You are on page 1of 8

-

How to Design and Report Likert Scale First How to Design one:
http://www.gifted.uconn.edu/siegle/research/instrument%20Reliability%20and %20Validity/Likert.html !ee the many different types of Likert !cale" http://www.socialresearchmethods.net/kb/scallik.htm a mini #ob$aid with good e%amples"

Likert scale: & Likert scale pronounced 'lick$ert'" is a type of psychometric response scale often used in (uestionnaires) and is the most widely used scale in sur*ey research. +hen responding to a Likert (uestionnaire item) respondents specify their le*el of agreement to a statement. ,he scale is named after Rensis Likert) who published a report describing its use Likert) -./2".

Sample Question presented using a five-point Likert Scale


& typical test item in a Likert scale is a statement) the respondent is asked to indicate their degree of agreement with the statement. ,raditionally a fi*e$point scale is used) howe*er many psychometricians ad*ocate using a se*en or nine point scale. Ice cream is good for breakfast -. 2. /. 2. 3. !trongly disagree 0isagree 1either agree nor disagree &gree !trongly agree

Likert scaling is a bipolar scaling method) measuring either positi*e and negati*e response to a statement. !ometimes Likert scales are used in a forced choice method where the middle option of 41either agree nor disagree4 is not a*ailable. Likert scales may be sub#ect to distortion from se*eral causes. Respondents may a*oid using e%treme response categories central tendency bias"5 agree with statements as presented acquiescence response bias"5 or try to portray themsel*es or their group in a more fa*orable light social desirability bias".

Scoring and analysis: http://www.answers.com/topic/likert-scale


&fter the (uestionnaire is completed) each item may be analy6ed separately or item responses may be summed to create a score for a group of items. 7ence) Likert scales are often called summati*e scales. Responses to a single Likert item are normally treated as ordinal data) because) especially when using only fi*e le*els) one cannot assume that respondents percei*e the difference

2 between ad#acent le*els as e(uidistant. +hen treated as ordinal data) Likert

responses can be analy6ed using non$parametric tests) such as the 8ann$ +hitney test) the +ilco%on signed$rank test) and the 9ruskal$+allis test.:-; Likert !cale <nter*al 0ata"
+hen responses to se*eral Likert items are summed) they may be treated as inter*al data measuring a latent *ariable. <f the summed responses are normally distributed) parametric statistical tests such as the analysis of *ariance can be applied. =%amples:

&ttitudes toward >omputer 20 (uestions) but =ach participant gets one score) summed"

!ample ?uestions: -. < feel happy when using a computer. 2. 8y hands will sweat e*ery time when < use a computer /. < feel an%ious when < ha*e to type up a paper on a computer 2. @.. !ome of the personality test same way" 9olbAs Learning !tyle <n*entory >heck it out online" 0ata from Likert scales are sometimes reduced to the nominal le*el by combining all agree and disagree responses into two categories of 4accept4 and 4re#ect4. ,he >ochran ?) or 8c1emar$,est are common statistical procedures used after this transformation.

A Note from Marcie:


Plenty of people treat ordinal data as continuous --- especially when survey data are supplemented with other measures

!"ample of a Likert #cale $ordinal% #urvey and &ata Analysis


0ata set: the one posted on >ourse +ebsite: Cultural differences in online Learning See the Survey at: http://sur*eymonkey.com/s.aspBuCD-20-EE/32/ ?uestion 2F !tudentsA Gerceptions on ,eachers and ,eaching in Heneral" Practice of statistical Analysis for two Class Periods: Set the data correctly (Data, aria!le" Analy#e Data$Practice$contin%o%s (Descripti&e Analysis, first ti'e, lea&e (factor) %nchecked* second ti'e, check it, co'pare the res%lts" Look at the res%lts and see what concl%sions can yo% draw+ Analy#e Data$Practice$,rdinal (-nferential statistics" #ummary of &ata: $&escriptive analysis%'generated by #urveyMonkey . Iollowing are se*eral (uestions about your perceptions on or e%pectations about teachers and teaching in general5 please click on the button to indicate your choice: #trongl )espons #trongl Agre (ndecide &isagre y e y agree e d e disagre Average e < typically consider my teachers to ha*e wisdom. < usually ha*e a great deal of respect for my teachers. /2% 23" 2.% /." *+, $-+% *+, $-+% J% ." -% 2" 0% 0" . -/

.% -2"

2% 3" 2/% /0"

0% 0"

. +-

< feel me and my teachers 01, are essentially e(uals. .% -2" 2D% /3" $*2% < think there should be e%press rules of conduct in e*ery class which all students should follow. < e%pect my teachers to be recogni6ed e%perts in the field which they teach.

2% /"

3 /4

2J% /D"

*1, -J% 2/" $//%

D% E"

0% 0"

3 13

0*, $/1%

2/% 3J"

E% --"

/% 2"

-% -"

. -.

