When astrology’s defenders are asked for their best scientific evidence, one name comes up more often than almost any other: Vernon Clark. A clinical psychologist working in the 1960s, Clark designed a series of matching experiments that — unlike the Carlson experiment two decades later — produced results favoring astrology’s claims. Astrologers outperformed chance. Astrologers outperformed a comparison group of non-astrologer professionals.
These studies are real, they were published, and their results were genuinely positive for astrology. They’re also, by any modern standard, the kind of study that would need substantial follow-up before being treated as conclusive — and that follow-up mostly didn’t happen in the form it needed to. Understanding why the Vernon Clark studies produced the results they did, and why they didn’t end the debate, is a useful case study in how early positive results in a contested field tend to play out.
What Clark Did
Vernon Clark ran several related experiments between 1960 and 1968, the most cited of which involved a matching paradigm broadly similar in spirit to what Carlson would later test — but with a different structure and a smaller scale.
In one widely cited version, Clark recruited pairs of subjects matched for age, sex, and general background but who had pursued very different life paths — for instance, one subject who had become a successful professional in a creative field and one who had spent time in a psychiatric institution, or pairs with similarly contrasting life outcomes. Astrologers were given the natal charts for each pair (without identifying information) and asked to determine which chart belonged to which life outcome, based on chart features alone.
In Clark’s most-cited study, a group of astrologers correctly matched charts to case histories at rates significantly above the 50% chance level expected for a two-option matching task — with some versions of the experiment reporting accuracy in the 70-80% range across multiple trials. A comparison group of non-astrologers (in some versions, social workers or psychologists without astrological training) given the same case histories and charts performed at approximately chance levels, unable to extract meaningful signal from chart data they didn’t know how to read.
Clark interpreted this as evidence that astrologers were extracting genuine information from natal charts that non-astrologers could not — information that correlated with real differences in life outcomes.
Why the Results Looked Strong
Several features of Clark’s design contributed to the positive results, and identifying them is not an accusation of fraud — it’s how methodology works, and it’s exactly the kind of thing follow-up research is supposed to identify and control for.
The case pairs were often extreme. Matching a chart to “became a successful architect” versus “spent significant time in psychiatric care” is a coarser task than matching a chart to one of many similar personality profiles, as in Carlson’s CPI-based design. Extreme contrasts are easier to detect through almost any systematic approach, including ones that don’t depend on the proposed mechanism being true.
Sample sizes were small. Several of Clark’s experiments involved single-digit or low double-digit numbers of astrologers and case pairs. Small samples are more susceptible to results that don’t replicate — not necessarily through fraud, but through the ordinary statistical reality that small samples produce noisier estimates, and noisy estimates sometimes land on results that look dramatic.
The astrologers may have had access to cues beyond the chart itself. Some critiques of Clark’s methodology have noted that case history materials, even with identifying information removed, can contain stylistic or contextual cues — the way a case is written up, details that correlate with the subject’s background — that a skilled reader could pick up on without using astrology at all. This is a known issue in matching-paradigm studies generally and is part of why double-blind protocols with carefully sanitized materials (as Carlson later used) became the standard.
Publication and citation dynamics favor positive results. This is true across all of science, not specific to astrology research — studies with positive, interesting findings get published, cited, and remembered more than null results, which often don’t get published at all. The “file drawer problem” means that we don’t have a clear picture of how many similar matching attempts produced null results and were never written up.
What Happened to Replication
The crucial question for any early positive result is: did it replicate under tighter controls? For the Vernon Clark studies, the honest answer is: not cleanly, and not at the scale that would settle the question.
Some subsequent attempts to replicate Clark-style matching studies with larger samples and tighter controls on case material produced weaker or null results. A notable feature of the broader literature, summarized in reviews by researchers including Geoffrey Dean, is that matching study results tend to attenuate — get smaller and less significant — as sample sizes increase and controls tighten. This pattern (strong early results in small studies, weaker results in larger follow-ups) is not unique to astrology research. It’s a well-documented phenomenon across many fields and is one of the core findings discussed in the literature on the replication crisis in psychology more broadly.
Carlson’s 1985 study can be read, in part, as the large-sample, tightly-controlled version of the Clark paradigm that the field needed — and it produced a null result. Whether Carlson’s study and Clark’s studies are measuring the same underlying claim closely enough to treat one as a “correction” of the other is debatable; the case materials, matching tasks, and astrologer populations were different in each. But the overall trajectory — promising small studies followed by a large, rigorous study finding nothing — fits a pattern that shows up across many areas of contested research.
What Astrology Proponents Say About This
It’s worth representing the position of astrologers who continue to cite Clark’s work, because it’s not simply “ignore the larger study.”
Some practitioners argue that Clark’s matching task — extreme life-outcome contrasts — is actually closer to what astrology claims to be good at than Carlson’s CPI-matching task. Natal astrology, in this view, is better suited to identifying major life-direction signatures (a chart suggesting significant struggle versus significant achievement) than to matching fine-grained personality trait profiles. If that’s correct, Clark and Carlson may not be testing the same claim at different scales — they may be testing different claims, one of which (the Clark version) hasn’t been adequately retested with modern controls.
This is a coherent argument, and it points to a real gap: a large, well-controlled, double-blind study using Clark’s extreme-contrast matching design, with sanitized case materials and a substantial sample of astrologers, does not appear to exist in the literature. Until it does, the question Clark’s studies originally asked remains genuinely open — not because Clark’s results stand as proven, but because they were never properly tested at scale with the rigor the field eventually developed for other questions.
The Honest Position
The Vernon Clark studies are real, were published, and produced results favorable to astrology under the conditions tested. They are not fraudulent, and dismissing them outright misrepresents the history of the research.
They are also small studies from the 1960s, using a task design (extreme-contrast matching) that has not been adequately retested at scale with modern methodological controls. The broader pattern in matching-paradigm research — strong small-sample results attenuating under larger, tighter studies — applies here as it does elsewhere, and the one large, rigorous matching study that exists (Carlson’s) used a different task design and found nothing.
What this leaves us with is not “astrology has been proven” or “Clark was wrong.” It’s a specific, well-defined research gap: does the extreme-contrast matching paradigm that produced Clark’s positive results hold up under Carlson-level rigor? Nobody has run that study. Until someone does, both the Clark results and the question they raised remain part of the unfinished business of divination research — cited frequently, resolved rarely, and more useful as an illustration of how science handles early positive results than as evidence for or against astrology itself.
For the contrasting case — what happens when a matching study is run at scale with tight controls — see the Carlson experiment, and for the broader statistical issues that make this kind of research difficult to interpret, the p-value problem in astrology research.