This is a different issue, and is also not quite the way you stated here. The issue is not falsification of results (which is outright fraud and illegal), but of irreproducibility of results (which is sloppy science). Also, this is not so much an issue of basic research, but of a different kind of research. Basic research reveals fundamental processes, such as the effect of changing temperature on the production of a specific enzyme within a specific cell type. This kind of research is highly reproducible; if I describe my work, and another scientist wants to use it as a basis for his work, his first step will be to reproduce my work. Basic research is very straightforward and reproducible.
Where the issue of non-reproducibility comes up most often is in large-scale clinical/observational studies. These studies rely on large sample sizes, because the systems being studied have a lot of background variability. In order to analyze the study, heavy-duty statistics are applied. Where I use statistics to show whether the effect I measured in experimental group A is really different than what I measured in control group B, very often in large-scale studies, statistics are used to determine whether there is an effect at all--not whether an observed effect is real, but if an effect even exists. That is a huge difference from basic research.
In numerical terms, that is the difference between my using statistics to validate a 500% difference between experimental and control groups, while the large scale studies use statistics to reveal a 1% difference between the two groups--a difference that is probably not even real.
There are other **major** problems with large-scale studies that depend on statistical analysis to determine if the study shows anything. This kind of research is a pet peeve of mine, and I could go on about it for a very long time. People engaged in this kind of research very often are not trained as scientists--they are MDs, usually--and do not understand basic biochemical mechanisms and over-estimate the power of statistics. I am actually glad that objections to this kind of "science" are getting some publicity--that kind of attention is *long* overdue. Large scale studies cannot be avoided, but they desperately need methodological reform.
The bottom line is that none of this has anything to do with the fact that government funds the majority of basic research. Given the nature of how science is funded, I only see that basic research would halt if government stopped funding, because it has no commercial application and is not funded by the private sector.
You are correct that article I read dealt with clinical studies. But I also recall some news recently of basic researchers being caught in falsifying research.
But I will have to stick to my guns and say that it would be better for science if the government was out of the business of selecting what research gets funded and what does not. Government could give tax breaks to industry to fund basic research at universities.
Government funded basic research gave us Anthropomorphic Global Warming and I would rather not have another round of that kind of science.