Personal Reflections on the Mars Effect Controversy

Mistaken Escalation

The very name "Mars Effect" is a misnomer. The controversy centrally surrounds data purporting to show evidence for a statistically significant and non-chance correlation between persons emerging as sports champions and having Mars in certain positions at the time of birth. But, alas, both the Gauquelins and their critics have treated this correlation as though it demonstrated a causal relationship. Such a causal relationship would indeed be highly extraordinary and at least to some degree lend support to the notion of planetary "influence" as found in astrology. Thus, such a causal claim is not merely implausible to most scientists, it represents a kind of aid and comfort to what most scientists consider to be a pseudoscience that was supposedly discredited by astronomy. Gauquelin in effect reopened an old wound from an old battle which the astronomers thought they won long ago. Yet, in fact, this greatly exaggerates the evidence Gauquelin has actually put forward. It is fundamental that a correlation may be valid while due to any number of third factors; and Gauquelin has merely demonstrated (at best) the existence of a Mars Correlation (rather than Effect). Seen in this light, his evidence is really not that extraordinary at all. It does, of course, remain an anomaly, and it may be worthwhile to pursue its causes; but the evidence claimed really generates great excitement and passion if we prematurely leap to the conclusion that its validity demonstrates a causal connection supportive of astrology.

Why, then, has this anomalous correlation been persistently presented as though it necessarily was linked to astrology? In large part, of course, it is due to Gauquelin's work having originally centered around his empirical investigations into the claims of traditional astrology. The Gauquelins themselves do think that they are dealing with some sort of mysterious causal connection and the possible birth of a new science they term neoastrology. However, Gauquelin has consistently acknowledged the he might be quite wrong, that the correlations he finds may be due to some mysterious third factor we have not yet been able to isolate and establish. But, perhaps more significantly, some (but by no means all) astrologers have greeted Gauquelin's claim of a Mars Effect with great joy and have offered him support for his research (e.g., the use of computers), while his fellow "normal" scientists have met his claims with either indifference or severe antagonism (far beyond what one normally obtains from critical peers in science). So, Gauquelin's alliance with some astrologers is understandable and largely the result of his having to find support for his work where he can. But the great irony in all this is that the vast majority of Gauquelin's work is severely damaging to traditional astrology, and he has always made this quite clear. If Gauquelin had not had the "misfortune" to stumble on a few anomalous correlations during his research into the claims of the astrologers, he would today be hailed by his critics as the greatest debunker of astrology science has so far produced. Even his critics today must admit that his empirical research constitutes the best body of evidence against traditional astrology to be found anywhere (And, of course, Gauquelin's vast evidence against astrology has not received anything like the criticism levelled against the similarly derived evidence for the Mars Effect.).

Although we can readily understand Gauquelin's framing his anomalous finding in the context of astrology, the more interesting question for the sociologist of science is: Why have the critics inflated this "mere correlation" into an "effect" which thereby increases the extraordinariness of the claim rather than diminishes it? Certainly, two functions emerge from such an escalation. (1) By increasing the extraordinariness of the claim, one can call for stronger evidence for it than Gauquelin has to offer; if degree of proof must be commensurate with the extraordinariness of the claim, this places a greater burden of proof on Gauquelin. (2) The importance of the critics' challenge is increased if the object being attacked is more important. This has two components. (a) There is little point in attacking a mere correlation which results only in discrediting a minor anomaly. The Mars Correlation is only significant as an issue when it is tied to the claims of the astrologers, the real targets of the critics who have rallied to "fight pseudoscience." And (b) by attacking astrology and tapping the hostility against that "irrational superstition," the critics make themselves newsworthy and will be viewed as "heroes" by those want to see such "nonsense" crushed and the public "educated." But such escalation of the basic claim is really inexcusable for scientific critics, whose first order of business in dealing with an extraordinary claim should be to seek to minimize its extraordinary character wherever that can reasonably be done. A proponent of an anomaly may have good reasons (both practical and theoretical) for interpreting his anomaly as of maximum significance; but the critics of an anomaly have an obligation to "cut it down to size" so as not to exaggerate its importance. Here the opposite seems to have happened, and this has resulted in charges that the data must be spurious because the claim was painted as incredible. By inflating the claim, it becomes all the more important, too, to discredit it. This seems to have resulted in irresponsible attempts to debunk the claim.

