Competition in the private sector incentivizes upstanding research. But government-funded research works quite differently.
Fraud is a big problem in the science world in general. There is a huge amount of pressure to get positive results. Continued funding for researchers is often contingent upon those results. It isn't terribly surprising that fudging the numbers can be a large temptation.
However, at least in the private world, competition helps to keep such temptation in check, at least to some degree. For instance, there is a lot of competition among the hundreds of private scientific journals to be the best. For a journal to gain a reputation for publishing fraudulent or otherwise inaccurate data would be devastating.
On the other hand, educational institutes that are funded by government are motivated to shield researchers since large government grants are often dependent upon them. Politicians tend not to want to fund research that disagrees with their political stance on an issue whether motivated by ideology or economics. This is a big problem in the medical world (think COVID) but can affect research of all types (think climate change).
The numbers seem to back this up. Approximately 20% of the time, clinical trials are fraudulent. However, when looking specifically at government funded trials, nearly 44% contained false data according to an independent analysis. In fact, 34% of scientist receiving Federal funding have admitted to engaging in research misconduct of one kind or another in order to make their research match the ideological and/or economic agenda of politicians.
Guess how much funding there is for researching the accuracy of government-funded research...