The interpretations (Discussion section) and the conclusions should be based on the data. Trust your judgment here. If the interpretations and conclusions are not logical - if they appear to be unrelated to the findings or generalize beyond the realm of the investigation - then their value is questionable.
The extent to which the researchers can generalize to a broader population and real life circumstances will depend on the exact conditions of the study itself. For instance, in a study of the effect of antiplaque/antigingivitis mouthwash on plaque deposits and gingivitis, was the study sample limited to a group of 20-30-year old dental hygiene students who tend to brush their teeth at least twice a day and are extremely health conscious and compliant? Perhaps in the broader population of adults, that includes the full array of ages and brushing habits, the “effect” of the positive mouthwash would be overwhelmed by the generally lower compliance and attentiveness to oral hygiene found in the general population.