The Status of the Claim

The current evidence strongly indicates that (a) a Mars Correlation was validly found by the Gauquelins, (b) a correlation was found in several replications by the Gauquelins using different samples, (c) a similar correlation was found in replications by the CP and the first study conducted by Kurtz-Zelen-Abell (KZA). In regard to (a) and (b), the key question concerns the validity of the Gauquelins' data. It has repeatedly been incorrectly stated that there is no way to check this data. Not only have the Gauquelins published all their data (so computations can easily be checked), they have kept all original records from the birth registries, and these have been made available to any serious researchers. In fact, the Gauquelins have urged critics to check this data (see Gauquelin's proposal in this issue of ZS). It should also be noted that checking the data would not necessitate checking every single case; one needs to look only at the sports champions for whom there was a Mars Correlation (only about 23% of the sports champions). However, this is largely a false issue. Based only on Gauquelin's work, a critic can properly say that the evidence is simply not strong enough until there are independent replications, and this does not necessitate calling the Gauquelins' data into question. In science, we can take an agnostic, wait-and-see position and speak of evidence as "unconvincing" rather than leap to the conclusion that something spurious is present. If you don't believe that what someone tells you he saw actually happened, you may think him mistaken and insist on further witnesses without calling the original narrator a liar. Once a critic claims Gauquelin's work is spurious, then that critic has the burden of proof on him to show how it is spurious. That simply has not yet been done successfully, so we must proceed on the presumption that the Gauquelins' data is legitimate. The claims put forward by the Gauquelins are falsifiable, but some critics' vague and unsupported accusations of spurious data are not. All statements claiming scientific status must be falsiflable, whether made by proponents or critics.

Although the CP and KZA may wish to reanalyze their data in a way that questions the existence of a Mars Correlation, the fact remains that both their studies replicated Gauquelin's own work. That is, if the analysis applied to Gauquelin's studies is applied to the data used by the CP and by KZA, those results support the claim of a Mars Correlation. In the case of the CP, they questioned the theoretical chance expectation level used by Gauquelin in his own work and concluded that their study did not therefore show evidence for a Mars Effect. But it is apparent that this is post hoc reasoning on their part since the same reasoning could have been applied to Gauquelin's own work in the first place. If they had originally reasoned this way, they would never have felt it necessary to replicate Gauquelin's work at all. They could have explained his results away the same way they explained away their own. (Of course, this does not prove that the CP criticism of the chance level used is incorrect; but it does demonstrate that it was a criticism that must have been "realized" only after they got results that otherwise would force them to admit support for the Mars Effect.)

In the first KZA test, it is clear that the total sample they used did show the Mars Effect. Only through reanalysis of that data (sample splitting, etc.) can that data be made to fall short of a demonstration of the claimed correlation (a reanalysis that Kammann, Dennis Rawlins, and others have criticized). Though KZA reach different conclusions from Gauquelin (as did the CP), their data results unequivocally constitute a replication of Gauquelin's own finding. They did not replicate his analysis, but they certainly did replicate his own data pattern. (We should not confuse replication of a study with replication of the data results, and we should not confuse replication of data results with replication of the analysis of such data.)

So, whether or not such replications have faults similar to those that might be present in Gauquelin's early studies, they do indeed constitute replications of his work. All these replications may be flawed just as Gauquelin's work may be flawed, but the critics who have repeatedly asserted that Gaujuelih's work has not been replicated by others simply misstate the facts. And the fact that Supposed "flaws" found during these replications --resulting in new analyses to discount the Mars Correlation -- follow rather than precede the data these critics have produced but which support the Mars Correlation, suggests strongly that the critics are really rationalizing away embarrassing findings. Clearly these replications were initiated by the critics with the expectation that no significant Mars Correlation would emerge in their studies; for until the results came in which supported Gauquelin, these "flaws" had gone unnoticed.

