Manajemen | Fakultas Ekonomi Universitas Maritim Raja Ali Haji joeb.84.1.55-61
Journal of Education for Business
ISSN: 0883-2323 (Print) 1940-3356 (Online) Journal homepage: http://www.tandfonline.com/loi/vjeb20
On the Validity of RateMyProfessors.com
Thomas Timmerman
To cite this article: Thomas Timmerman (2008) On the Validity of RateMyProfessors.com,
Journal of Education for Business, 84:1, 55-61, DOI: 10.3200/JOEB.84.1.55-61
To link to this article: http://dx.doi.org/10.3200/JOEB.84.1.55-61
Published online: 07 Aug 2010.
Submit your article to this journal
Article views: 68
View related articles
Citing articles: 18 View citing articles
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=vjeb20
Download by: [Universitas Maritim Raja Ali Haji]
Date: 11 January 2016, At: 22:44
OntheValidityofRateMyProfessors.com
Downloaded by [Universitas Maritim Raja Ali Haji] at 22:44 11 January 2016
THOMASTIMMERMAN
TENNESSEETECHNOLOGICALUNIVERSITY
COOKEVILLE,TENNESSEE
ABSTRACT.Thepurposeofthisstudy
wastoaddresssomeofthemostcommon
questionsandconcernsregardingRateMy
Professors.com(RMP).Datafrom5differentuniversitiesand1,167facultymembers
showedthat(a)theratingsarenotdominatedbygriping,(b)thesummaryevaluation
correlateshighlywithsummaryevaluations
fromanofficialuniversityevaluation,(c)
substantiverelationsaregenerallythesame
whenonlyasingleratinghasbeenprovided,(d)therelationbetweenRMPEasiness
andRMPQualityispartiallyexplained
bythefactthatlearningisassociatedwith
perceivedeasiness,and(e)thesubstantive
findingsgeneralizetobusinessfacultyin
differentuniversities.Theauthordiscusses
thepossiblevalueofRMPwithoutendorsingitsunlimiteduseforadministrative
purposes.
Keywords:RateMyProfessors.com,student
evaluationofteaching,validity
Copyright©2008HeldrefPublications
M
uch to the surprise (and chagrin)
of some educators, an enterprising college student took a routine
administrative university function (i.e.,
faculty evaluation) and turned it into a
moneymaking venture. John Swapceinski founded RateMyProfessors.com
(RMP) in 1999 after a particularly bad
experience with a faculty member. As
of March 2007, more than 1 million
professorsfrommorethan6,000schools
hadreceivedmorethan6.75millionratings. Swapcienski is now turning his
sights to other professional fields such
as medicine and law through a network
ofsimilarratings-basedWebsites(www.
ratingz.net). Not surprisingly, RMP has
generated a great deal of controversy
in the academic world and a great deal
of publicity in the nonacademic world.
WiththegrowingpopularityofRMPand
similar sites, business faculty members
maybenefitfromempiricalevidencethat
addressesthemostcontroversialaspects
ofsuchratings.
Althoughotherstudieshaveaddressed
theintercorrelationsofRMPdimensions
(e.g.,Felton,Koper,Mitchell,&Stinson,
2006;Felton,Mitchell,&Stinson,2004;
Riniolo, Johnson, Sherman, & Misso,
2006), this study examined three new
questionsthatsurroundRMP.First,some
suggest that RMP ratings are biased by
a disproportionate number of negative
ratings.The most prevalent argument is
that students are not motivated to visit
the Web site and make a rating unless
they have had a bad experience with a
professor. If this is true, RMP ratings
should be positively skewed (i.e., more
negative ratings), and there should be a
negative correlation between the num-
ber of ratings received and the reported
qualityofaprofessor.Asecondquestion
concernstheoverallvalidityoftheRMP
ratings. How well do the RMP ratings
correlate with official student evaluations of teaching (SET) conducted by
universities?A final concern with RMP
isthemeaningofratingsderivedfroma
small sample of students. Critics argue
(andpsychometrictheoryconfirms)that
smaller samples are less reliable and
therefore less valid. Evidence about the
validityofRMPmayhelpbusinessfacultymembersdeterminehowtointerpret
their individual ratings. Such evidence
mayalsohelpbusinessprofessorsinform
the public about the value (if any) of
suchsites.
BACKGROUNDAND
HYPOTHESES
LongbeforeRMPcamealong,SETs
were controversial (Theall & Franklin,
2001).Arestudentsqualifiedtoevaluate
collegefaculty?Dostudentevaluations
simplyreflectexpectedgrades?Arestudent evaluations influenced by faculty
personality? With so many sources of
contamination, should student ratings
be used for any decision-making purposes? These are only a few of the
most common questions regarding student ratings. Despite the fact that an
abundance of research has addressed
these questions (e.g., Marsh & Roche,
2000), supporters and critics are not
easilyswayedbydatafromonesideto
theother.
