Language Attitude Survey

A direct survey asks respondents openly to report their attitudes toward a named language or variety using rating scales and questions, so it measures consciously held, reportable beliefs and is transparent and easy to scale. The indirect matched-guise technique conceals its purpose by having listeners rate 'different speakers' who are in fact one bilingual speaker in two guises, so it captures covert, often socially undesirable attitudes that people would not state directly. Direct surveys trade depth of concealment for transparency, scale, and ethical simplicity; indirect techniques trade scale and transparency for access to private attitudes. The two are complementary and often combined.

The most common formats are Likert items (agree–disagree responses to statements) and semantic-differential scales (ratings between bipolar adjective pairs such as beautiful–ugly), often supplemented by open questions. Because a single item is noisy, several items are combined into a scale; their internal consistency is then assessed, most often with Cronbach's alpha, and the underlying dimensions are examined with factor or principal-components analysis to confirm that the items measure the intended constructs before scores are interpreted.

Common safeguards include assuring and ensuring anonymity, wording items neutrally and balancing positive and negative statements to offset acquiescence, embedding sensitive items among neutral ones, using self-administered or online modes that feel less personal than an interviewer, and piloting the instrument to detect items that provoke guarded answers. Where attitudes are highly sensitive, researchers triangulate the direct survey with an indirect technique rather than relying on self-report alone.