What, then,can we conclude from all this? First, that the Mars Effect is really a claimed correlation, an anomaly far less extraordinary -- and thus requiring far less evidence -- than a claim for an incredible causal relationship. Second, what evidence exists for the Mars Correlation does persistently show up if we accept the basic methodology proposed by Gauquelin. Third, if there is something fundamentally wrong with Gauquelin's methodology, that error has not been clearly stated or demonstrated in such a way as to convince skeptics toward such a claimed error (e.g., even critics of the Mars Effect like George Abell, Dennis Rawlins, and Richard Kammann have said that they do not understand the reasoning put forward by the CP, and Ray Hyman has raised criticisms of the KZA study rather similar to those put forward by Richard Kammann). Those independent investigations that have been conducted into Gauquelin's data (e.g., by Hans J. Eysenck, but we should recall that checks were also conducted by the CP and by Paul Kurtz) indicate that no irregularities are present there. (It might also be mentioned that privately circulated "critiques" of Gauquelin's analyses have been seen by many of us, but such critiques --e.g., those by Lawrence Jerome and Colin James -- have not been endorsed by their fellow critics who have found them unconvincing or in error.) Fourth, Gauquelin's own work has withstood critical appraisal thus far levied against it by responsible critics. And, fifth, following the normal procedures of science, we should accept the evidence for the Mars Correlation while recognizing that this evidence may yet be superseded by new research and data analysis that may establish it is invalid. But the burden for such replications and reanalyses must now fall upon the critics of the Mars Correlation. The Mars Correlation may or may not really be present, but the case ultimately must rest on the reasons given for any conclusion. So far, the case against the Mars Effect rests largely on spurious reasoning. The burden of proof is on the claimant, and Gauquelin has accepted that burden by conducting his studies. And if extraordinary claims require extraordinary proof, Gauquelin's meticulous research (which goes well beyond the scientific norm of openness since all his data is publicly available), plus the replications by his critics, surely constitutes such extraordinary evidence (especially when we recognize that the claim of a correlation rather than a causality is simply not that extraordinary).

The CP's Study

The Belgian CP, as I have noted above, replicated Gauquelin's experiment and his results but have argued that these do not support his conclusions because the CP questions the validity of the theoretical chance expectation level that Gauquelin used to compare with his (and their) group of sports champions. It is essential to note that, first of all, the CP unequivocally has accepted the data base of athletes they used. It has been pointed out that Gauquelin was asked to gather the data for them on these sports champions, and this has been brought up to question the authenticity of the sample. But (1) the CP chose the athletes, (2) Gauquelin merely acted in a secretarial capacity, at their request, in writing to the birth registries, (3) the data from the registries was directly transmitted to the CP, (4) these documents remain available for checks, and (5) the CP almost certainly checked this documentation since they flatly endorse the data and explicitly have exonerated Gauquelin of any charges related to error on his part regarding that data. Thus, questions raised about this data by Paul Kurtz and George Abell are clearly irrelevant unless they are really challenging the claims of the CP which sides with Gauquelin on this specific issue.

The main reason the CP has given for discounting the pro-Mars Effect interpretation of their own study concerns the matter of the theoretical chance expectation level that Gauquelin has used with which to compare the sports champion sample. This explanation is at this point still unclear to almost all critics of the Mars Effect outside of those connected with the CP. This may be due to problems of translation, and this will eventually be clarified once an authorized translation becomes available (ZS is now seeking to obtain such a translation in English). However, in principal, the original CP argument had merit since we had no empirical basis n1independent of Gauquelin's own data on non-athletes which supported his theoretical expectation for knowing what the chance rate for non-athletes (the general population) should really be. What has been largely overlooked is that the Zelen test (the KZA study) was constructed to avoid this very problem since the Zelen approach used an empirical (rather than theoretical) control group of non-champions. In the KZA test, the non-champions empirically demonstrated a Hars Effect of about 17%, the very level that Gauquelin and others had theoretically expected. Thus, the KZA test would seem --if nothing else-- to demonstrate that the CP test really did show evidence for the Mars Effect and that the CP interpretation is probably wrong. In light of the KZA test, the CP test clearly corroborates Gauquelin's interpretation. In this sense, the CP test and the KZA test are contradictory. Kurtz and Abell have chosen to remain silent about this fact and have misled their readers into thinking the CP test conclusions are congruent with their own negative conclusions about the Mars Effect. (For details about the CP test, I refer you to the exchanges between Gauquelin and J. Dommanget in ZS#9 and this issue.)

