Analysis of Likert-Scale Data

Analysis of Likert-Scale Data - Psychological B ulletin...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Psychological B ulletin 1989, V ol. 105, N o. 1 ,1 6 7 -1 7 0 C opyright 1989 by the A m erican Psychological A ssociation, In c. 0033-2909/89/S 00.75 Analysis of Likert-Scale D ata: A R einterpretation of G regoire and D river Jeffrey Lee Rasmussen P u rd u e S ch o o l o f S cience In d ia n a U n iv e rsity P u rd u e U n iv e rsity a t In d ia n a p o lis M ost studies that have investigated the use of coarsely grained scales have indicated that the accuracy of statistics calculated on such scales is not com prom ised as long as the scales have about 5 or m ore points. G regoire and Driver (1987), however; found serious perturbances of the Type I and Type II error rates using a 5-point scale. They carried out three com puter sim ulation experim ents in which continuous data were transform ed to Likert-scale values. Two of the three experim ents are shown to be flaw ed because the authors incorrectly specified the population m ean in their sim ulation. This article corrects the flaw and dem onstrates that the Type I and Type II error rates are not seriously com prom ised by the use of ordinal-scale data. Furtherm ore, G regoire and D river's results are rein- terpreted to show that in most cases, the param etric test of location equality shows a power superior- ity to the nonparam etric tests. Only in their most nonnorm al sim ulation does a nonparam etric test show a power superiority. O ften, in psychological research th e d ep en d en t v a ria b le m ay tak e o n only a lim ite d n u m b e r o f v alu es. T h e u b iq u ito u s 5 - o r 7 -p o in t L ik e rt scale is a good e x a m p le o f su c h a d e p e n d e n t v ari- a b le . A g reat deal o f research h as b een c a rrie d o u t co n cern in g th e effects o f th e n u m b e r o f scale p o in ts o n th e reliab ility an d o n th e T y p e I/T y p e II e rro r ra te s o f statistical te c h n iq u e s (B en - dig, 1 9 5 4 ; C ic c h e tti, S how alter, & T yrer, 1 9 8 5 ; C o h e n , 1983; C o n n o r, 1 9 7 2 ; J e n k in s & T aber, 1 9 7 7 ; K o m o rita & G ra h a m , 1965; L issitz & G re e n , 1 9 7 5 ; L u n n e y , 1970; M a te ll & Ja c o b y , 1 9 7 1 ;R a m sa y , 1973). In g en eral, stu d ies h av e in d icated th a t th e reliab ility o f a test in creases w ith an in crease in th e n u m b e r o f scale p o in ts. F o r e x a m p le , C ic c h e tti et a l. (1 9 8 5 ) in v estig ated th e effects o f th e n u m b e r o f scale p o in ts from 2 to 100 o n th e a c c u ra c y o f th e in tra c la ss co rrelatio n . T h e y found th a t re lia b ility in creased stead ily from a 2 -p o in t scale to a 7 -p o in t scale a n d th e n leveled off after th e 7 -p o in t scale. T h e m o st d ra m a tic in crease w as show n b etw een th e 2 -p o in t scale an d th e 3 -p o in t scale....
View Full Document

This note was uploaded on 05/04/2010 for the course MARKETING ME501 taught by Professor Mark during the Spring '10 term at Assumption College.

Page1 / 4

Analysis of Likert-Scale Data - Psychological B ulletin...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online