Rex Kerr answers your question well to the extent that you ask about the desires, perceptions, and knowledge of individual inquirers or scientists. We may be able to will ourselves to perceive things in a certain way out of sheer force of desire, up to a point. What point that is exactly becomes an empirical psychology question. I suspect that for some people, experiencing psychosis, little desire is necessary to adjust their sense of reality such that it doesn't conform with perceptions, while for others it may take considerable will or unconscious desire.
However, that is all to discuss the individual, which is the wrong unit of analysis for discussing natural science. Philosopher Helen Longino, for instance, has argued persuasively that an individual person cannot do science in isolation, that science is a fundamentally social process. Individuals can have knowledge they acquire on their own. I see that I am typing now. I know that I am typing. By "know" I mean that I am defeasibly very confident, but not that I am objective. That knowledge is not yet scientific, and we should understand that knowledge as neither certain nor objective.
Your scientist would therefore be wrong to say that "only the hard sciences produce true knowledge," but right to say that scientific methods produce better-justified knowledge than pre-scientific knowledge, because only science, by definition, can lend objectivity to knowledge. Science simply is the process through which beliefs are tested and criticized in ways that justify our treating the few beliefs that pass our tests as more objective.
How? Knowledge-generating methods and procedures are scientific precisely if and because they neutralize the desires of individuals or groups to believe one thing or another. An individual's perceiving and believing independently do not have those features. When methods and procedures involving different, independent assumptions converge on the same results, we describe those results as more "robust" or more "objective."
All of this is consistent with the idea that yes, our desires can influence — though I'm not sure I would agree they can "determine" — our beliefs. But no amount of desire per se can render our beliefs scientific or objective. We should only understand our beliefs as hewing closer to reality than other beliefs do when our beliefs have been produced by desire-neutralizing processes rather than by desire itself.