graph for homicides. In other words, the
Page | 32

work function for suicides is essentially constant across county population sizes. This also means that a lot more can clearly be done with more effective gun laws, and also with health care initiatives to address mental health issues, to reduce this tragic and preventable cause of self-inflicted violence. The effect of the nonzero intercept c, or the work function, is also revealed in the variation in the firearms-related suicide rate, across different Michigan counties. This is not a random variation but a very systematic one, following the hyperbolic law, y/x =h + (c/x), which is a consequence of the linear law y = hx +c relating suicides y and the county population x.

8.00

Firearms-related Suicide rate, y/x Per 100,000

7.00 6.00 5.00 4.00 3.00 2.00 1.00 0.00 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0

y/x = 4.67 + (1.03/x) r2 = 0.9778 Michigan counties (1999-2003) Hyperbolic law

County population, x [in 100,000s]


Figure 10: The firearms-suicide rates for Michigan counties (1999-20003) is seen to decrease systematically with increasing population following the hyperbolic law y/x = h + (c/x).
Page | 33

Reference List
1. Brady Campaign to Prevent Gun Violence, 2011 Scorecards, http://www.bradycampaign.me/sites/default/files/2011_Brady_Campaig n_State_Scorecard_Rankings.pdf 2. American Under the Gun, http://www.americanprogress.org/wpcontent/uploads/2013/03/AmericaUnderTheGun.pdf Full report here. 3. A 50 State Analysis of Gun Violence and its Link to Weak State Gun Laws, by Arkadi Gerney, Chelsea Parsons, and Charles Posner, April 4, 2013, Center for American Progress Report, Brief discussion here, http://truth-out.org/news/item/15524 I was able to use the comments section and posted the following yesterday (May 10, 2013) with my enquiry about the Law Center state ranks for gun law effectiveness, after I had completed my analysis described in Refs. [1, 2].

Vj Laxmanan 9 hours ago (as of 4:30 AM on May 11, 2013)

I already posted one comment on this topic. Here's one more for clarity. To quote from above, "Finally, we compare this overall state gun-violence ranking with a Law Center to Prevent Gun Violence ranking of states based on the strength of their gun laws." I am looking for a nice table which gives the "Law Center to Prevent Gun Violence ranking of states based on the strength of their gun laws." I found nice table for each of the 10 gun violence indicators but cannot find anything like that for the above on strength of gun laws. If there is no table already, perhaps, one should be created like the other 10 tables for gun violence indicators.

Vj Laxmanan 9 hours ago (as of 4:30 AM on May 11, 2013)

Page | 34

Can someone point me to a table which gives the ranks (or points) for each state for effectiveness of gun laws. I was unable to find such a table either in the study America Under the Gun or at the Law Center website (Gun Laws Matter). All I got was the methodology for ranking the effectiveness of gun laws (210 points total) but no tabulation of the ranks for each state. Thanks and appreciate any feedback.

4. America Under the Gun, by Arkadi Gerney, Chelsea Parsons, and Charles Posner, April 2, 2013, http://www.americanprogress.org/issues/civilliberties/report/2013/04/02/58382/america-under-the-gun/ Brief discussion of the Report. 5. Law Center to Prevent Gun Violence, http://smartgunlaws.org/ Click on Studies & Statistics to get methodology for ranking states based on the effectiveness of gun laws (29 factor considered, maximum points 210). However no table of state ranks or points is available. 6. Gun Death Statistics and the Method of Least Squares and the Forgotten Property of a Straight Line, Published May 8, 2013, http://www.scribd.com/doc/140152581/Gun-Death-Statistics-and-theMethod-of-Least-Squares-and-the-Forgotten-Property-of-a-Straight-line 7. Comparison of the Strong and Weak Gun Law States and the Ten States with Highest Level of Gun Violence: Least Squares Analysis of the Data, Published May 10, 2013, http://www.scribd.com/doc/140536622/Comparison-of-the-Strong-andWeak-Gun-law-States-and-the-Ten-States-With-Highest-Levels-of-GunViolence-Least-Squares-Analysis-of-the-Data 8. Einsteins Photoelectric Equation and the Electromotive Force, by R. A. Millikan, Phys. Rev. 1916, Vol. VII, No. 1, Second Series, pp. 18-32. http://www.ffn.ub.es/luisnavarro/nuevo_maletin/Millikan_1916_1.pdf 9. A Direct Photoelectric Determination of Plancks h, R. A. Millikan, Phys. Rev. 1916, vol. VII, No. 3, pp. 355-390. http://mapageweb.umontreal.ca/leonelli/PHY3320/millikan.pdf 10. The Electron and the Light Quanta from the Experimental Point of View, by Robert A. Millikan, Nobel lecture, May 23, 1924, see pages 61 to 63. The graph for sodium is on page 63. The oil drop experiment is discussed in the first part of the lecture. An analysis of the experiment described in Table 1 (page 57) shows that in this experiment, which Millikan must obviously be very proud of, there was only ONE SINGLE
Page | 35

11. 12.

