In an earlier study we investigated the development of noun plurals in Danish children aged 0-10 years using a multi method research approach comparing five different data types: 1) lexical data; 2) reported data; 3) naturalistic spontaneous child language input and output; 4) semi-naturalistic/semi-experimental data; 5) experimental data. Our results showed very interesting patterns (both similar and diverse) across data types. These results lead us to our main research question for the present study: Which impact does the research method have on the results of an empirical study? We predict: 1) Parents do not perceive all the linguistic output that their child produces and we therefore expect to find nouns and noun PL forms earlier in naturalistic spontaneous child speech data than in reported data. 2) The semi-naturalistic/semi-experimental data are collected in a more naturalistic setting than the experimental data and we therefore predict a lower percentage of incorrectly produced PL forms in the semi-naturalistic/semi-experimental than in the experimental data. 3) In the experimental data the children are to produce the PL form of nouns given by the investigator. We expect children to produce a large amount of sg. instead of PL forms, either as a repetition of the sg. form given by the investigator or as a PL error form irreguralizing the “Ø” PL marker (pure zero, e.g. sg. mus ‘mouse’ – PL mus ‘mice’). 4) The experimental data is based on a fixed number of specific pre-selected items and we therefore expect the amount of correctly produced PL forms in the experimental data to be sensitive to the specific items selected. Furthermore, we will discuss how a multi-method approach can reveal interactions between different factors in the acquisition of morphology that in the past could not be revealed with just one method.