The CSICOP Connection

KZA conducted two separate studies, the first was published in The Humanist (the KZA or Zelen test), the second was published in The Skeptical Inquirer (the U.S. test). When Dennis Rawlins first brought charges against these studies in his FATE article ("sTarbaby"), the initial reaction of CSICOP's Councilors was that these simply were not CSICOP studies at all, merely studies conducted by three members of CSICOP, one of whom happened to be its chairman and one other a Councilor. It was quickly pointed out that this defense was completely wrong as far as the second test was concerned for CSICOP clearly sponsored the U.S. test, paid for it, and took credit for it via public statements at the time it was released. No one has claimed that all the members of CSICOP should be held responsible for the CSICOP related tests. The charges have been made against the CSICOP leadership, its Executive Council, and mainly by ex-Fellows of the CSICOP who wished to bring these faults to the attention of the CSICOP general membership. The CSICOP leadership has consistently tried to make it appear that the critics were attacking CSICOP rather than them. They have tried to make it appear that we critics have been out to destroy rather than reform CSICOP, a thoroughly untrue picture of the motives of the critics. Unfortunately, this picture of the critics has apparently been accepted by most of the CSICOP Fellows since they thus far have been remarkably indifferent to the whole affair. (And this indifference has actually been cited by Paul Kurtz and Hendrick Frazier as good reason for why further material on this controversy should not appear in The Skeptical Inquirer. CSICOP is interested in the education of the general public but perhaps less interested in the education of its members.) The fact that the CSICOP Council recently placed an announcement in The Skeptical Inquirer (following Rawlins' public critique) that CSICOP would no longer conduct or endorse research, is easily read as an admission, in fact, that they had sponsored the second KZA study (the U.S. test), and that this mistake would not be made again. (An unfortu::ate way of handling this embarrassment since one of the original reasons for forming CSICOP was so that they could conduct research and not merely encourage its being done by others.)

If the connection between the CSICOP and the U.S. test is clear, the connection between it and the first KZA study is far less so. It has been repeatedly pointed out by the Councilors that this first study was not even sponsored by CSICOP. That defense is not congruent with the facts. It is certainly true that the KZA study was published in THE HlJWIST and not even in The Skeptical Inquirer. But what is overlooked is that in those early days of CSICOP (when in fact I was co-chairman of CSICOP and on the Council) there was no clear distinction at all between the American Humanist Association and CSICOP. The AHA was the sponsor and creator of CSICOP (funding to initiate CSICOP came from the AHA). Paul Kurtz was the editor of The Humanist as well as chairman of the CSICOP. Funding seems to have been totally intertwined. CSICOP only became a separate organization after the first test was initiated, mainly because Paul Kurtz left his position as editor of The Humanist. This blending of the AUA and CSICOP matters was well known and a major reason for my own early problems with CSICOP since I had from the start disapproved what I saw as the authoritarian manifesto published by Kurtz against astrology in The Humanist, publicity for which actually promoted the initiation of CSICOP. Certainly, to now claim no serious connection existed between CSICOP and the original KZA test is remarkable post hoc reasoning that is more convenient than accurate.

But whatever the exact relationship between CSICOP and these anti-Mars Effect experiments, the clear fact remains that the Council of CSICOP has generally allowed the public to believe these studies are competent and that the criticisms levelled by Rawlins and others have been met. Their silence has generally been interpreted as consent, and that has been intentional. In other words, in my view, the CSICOP leadership has engaged in stonewalling their critics' charges while also resorting to ad honinem attacks (especially against Pawlins) and seeking to make it appear the CSICOP "mission" and not their own bungling is under attack. This is a tragedy since the major criticism against the CSICOP's handling of this affair has come not from allies of Gauquelin but from fellow rationalists who are in fact skeptical about the validity of the Mars Effect. In short, the CSICOP has been -- in my view -- guilty of the very pathological science that they were set up to attack. Instead of exemplifying a rational approach to an anomalous claim,CSICOP has descended into protecting orthodoxy and its own reputation as a goal more important than finding the truth. The inquirers have indeed become the Inquisitors that some feared they might. This is a great loss since the world truly needs a responsible group of critics to challenge the real nonsense and pseudoscience that competes with science.

All of this is not to say that the CSICOP has not otherwise done some valuable work. Nor do I suggest that most of the errors committed by its Council have been intentional. The sad part of all this is that honorable men could make such mistakes, often probably with good intentions and perhaps short-sighted high motives. Seeing their efforts as a Great Crusade Against the Irrational. I think they just got carried away, found themselves on the defensive, and then mistakenly tried to rationalize or ignore the errors -- all in the name of the Good Fight and for the Just Cause. Alas, as Coya put it so well: "The sleep of reason produces monsters."

Finally, on this issue, it should be noted that a number of Fellows of CSICOP were outraged by the behavior of the Council, and these Fellows resigned. A few others have expressed their concern. A few of the Technical Consultants to the CSICOP also resigned. But the vast majority of the Fellows remain apathetic or support the CSICOP (on this, see the early poll by Robert McConnell, whose results are published in this issue). So far there is little sign of reform demands from the CSICOP membership.