2 < am more comfortable when my teacher conducts class in a formal manner rather than J% ." informally.

/0% 20"

2/% /-"

2*, $0-%

3% J"

2 13

5otal )espondents $filtered out% $skipped this 6uestion% Note: however7 this summary does not give us Mean and #&7 still I needed to analy8e the raw data

.20 2 .

&ata: $Part of the )aw &ata-#urveyMonkey will give you a numerical version%
Culture Open-Ended Response QS Q1Wisd om Q2Resp ect Q3Equ al Q4rulescond uct Q5expe rts Q6 ormalman ner

C!inese C!inese C!inese C!inese C!inese C!inese C!inese "merican "merican "merican "merican american "merican "merican american "merican "merican

2 2 2 1 1 3 1 2 1 2 1 2 1 2 2 2 2

1 2 2 1 1 1 1 2 1 4 1 2 1 2 2 2 2

2 2 2 2 2 2 1 4 3 4 4 3 2 3 2 4 1

2 1 2 1 2 2 1 2 1 2 1 1 2 2 2 1 4

1 1 2 1 1 4 2 1 2 3 1 1 1 2 3 2 1

5 4 2 2 2 4 4 3 2 4 1 3 4 2 5 1 4

&escriptive Analysis: $from . strongly agree to * strongly disagree%


n Q1Wisdom4 Q2Respect4 Q3Equal4 Q4rulesconduct4 Q5experts4 Q6formalmanner4 12# 12# 12% 12% 12% 12# Mean 1$%5& 1$#%5 2$66' 2$&31 1$%&1 3$&4% SD &$614# &$6'#% &$'6#2 &$#351 &$%5'2 1$&33#

Note !"i#ert!Scale!$as!to!%e!set!as!&ontinuous!Data!in!order!for!'nal(se)it!to!run!descripti*e!Statistics+!not! accurate!%ut!'ccepta%le,

3 Plenty of people treat ordinal data as continuous --- especially when survey data are supplemented with other measures

The rigorous analysis is to get a Weighted mean, which Analyse-it does not do. Often times, researchers go right into Inferential Statistics and Skip the Descripti e Statistics, since it is less informati e.!

:=nd of 0escripti*e &nalysis;

<nferential &nalysis on the differences among the three groups &merican) >hinese) 9orean" For later discussion KouAll try to create the charts posted below) using &nalyse$it"
1) Participants perceptions on teacher and teaching in general (pre-survey) : <tem 2 on the pre$ sur*ey assessed participantsA perceptions and e%pectations on teacher and teaching in general. ,he three (uestions that are closely related to sense of Gower 0istance were analy6ed inferentially with the 9ruskal$+allis &nalysis of Variance test) with cultural identity being the independent *ariable. ,he results indicate that: a" there were significant differences in participantsA perceptions about being e(ual with their instructor. ,he 9orean group had the highest mean rank 23.3/" on a scale of - strongly agree" to 3 strongly disagree". Ly contrast) the &nglo$&merican group had the lowest mean rank 2..JJ" and therefore percei*ed their instructors more as e(uals. b" ,here was no significant difference in participantsA perceptions about rules of conduct in online classes. ,he >hinese group had the lowest mean rank 2../D") an indication of a stronger agreement about implementing specific rules of conduct. ,his result aligned with some of their narrati*e comments about Mfeeling lostN and hoping for more guidance. &nd c" ,here was highly significant differences in their perceptions on course conduct. &gain the >hinese had the lowest mean rank) an indication of a stronger agreement about conducting courses in a formal manner. 2) Post-survey: approaching superior and peer when completing individual assignments and team work: Other Responses to the post sur*ey that reflect the impact of Gower 0istance include: a" LearnersA comfort le*el in approaching the instructor/facilitator/,& for help with indi*idual assignments and/or teamwork5 and b" ,heir comfort le*el in approaching the peers for help with indi*idual assignments and/or teamwork. Garticipants rated their comfort le*el from *ery comfortable -") to somewhat comfortable 2") uncomfortable /") and to *ery uncomfortable 2". ,he lower their mean rating) the higher their comfort le*el. 9ruskal$+allis &nalysis of Variance was used again to compare the mean differences in participantsA ranking of comfort le*el in approaching MsuperiorN or their peers) when completing indi*idual assignments and team work if applicable. 9Note: :ecause the regular Mean of Likert #cale does not make much sense7 I skipped the descriptive Analysis and ;ent right into Inferential Analysis'Analysis of <ariance using the Non-Parametric =ruskal-;allis statistic> <tem -. <ndi*idual &ssignment: &pproaching !uperior for 7elp two$tailed test" Rank 8ean O. <&: &pproach 4!uperior4 D n sum rank &mericanD /.30.0 /0.D3 >hineseD -3 DE2.3 23.30 9oreanD 2. -2-J.3 2-..E 9ruskal$+allis statistic J.-3 chis(r appro%imation) pD 0.02E0 corrected for ties" +hen the le*el of significance is set at 0.03 a") the small p *alue 0.02" indicates significant difference in participantsA rating for approaching MsuperiorN in indi*idual assignment. ,he