13.

14.

ELEMENTARY UNIT OF ELECTRICAL CHARGE attached to the drop. In other experiments, many units of charge were attached (Nq where N is the number of units). Millikan correlated the differences in the velocities of the drop (as it rises or falls under the combined action of gravity and electric fields) to the sudden changes in N, from N1 to N2. Millikan gives a fascinating account of this discovery and the analysis that follows is indeed remarkable and insightful. Everyone who wants to under science, and apply scientific methods, must study Millikans 1909, 1911, and 1913 papers on this subject. Babe Ruths 1923 Batting Statistics and Einsteins Work Function, Published April 17, 2013, http://www.scribd.com/doc/136489156/BabeRuth-s-1923-Batting-Statistics-and-Einstein-s-Work-Function Babe Ruth Batting Statistics and Einsteins Work Function, To be Published April 17, 2013, http://www.scribd.com/doc/136556738/BabeRuth-Batting-Statistics-and-Einstein-s-Work-Function The Method of Least Squares: Predicting the Batting Average of a Baseball Player (Hamilton in 2013), Published May 7, 2013, http://www.scribd.com/doc/139924317/The-Method-of-Least-SquaresPredicting-the-Batting-Average-of-a-Baseball-Player-Hamilton-in-2013 Growth in a Time Before Debt, by Carmen Reinhart and Kenneth S. Rogoff, http://www.nber.org/papers/w15639.pdf NBER Series, Working Paper 15639. On page 11, Over the past two centuries, debt in excess of 90 percent has typically been associated with mean growth of 1.7 percent versus 3.7 percent when debt is low (under 30 percent of GDP), and compared with growth rates of over 3 percent for the two middle categories (debt between 30 and 90 percent of GDP). Of course, there is considerable variation across the countries, with some countries such as Australia and New Zealand experiencing no growth deterioration at very high debt levels. It is noteworthy, however, that those high-growth high-debt observations are clustered in the years following World War II. And, on page 23, Why are there thresholds in debt, and why 90 percent? This is an important question that merits further research, but we would speculate that the phenomenon is closely linked to logic underlying our earlier analysis of debt intolerance in Reinhart, Rogoff,
Page | 36

15.

16.

17.

18.

19.

20.

and Savastano (2003). As we argued in that paper, debt thresholds are importantly country-specific and as such the four broad debt groupings presented here merit further sensitivity analysis. A general result of our debt intolerance analysis, however, highlights that as debt levels rise towards historical limits, risk premia begin to rise sharply, facing highly indebted governments with difficult tradeoffs. Does High Public Debt Stifle Economic Growth? A Critique of Reinhart/Rogoff, by Thomas Herndon, Michael Ash, and Robert Pollin, http://www.peri.umass.edu/236/hash/31e2ff374b6377b2ddec04deaa63 88b1/publication/566/ April 15, 2013. Guest Post: The Grad Student who took down Reinhart-Rogoff Explains Why Theyre Fundamentally Wrong, Business Insider, by Thomas Herndon, April 22, 2013, http://www.businessinsider.com/herndon-responds-to-reinhart-rogoff2013-4 Guest Post, Reinhart/Rogoff and Growth in a Time Before Debt, by Arindrajit Dube, April 17, 2013, The Next New Deal, The Roosevelt Institute, http://www.nextnewdeal.net/rortybomb/guest-postreinhartrogoff-and-growth-time-debt See also references to internet blogs cited by Dube. The Method of Least Squares: The Debt-GDP Relation for the Trillionaire Club of Nations, Published May 4, 2013, http://www.scribd.com/doc/139348541/The-Method-of-Least-SquaresThe-GDP-Debt-Relation-for-the-Trillionaires-Club-of-Nations An MIT Non-Economists View of the Harvard-UMass Debt/GDP Ratio and Economic Growth Debate, Published April 26, 2013, http://www.scribd.com/doc/138076426/An-MIT-Non-Economist-s-Viewof-the-Harvard-UMass-Debt-GDP-Ratio-and-the-Economic-Growth-Debate Iceland Votes Against Austerity: Analysis of Icelands Debt-GDP, Published April 28, 2013, http://www.scribd.com/doc/138345921/IcelandVotes-Against-Austerity-Analysis-of-Iceland-s-Debt-GDP-Data-2002-2012

