Journal of Oral Health and Community Dentistry

Register      Login

VOLUME 9 , ISSUE 3 ( September-December, 2015 ) > List of Articles

RESEARCH ARTICLE

Adjusting the Oral Health Related Quality of Life Measure (Using Ohip-14) for Floor and Ceiling Effects

M Andiappan, FJ Hughes, S Dunne, W Gao, ANA Donaldson

Citation Information : Andiappan M, Hughes F, Dunne S, Gao W, Donaldson A. Adjusting the Oral Health Related Quality of Life Measure (Using Ohip-14) for Floor and Ceiling Effects. J Oral Health Comm Dent 2015; 9 (3):99-104.

DOI: 10.5005/johcd-9-3-99

License: CC BY-NC 3.0

Published Online: 01-09-2015

Copyright Statement:  Copyright © 2015; Jaypee Brothers Medical Publishers (P) Ltd.


Abstract

Introduction

The influence of floor (lowest) and ceiling (highest) effects on the outcome measure is of serious concern particularly when the outcome is measured using Likert scales. Conventional regression methods yield biased results and hence tobit regression is to be used to adjust for these effects. This paper is an attempt to use tobit regression in finding the predictors of oral health related quality of life after adjusting for floor and ceiling effects.

Methods

A sample of 360 participants were asked to self asses their oral health related quality of life (OHRQoL) using Oral Health impact profile with 14 items which forms the data for this study. Apart from descriptive statistics, Ordinary Least squares regression and tobit regression were used to find the significant predictors of OHRQoL and the results of both methods were compared.

Results

The sample comprised of 41.1% men and 58.9% women. Majority of the participants (68.3%) were whites. The average item difficulty was 0.4 and the average item easiness was 0.03. The floor and ceiling values for the composite scores were 14 and 56 respectively. Age and gender were not statistically significant both in Ordinary Least Squares (OLS) regression and Tobit regression. Full time employment, student and retired have significantly lower scores in OLS but only retired had significantly lower scores in the tobit model.

Conclusion

Tobit model, after adjusting for floor and ceiling effect, gives higher values for the predictors and the OLS model underestimates the effects of predictors on OHIP scores.


PDF Share
  1. Generic and oral quality of life is affected by oral mucosal diseases. BMC Oral Health 2012;12:2.
  2. The Paradox of Better Subjective Oral Health in Older Age. J Dent Res 2011;90(11):1279–85.
  3. Quality of life and self-perceived oral health among workers from a furniture industry. Braz J Oral Sci 2011;10(4):226–32.
  4. Investigating Ceiling Effects In Longitudinal Data Analysis. Multivariate Behav Res 2009;43(3):476–96.
  5. A note on the use of the Tobit approach for tests scores with floor and ceiling effects. British Journal of Mathematical and Statistical Psychology 1997; 50:3510364.
  6. Modeling Outcomes with Floor or Ceiling Effects: An introduction to the Tobit Model. Gifted Child Quarterly 2010; 54(4): 314–20.
  7. The magical number of seven, plus or minus two: some limits on our capacity for processing information. Psychological Review 1956;63:81–97.
  8. Evaluation of the spiritual well-being scale in a sample of college students. The international journal for the psychology of Religion 2001;11(1): 25–33.
  9. An evaluation of the research and usefulness of the spiritual we-being scale. Journal of Psychology and Theology 1991;19:49–55.
  10. Applied Longitudinal analysis. Wiley; New York: 2004.
  11. A Random-effects model for longitudinal data. Biometrics 1982;38:963–74.
  12. Estimation of relationships for l imited dependent variables. Econometrica 1958;26(1):24–36.
  13. TOBIT factor analysis. British Journal of Mathematical and Statistical Psychology 1989;42:241–50.
  14. Congeneric modeling of reliability using censored variables. Applied Psychological Measurement 1989;13:151–159.
PDF Share
PDF Share

© Jaypee Brothers Medical Publishers (P) LTD.