In this paper, based on the research of paronym use, we discuss the relationship between the norm and usage, and the challenges of researching norm and usage. Paronyms, as words that look similar but have a different meaning are an interesting standard language issue – and a linguistic issue as well because paronymy is a type of lexical relationship. When researching paronyms as a standard language issue we are faced with the challenge of researching norm and usage; therefore, we think about the normative sources in the Croatian language, about what the norm should be based on, and about the possible ways researching usage. The research was carried out on the sample of 296 participants via an anonymous questionnaire that comprised a sentence completion task. The results of the research show that there are paronym pairs which a quite high percentage of participants uses in a meaning that is different from what is prescribed by the norm; the results also show that the norm itself does not always treat certain pairs as paronyms, but rather as synonyms.