21. A Brief Survey of the Debt-GDP Relations for Some Modern 21st Century Economies, Published May 1, 2013,

Page | 37

http://www.scribd.com/doc/138912093/A-Brief-Survey-of-the-DebtGDP-Relationship-for-Some-Modern-21st-Century-Economies 22. Legendre, On Least Squares, English Translation of the original paper http://www.york.ac.uk/depts/maths/histstat/legendre.pdf 23. Line of Best-Fit, Least Squares Method, see worked example given http://hotmath.com/hotmath_help/topics/line-of-best-fit.html The formula for h used in this example is an actually approximate one and was used, before the advent of modern computers, since it only involves the determination of x2 and xy and the sum of all the values of x, y, x2 and xy. The exact formula, is given below, with xm and ym denoting the mean or average values of x and y in the data set, and ym = hxm + c since the bestfit line always passes through the point (xm , ym). h = (x xm)(y ym)/ (x xm)2 Determine the deviations of the individual x and y values from the mean, or average, (x xm) and (y ym). Determine the product (x xm)(y ym) and their sum. This gives the numerator in the expression for h. Determine the square (x xm)2 and the sum. This gives the denominator in the expression for h. This also fixes the intercept c via ym = hxm = c . Then, using the regression equation, determine the predicted value yb on the best-fit line and the vertical deviation (y yb) and the squares (y- yb)2. The sum of these squares is a minimum. This can be checked by assigning other values for h (using any two points) and allowing the graph to pivot around (xm, ym). The regression coefficient r2 = 1 - { (y- yb)2 / (y- ym)2 } is a measure of the strength of the correlation between x and y (or y/x versus x). For a perfect correlation, when all points lie exactly on the graph, r2 = +1.000. 24. The Discovery of Ceres, Johannes Kepler and the door to science, http://www.keplersdiscovery.com/Asteroid.html 25. How Gauss determined the orbit of Ceres, by Jonathan Tennenbaum and Bruce Director, http://www.schillerinstitute.org/fid_9701/982_orbit_ceres.pdf The discussion starting with page 79 on the method used by Gauss is noteworthy.
Page | 38

26. Asteroids at Mathematics of Planet Earth 2013, by Professor Robert Miller, http://mpe2013.org/2013/02/20/asteroids/ 27. Savante Arrhenius, http://en.wikipedia.org/wiki/Svante_Arrhenius The Nobel laureate (in chemistry) published a paper in 1896 which showed that (if the earth were in complete radiative equilibrium with the sun, applying the newly developed mathematical laws of radiation due to Stefan and Boltzmann) all of the earths ocean would be frozen solid if it were not for the small concentration of carbon dioxide in its atmosphere. Then, the famous astronomer Carl Sagan, then a young PhD student, convinced NASA to send probes to Venus to measure the surface temperatures of the planet. This led to the theory of a runaway greenhouse effect on Venus. The fate of Venus awaits the Earth, said Sagan, if we do not control rising carbon dioxide (and other greenhouse) gases; see also Ref. [31]. 28. The Discovery of Global Warming, http://www.aip.org/history/climate/Venus.htm 29. Carbon dioxide has reached record levels: Disaster awaits many countries, May 10, 2013, Posted by Ray, http://reinep.wordpress.com/tag/carl-sagan/ 30. Atmospheric Carbon dioxide variations in at Mauna Loa Observatory, Hawaii, First published online March 18, 2010, http://onlinelibrary.wiley.com/doi/10.1111/j.21533490.1976.tb00701.x/abstract 31. Stephen Hawking and Carl Sagan on Global Warming: The Hell of Venus A Valuable Reminder to Take the Increasing Greenhouse Effect on Earth Seriously, by Christine, July 23, 2010, in 350 or Bust, http://350orbust.com/2010/07/23/carl-sagan-and-stephen-hawking-oneffects-of-global-warming-the-runaway-greenhouse-effect-on-venus-is-avaluable-reminder-to-take-the-increasing 32. Did the contentious global warming hockey stick graph get it right after all? By Geoffery Lean, The Telegraph, April 26, 2013, http://blogs.telegraph.co.uk/news/geoffreylean/100213990/did-thecontentious-global-warming-hockey-stick-graph-get-it-right-after-all/

