It is common to make claims about the relationship between scientific positions and philosophical positions, but the arguments that underwrite such claims are not always the same. There are at least two major groups:
(A) Analogy of Theory: We have some scientific theory, call it T, which naturally suggests some metaphysical or natural-philosophical position. For instance, Gilson notes somewhere that historically people who tended to assume that the physical world is deterministic did so because determinism seemed the natural philosophical analogue of how they understood Newtonian physics (T).
(H) Historical Induction: Our best scientific theory, call it T, was reached by assuming some metaphysical or natural-philosophical position, sometimes such that even those who had a different position in fact proceeded as if this philosophical position were true. For instance, it is sometimes argued that naturalism is the best philosophical position because our modern physical theories (T) were developed by people assuming naturalism or, even if they did not in fact believe naturalism true, doing their investigations as if naturalism were true.
(There are other kinds of arguments; for instance, both of these go from scientific theory to philosophical position, but it is also the case that people accept scientific theories that best fit their philosophical positions. I am more interested here, however, in the science-to-philosophy direction.)
In practice, people tend to use either kind of argument as it suits them, as is perhaps inevitable, but there are complications that are often not recognized. First, both kinds of argument are sensitive to what we might focus on; for instance, using H on neuroscience for the past fifty years, the philosophical position that might be suggested is materialism about the mind, but if one uses the entire history of neuroscience, the sure winner is substance dualism. I've noted before that not only were substance dualists primarily the founders of the field, most scientists studying the brain were substance dualists well into the twentieth century, and two people who have received the Nobel prize for major work in neuroscience (Sherrington and Eccles) were vocal substance dualists of one kind or another. And this makes sense in itself -- if you want a philosophical position that both recognizes the relevance of the brain to the mind but at the same time does not make assumptions that would affect your interpretation of how any particular discovery about the brain bears on questions about the mind, substance dualism is a very obvious candidate, and in the development of a difficult science a philosophical position that recognizes a connection without prejudging it will have obvious advantages over its rivals.
Second, assessment sometimes depends on your reference points. For instance, most modern biologists would take modern biological theories to suggest materialism/mechanism by A and H alike, but it is clear that by nineteenth-century standards modern biology would have to count as weakly vitalistic and the actual history of biology is much more checkered. Modern biology is partly built out of mechanistic triumphs (such as the discovery that there is no sharp line between organic and inorganic compounds), but it is also partly built out of vitalistic triumphs (such as the organic theory of fermentation and the germ theory of disease), and the ease with which biologists fall back into information-based metaphors is a vitalistic heritage. The reason modern biologist don't recognize their field as vitalistic is that when they think vitalism, they are largely thinking of vitalism as Driesch's entelechies and things of that sort,
and take more modern theories to have developed solutions to problems that nullify the kinds of problems late vitalists insisted upon; they think of strong forms of vitalism when thinking of vitalism and weak and partially agnostic forms of mechanism when they think of materialism or mechanism; and they organize their thoughts on the history of the field, when they think of vitalistic triumphs at all, as a largely mechanistic framework with vitalistic qualifications rather than a largely vitalistic framework with mechanistic qualifications.
Another example is that while there were certainly Newtonian physicists who were determinist, this was not a historically popular position; most physicists who laid the groundwork of the field thought that there were causes not covered by the laws of motion, such as minds with free will, and thus tended to read Newtonian theory as ceteris paribus, something you use when you don't have to worry about such other causes, rather than as absolute.
A third complication is that A and H are both based on highly defeasible forms of inference, and therefore can never on their own get you certain conclusions. A is based on analogical inference, and, what is more, most often an analogical inference based not on any rigorous standard of similarity but a general 'feel' as to how this suggests that. As some have noted in recent decades, it's not clear how strictly deterministic Newtonian physics is even when taken strictly; there are situations definable in the theory, like Norton's Dome, for which arguably there is an answer but no deterministic one, and while there are counterarguments, there is no general agreement about whether they are successful. But the way Newtonian physics tends to be taught makes the billiard-ball universe seem inevitable.
Likewise, just as A depends on an assumption of a standard of similarity that might not be accepted, so does H depend on an assumption of uniformity that might be controverted. Major sea-changes do occur, after all; the history of neuroscience is in fact very favorable to substance dualism, but one could well argue that once it reached a certain point, enough changes had accumulated to tip the scales in a different direction.
These complications are unavoidable; there is no kind of A or H you could make that would enable you to avoid them completely. And thus we have what we have, with everybody using whichever one, and whichever variation of one, opportunistically.