J
&merican group) not surprisingly) had the lowest mean rank /0.D3") an indication of greater comfort le*el in approaching the instructors for help5 and the >hinese group had the highest mean rank 23.30" and thus lower comfort le*el in approaching their instructors. <tem 2. <ndi*idual &ssignment: &pproaching Geer for 7elp two$tailed test" cases e%cluded: 2 due to n J/ missing *alues" G. <&: &pproach Geer by HroupJ n Rank sum 8ean rank &mericanJ /.--.3 2..20 >hineseJ -3 /J2.3 22.E/ 9oreanJ 2J -2-J.0 32.2E 9ruskal$+allis statistic 2D.2D chis(r appro%imation) pJ P0.000- corrected for ties" +hen aC0.03) the small p *alue P0.000-" indicates highly significant differences in participantsA comfort le*el in approaching peers for help with indi*idual assignments. ,he >hinese group had the lowest mean rank 22.E/$$higher comfort le*el") while the 9orean group had the lowest mean rank 32.2E$$lower comfort le*el". <tem /. ,eamwork: &pproaching !uperior for 7elp n Q. ,eam: &pproach 4!uperior4 by HroupJ &mericanJ >hineseJ 9oreanJ 9ruskal$+allis statistic pJ cases e%cluded: -J due to missing 3E *alues" n /-3 -2 2.EE Rank sum E-2.3 30..0 /EJ.3 8ean rank 2D.2J //../ /2.2.

chis(r appro%imation) corrected for 0.2/D2 ties"

GC0.2/D RaC0.03" indicates no significant difference in participantsA comfortableness in approaching superiors for help when completing teamwork. <tem 2. ,eamwork: &pproaching Geer for 7elp two$tailed test" cases e%cluded: -J due to missing n 3E *alues" V. ,eam: &pproach Geer by HroupJ n Rank sum 8ean rank &mericanJ /E0D.3 2D.02 >hineseJ -3 /-..3 2-./0 9oreanJ -2 3E3.0 2E.J3 9ruskal$+allis statistic 22.E0 chis(r appro%imation) corrected for pJ P0.000- ties" ,he high 9ruskal$+allis statistic 22.E" and the small p *alue P0.000-" again indicates highly significant difference in participantsA comfort le*el in approaching peers for help with team work. ,he 9orean group mean rank C 2E.J3" contributed greatly to this difference. 7owe*er) the statistical power might ha*e been reduced in this test because of the -J missing rating *alues

E
from the 9orean group. &s mentioned in the curriculum analysis) many of the 9orean courses did not in*ol*e team work and many chose Mnon applicableN for this sur*ey (uestion. Summary: In luence o power distance evidenced !y the our tests: >onforming to the e%isting findings about Gower 0istance) the &merican group mainly &nglo$&merican" had the lowest G0< score) while the >hinese group had the highest G0< score. Gossibly because of their sense of G0<) the &merican group felt the most comfortable in approaching their instructors for help) while the 9orean group felt most uncomfortable in doing so. >hinese students) because of their large class si6e) did not ha*e much opportunity to interact with the instructors. !till) their reported comfort le*el in approaching the instructors was low. &s to approaching their peers for help) the >hinese group felt the most comfortable in completing both indi*idual assignments and team work) the &merican group felt comfortable) while the 9orean group felt the least comfortable in completing both indi*idual assignments and teamwork. &gain) the 9oreansA cultural perceptions on >8> might ha*e influenced their ratings here. &s some of the 9orean participants commented) peers or classmates online can be Mstrangers.N &s to the high comfort le*el of the >hinese) it is worth noting that most of these >hinese students worked in self$formed teams and they therefore were comfortable about approaching their peers for help. ,he four 9ruskal$+allis analyses on the post$sur*ey items had re*ealing results. &lthough there was no significant difference in the three groupsA comfort le*el in approaching superiors for help with team work) there were significant differences in their rating for approaching superiors in indi*idual assignments) and there were highly significant differences in their le*els of comfort in approaching peers for help with indi*idual assignments and with team work. Gower 0istance indeed affected studentsA ways in approaching instructors and their peers. Ly contrast) indi*iduals were able to o*ercome their sense of Gower 0istance when working as a group. <n other words) indi*iduals became Mbra*erN when working as a team to approach their instructors for help. Irom ;ang?s @ultural #tudies of Anline Learning7 "ritish #ournal o $ducational %echnology"

You might also like