Page | 39

33. New graph shows unprecedented global warming over the last 11,000 years, March 26, 2013, http://kottke.org/13/03/new-graph-showsunprecedented-global-warming-over-past-11000-years 34. Ten Charts that Make Clear the Planet Just Keeps Warming, by Joe Romm, October 15, 2012, http://thinkprogress.org/climate/2012/10/15/1014151/ten-charts-thatmake-clear-the-planet-just-keeps-warming/ 35. Airline Quality Report: An Analysis of On-Time Percentages, Published April 18, 2013, http://www.scribd.com/doc/136760664/Airline-QualityReport-2013-Analysis-of-the-On-Time-Percentages 36. Airline Quality Rating 2013, Purdue University, e-Pubs, April 8, 2013, by Dr. Brent D. Bowen (Purdue University, College of Technology) and Dr. Dean E. Headley (Wichita State University, W. Frank Barton School of Business) http://docs.lib.purdue.edu/aqrr/23/ 37. Airline Quality Report 2013: An Analysis of On-Time Percentages, Published April 18, 2013, http://www.scribd.com/doc/136760664/Airline-Quality-Report-2013Analysis-of-the-On-Time-Percentages 38. Bibliography, Articles on Extension of Plancks Ideas and Einsteins Ideas beyond physics, Compiled on April 16, 2013, http://www.scribd.com/doc/136492067/Bibliography-Articles-on-theExtension-of-Planck-s-Ideas-and-Einstein-s-Ideas-on-Energy-Quantum-totopics-Outside-Physics-by-V-Laxmanan 39. Gun Control 2013: Suicide Stats Irrelevant to Gun Control Policy, Matt MacBradaigh, in Politics, May 6, 2013, http://www.policymic.com/articles/38391/gun-control-2013-suicidestats-are-irrelevant-to-gun-control-policy 40. Gun crime statistics by US state: latest data, Datablog, Posted by Simon Rogers, December 17, 2012, http://www.guardian.co.uk/news/datablog/2011/jan/10/gun-crime-usstate Total firearm murders and the firearm murder rates (per 100,000 population) for all states is given here.

Page | 40

About the author V. Laxmanan, Sc. D.


The author obtained his Bachelors degree (B. E.) in Mechanical Engineering from the University of Poona and his Masters degree (M. E.), also in Mechanical Engineering, from the Indian Institute of Science, Bangalore, followed by a Masters (S. M.) and Doctoral (Sc. D.) degrees in Materials Engineering from the Massachusetts Institute of Technology, Cambridge, MA, USA. He then spent his entire professional career at leading US research institutions (MIT, Allied Chemical Corporate R & D, now part of Honeywell, NASA, Case Western Reserve University (CWRU), and General Motors Research and Development Center in Warren, MI). He holds four patents in materials processing, has co-authored two books and published several scientific papers in leading peer-reviewed international journals. His expertise includes developing simple mathematical models to explain the behavior of complex systems. While at NASA and CWRU, he was responsible for developing material processing experiments to be performed aboard the space shuttle and developed a simple mathematical model to explain the growth Christmas-tree, or snowflake, like structures (called dendrites) widely observed in many types of liquid-to-solid phase transformations (e.g., freezing of all commercial metals and alloys, freezing of water, and, yes, production of snowflakes!). This led to a simple model to explain the growth of dendritic structures in both the groundbased experiments and in the space shuttle experiments. More recently, he has been interested in the analysis of the large volumes of data from financial and economic systems and has developed what may be called the Quantum Business Model (QBM). This extends (to financial and economic systems) the mathematical arguments used by Max Planck to develop quantum physics using the analogy Energy = Money, i.e., energy in physics is like money in economics. Einstein applied Plancks ideas to describe the photoelectric effect (by treating light as being composed of particles called photons, each with the fixed quantum of energy conceived by Planck). The mathematical law deduced by Planck, referred to here as the generalized power-exponential law, might
Page | 41

actually have many applications far beyond blackbody radiation studies where it was first conceived. Einsteins photoelectric law is a simple linear law and was deduced from Plancks non-linear law for describing blackbody radiation. It appears that financial and economic systems can be modeled using a similar approach. Finance, business, economics and management sciences now essentially seem to operate like astronomy and physics before the advent of Kepler and Newton. Finally, during my professional career, I also twice had the opportunity and great honor to make presentations to two Nobel laureates: first at NASA to Prof. Robert Schrieffer (1972 Physics Nobel Prize), who was the Chairman of the Schrieffer Committee appointed to review NASAs space flight experiments (following the loss of the space shuttle Challenger on January 28, 1986) and second at GM Research Labs to Prof. Robert Solow (1987 Nobel Prize in economics), who was Chairman of Corporate Research Review Committee, appointed by GM corporate management.

Cover page of AirTran 2000 Annual Report


Can you see that plane flying above the tall tree tops that make a nearly perfect circle? It requires a great deal of imagination to see and to photograph it.

Page | 42

Vous aimerez peut-être aussi