Peer reviewed or refereed articles are scholarly works in academic journals which have been subjected to review by researchers, scholars, or experts in the field of study. Prior to publication the journal will send the article out, sometimes as "double-blind review" in which both author and reviewers' identities are concealed, to be vetted for quality, methodology, validity, use of evidence, etc.
At it's best peer review will stop erroneous research, unsubstantiated claims, blatantly biased, or weak scholarship from reaching publication.
News, magazine, and trade publication articles usually undergo some level of editorial review, but this is limited and the standards are not nearly as rigorous.
However, just because a researcher has name recognition or is affiliated with an impressive institution and their work has been published in an academic journal does not guarantee good scholarship or veracity. Aim to use the best information available, but be critical in your evaluation of all information. The ability to apply critical reasoning effectively is a skill that is developed over time (and continually honed over a lifetime) and as you drill down deeper into a field of study you will acquire the knowledge and reasoning abilities to do this more effectively.
While there are more and more open access journals cropping up with freely available content, by and large peer reviewed academic journals are expensive and only available to those with library or institutional access via subscription databases. Even if you use a search engine like Google Scholar many of the results will not link to the full text. You will still need access to databases to retrieve the full article.
Given the click bait culture of digital media there is an unfortunate impetus for journalists to exaggerate or misrepresent research studies in an attempt to grab readers' attention. It is always ideal to read the original study to see if the media characterization matches up with the actual findings of a study.
Sometimes bad research passes peer review and is published. But, unlike the web where unfounded non-information and dangerous misinformation abound, scholarly communities have a vested interest in ensuring the quality and integrity of the work being produced in their respective fields, particularly for scientific research. Scholarly publishers maintain policies for withdrawal or retraction for articles found to be fraudulent, bogus, biased, or to violate professional ethical standards. There are also sites like Retraction Watch which promote transparency in scholarly publishing.
There have been and will continue to be questionable research. One example of this is the issue of funding bias.
Lexchin, J. (2012, January 1). Sponsorship bias in clinical research. The International Journal of Risk & Safety in Medicine, 24, 4, 233-42. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/23135338
Goldacre, B. (2013, February 13). Trial sans error: How pharma-funded research cherry-picks positive results [excerpt]. Scientific American. Retrieved from http://www.scientificamerican.com/article/trial-sans-error-how-pharma-funded-research-cherry-picks-positive-results/