We should make clear that scientists are only as good at uncovering the truth as their methods are careful and unbiased. There are, unfortunately, a number of uncareful, biased scientists out there. It is not always easy to tell which scientists to pay attention to and which to tune out. The single easiest way to tell better science from worse is to follow the money supporting the work. When scientific funding comes from a group that doesn't have a vested interest in supporting any particular outcomes it is more likely that the results of the work so funded will be trustworthy than if the work is supported by a group who does have a vested interest in particular outcomes. Biomedical research supported by the federal government is generally more trustworthy than research which has been funded by a pharmaceutical group seeking to demonstrate the utility of a new drug, for example.
It is also a good idea to give more credence to work that has survived a process of peer-review vs. work that has been independently published. Work that has been peer reviewed has been carefully examined by other scientists looking for flaws prior to publication. That the work has been published, means that no serious flaws were found and that the results make sense to other scientists in the field. In a nutshell, a peer-reviewed study is something that can be trusted (assuming the journal in which the study has been published has a good reputation) and one that has not been peer reviewed should not be.
Peer reviewed, government funded scientific research may still suggest wrong or inaccurate answers to your issues, but will do so far less often than other alternatives.
(if you can't find good science to address your concern, look to the writings of experts in the field. Their work will be more biased and less reliable than scientific work, but it will also probably be more useful than nothing at all. Importantly, if you go in this direction, try to make sure that you are following the advice of a true expert, and not someone who is just pretending to be one. (degrees should be real and not purchased, experience should be "in the saddle" practical - if they say they are a therapist, they ought to work with clients)
We can sum all of this up by saying the following: The most effective self-help plans are those that are based on an individuals' cultural knowledge and expectations. Plans that are fitted to individuals' personal understandings of their problems will feel more correct and thus will end up getting acted upon more frequently. However, it is important that each person developing a self-help plan carefully reflect upon their cultural knowledge and expectations, keeping in mind that not everything they have learned is correct. Look up the results of peer-reviewed scientific research relevant to your problem or issue when you need to supplement or replace your personal understanding of that problem or issue. To the extent you can create a plan that draws upon scientific findings that inform and illuminate how to approach your problem and which feels right to you (one which fits with your cultural learnings and expectations), you will be in the best possible shape for creating positive change.