Online relationships versus face-to-face relationships. Proposal of a methodology for the holistic analysis of Likert-type scales

The questionnaire is a widely used instrument in social research for collecting information. One of the most important design decisions is the choice of one (or more) measurement scale(s). A common scale for rating/opinion questions is the Likert-type scale ranging from 0 or 1 (lowest rating/strongest disagreement) to 10 (highest rating/strongest agreement). Its success is based, among other things, on the advantage of being able to treat the scale from different perspectives (nominal, ordinal, or interval). Making the right choice is not an easy task for the researcher, as the perspective selected determines the methodology to be used for analysis and the potential results. This paper proposes a methodological strategy that facilitates this phase by presenting a protocol that allows the simultaneous analysis of two versions of the scale, nominal and interval. The proposed idea is based on the tabulation of the two scales in a mixed data matrix and it is advised to conduct an exploratory analysis through Multiple Factor Analysis. This technique maintains and compares the idiosyncrasies of each scale. The versatility and suitability of the proposal is illustrated by real research on social and affective relationships during the COVID-19 pandemic in Spain.