BACKGROUND: The primary purpose was to assess if VAS or Likert was relatively more suitable when assessing self-ratings of voice and speech in patients treated for laryngeal cancer. An additional purpose was to investigate whether the scales differed in measuring change after treatment. MATERIAL/ METHODS: Using both VAS and Likert, 71 patients answered three questions at baseline and 12 months after treatment began. For test-retest analyses, additional data were used from 21 subjects. Non-parametric statistical methods suitable for comparing scales with different numbers of response alternatives were applied. RESULTS: Test-retest reliability was strongest for the Likert scale. From baseline to follow-up, both scales detected significant improvements. However, for some Likert scores the associated VAS values varied across a wide range, and this was true for assessments at baseline, at follow-up, and for score changes. At baseline the inter-scale concordance showed a greater number of disordered pairs than at follow-up and on both occasions the older group produced a greater number of disordered pairs. CONCLUSIONS: No support for the superiority of VAS over Likert was found. However, VAS may produce instability in response patterns. When using self-ratings of voice and speech to identify a need of further rehabilitation, Likert scales seem better suited.
BACKGROUND: The primary purpose was to assess if VAS or Likert was relatively more suitable when assessing self-ratings of voice and speech in patients treated for laryngeal cancer. An additional purpose was to investigate whether the scales differed in measuring change after treatment. MATERIAL/ METHODS: Using both VAS and Likert, 71 patients answered three questions at baseline and 12 months after treatment began. For test-retest analyses, additional data were used from 21 subjects. Non-parametric statistical methods suitable for comparing scales with different numbers of response alternatives were applied. RESULTS: Test-retest reliability was strongest for the Likert scale. From baseline to follow-up, both scales detected significant improvements. However, for some Likert scores the associated VAS values varied across a wide range, and this was true for assessments at baseline, at follow-up, and for score changes. At baseline the inter-scale concordance showed a greater number of disordered pairs than at follow-up and on both occasions the older group produced a greater number of disordered pairs. CONCLUSIONS: No support for the superiority of VAS over Likert was found. However, VAS may produce instability in response patterns. When using self-ratings of voice and speech to identify a need of further rehabilitation, Likert scales seem better suited.
Authors: Giovanni Aschero; Flavio Fenoglio; Maria Giuseppina Vidili; Andrea Wussler Journal: Health Qual Life Outcomes Date: 2010-09-25 Impact factor: 3.186