TOPIC: Application of Multivariate statistics in questionnaire design
Dr. D. Dutta Roy
Psychology Research Unit
INDIAN STATISTICAL INSTITUTE
203, B.T. ROAD., Kolkata
Date: 11.3.2011
While recognizing the powerful position occupied by the scientific sample survey in social research, it is worth noting to the reliability and validity of the data. Modern research based on schedule technique sometimes fails to establish the data quality (reliability and validity) resulting conflict in use of these data for social policy and social research movement. Besides, this technique fails to provide integrated information about social cognitions related to social issues. In
implementation of any policy, social cognition like attitudes, beliefs, prejudice, stereotype ideas, social motivation, happiness with specific social policy etc of people play critical roles. Questionnaire overcomes this limitation as it provides more integrated reliable and valid information. It is less expensive, requires less skill to administer, and instills confidence because of its anonymity. It
helps in developing objective reliable and valid database in short time based on large number of people. This database is useful to determine distribution of variables across diverse groups of larger society. Questionnaires can be used to examine the general characteristics of a population, to compare attitudes of different groups, and to test theories. Due to these reasons, questionnaire is most widely used data collection instrument now-a-days. Good questionnaire construction is critical to the success of a survey. Inappropriate questions, incorrect ordering of questions, incorrect scaling, or bad questionnaire format can make the survey valueless. Sir Francis Galton in 1870 invented questionnaire and conducted survey for the study of human differences and heredity. Questionnaire is a comprehensive measure to assess individual differences in human faculty. Therefore, questionnaire is embedded in a well-established theory of individual differences.
CHARACTERISTICS OF QUESTIONNAIRE
Questionnaire possesses few characteristics as
1. Objectivity: It should be free from any subjectivity (personal judgment
regarding ability, skill, traits, knowledge to be measured and evaluated).;
2. Standardization: There should be uniformity in instruction, questions,
administration of questionnaire, and scoring procedures.
3. Reliability: There should be consistency in responses over the time and across the questions.
4. Validity: It shows the degree to which a questionnaire measures what it
purports to measure. Questionnaire should be related with the internal and external criteria.
5. Norm: A norm is the average or typical score (mean or median) on a
particular questionnaire made by a specified population. This is important to determine individual differences in responses across age, educational level etc.
APPLICATION
Questionnaire has wide application in exploring and testing hypotheses in the areas of different disciplines as Psychology, Education, Sociology, Human resource management, Personnel management, Labour welfare, Marketing management, Consumer behaviour research, Financial management etc.
APPLICATION OF MULTIVARIATE STATISTICS
Questionnaire includes sets of questions which are often to be linearly related. Therefore, to assess their underlying latent relationship, to estimate reliability and validity, several multivariate statistics are used.
MVS refers to the set of statistical tools that simultaneously analyze multiple measurements on each individual or object under investigation.
It is the linear combination of variables with empirically determined weights.
Variate value= w1X1+ w2X2+ w3X3 + wnXn
W=weight determined by the multivariate technique; X=observed variable
Common statistics are principal component, cluster analysis, multiple regression analysis, discriminant function analysis, correspondence analysis.
Analysis depends on specific objectives, pattern of relations among the measures and measurement scales used in the questionnaire.
When responding to the questions, individual is affected by several random variables. Multivariate statistics partials out random variation effect and accounts only fixed effect.
References:
- Dutta Roy, D. (2009). Principles of Questionnaire development with empirical studies. Prasad Psycho Corporation. New Delhi.
- Dutta Roy,D. - Construct validity of writing motivation questionnaire. International Journal of Psychological Research
- Dutta Roy,D.(2010).Cluster Analysis for Test-Retest Reliability. International Journal of Psychological Research ,(published from USA). 3,1,132-140.
- Dutta Roy,D.(2010).Construct validity of Reading motivation. Journal of the Indian Academy of Applied Psychology, (to be published in January 2011, vol.37,No.1).
- Dutta Roy, D. (2005).Clustering state anxiety scores across twelve months in the Antarctic expedition,Journal of Psychometry,Vol.XIX,1,January,pp 14-20.
- Dutta Roy, D. (2003).Cluster analysis of GHQ-12 items using Indian Antarctica Expeditioners' Responses,,Journal of Psychometry ,Vol.17, No.1&2,pp 38-44.
- Dutta Roy,D. (2002). Correspondence between item and rating on the checklist of relative importance of computer programming tasks., Journal of Psychometry, 16,2,67-76.