The first question addressed in this
studyisthedistributionoftheratings.An
Associated Press article in 2003 quoted
September/October2008
55
Downloaded by [Universitas Maritim Raja Ali Haji] at 22:44 11 January 2016
AmericanAssociationofUniversityProfessorsspokespersonJonathonKnightas
claiming,“Thesekindof[RMP]postings
willinevitablyfocusonstudentgripesand
have no credibility.” If RMP ratings are
dominated by gripes, it could mean that
thesampleofstudentswhopostmaynot
berepresentativeofthelargerpopulation
ofstudents.Thesamplemayhaveagreaterproportionofinferior,unmotivated,or
mean students. Such a sample would be
easy to spot because the distribution of
ratingswouldexhibitpositiveskew.Traditional SETs typically exhibit negative
skew(Tagomori&Bishop,1995;Wolfer
&Johnson,2003).Thus,positiveskewin
the RMP data may be a good indicator
thatthesampleofratersatRMPisdifferentthanthesampleofraterswhocompletetraditionalSETs.Thusthefollowing
hypothesesareoffered:
Hypothesis1a(H1a):Ratingsofoverall quality from RateMyProfessor.com
arepositivelyskewed.
H1b: Ratings of overall quality from
officialSETsarenegativelyskewed.
Aside from these predictions, the
possibilitythatstudentsaremorelikely
to visit RMP to complain would also
be evidenced by a negative relation
between the number of ratings and the
average level of the ratings. In other
words,accordingtotheaforementioned
logic, faculty members with many ratings should have more gripers and
therefore lower average ratings. The
secondhypothesisfollows:
H2: There is a negative relation
between the number of RMP ratings
andthemeanlevelofoverallquality.
The second hypothesis concerns the
concurrent validity of RMP ratings. Do
RMPratingsofOverallqualitycorrelate
withsimilartypesofsummarymeasures
from official university SETs? If RMP
ratingsarepositivelyskewedandprimarily capture complaints, whereas official
ratings are negatively skewed and capture something else, there is no reason
to expect a positive relation between
RMPratingsandofficialSETs.Thethird
hypothesisisthenullhypothesis:
H3:Thereisnosignificantcorrelation
betweenRMPratingsofOverallquality
andsummarymeasuresinSETs.
56
JournalofEducationforBusiness
A positive correlation between RMP
ratings and SETs could imply that both
measure the same source of contamination (Pike, 1999). If all ratings reflect
theperceivedeasinessofteachersorthe
perceivedattractivenessofteachers(Feltonetal.,2006),apositiverelationmay
emerge that does not imply any sort of
construct validity. In this study, I start
withtheassumptionthatvalidatedSETs
are meaningful measures of teaching
effectiveness. This assumption is based
onthelargeamountofevidenceconnecting objectively assessed student learning with student ratings (Cohen, 1981;
Marsh,1987;Theall&Franklin,2001).
The third hypothesis, introduced earlier,concernstheconclusionsthatcanbe
drawnfromRMPratingsbasedonvery
few ratings. A central tenet in psychometrictheoryisthatreliabilityincreases
withthenumberofobservations.Therefore, the relation between RMP ratings
andSETsdiscussedpreviouslymayexist
butmaybeattenuatedbytheunreliability of the ratings from faculty members
withonlyafewRMPratings.Thus,the
finalhypothesisisoffered:
H4: Relations between RMP ratings
and SETs increase as the number of
RMPratingsincrease.
Previous research on RMP ratings is
limited to a few studies that explore the
intercorrelations among RMP dimensions.Forexample,Feltonandcolleagues
(Felton et al., 2004; Felton et al., 2006)
foundthatRMPratingsofOverallquality
(created by combining Helpfulness and
Clarity ratings) were strongly correlated
withRMPratingsofEasinessandPhysicalattractiveness(Hotness).Theauthors’
interpretationofthesefindingsisthatstudent ratings of quality are contaminated
by factors unrelated to the actual effectivenessoftheinstructor.Inotherwords,
quality ratings are caused by perceived
easiness and attractiveness. Felton et al.
(2006)thenconcluded“theseself-selected evaluations from RateMyProfessors.
com cast considerable doubt on the usefulnessofin-classstudentopinionsurveys
for purposes of examining quality and
effectivenessofteaching”(p.13).
Itisimportanttonotetwolimitations
of those studies. First, the studies by
Felton et al. (2004; 2006) only examined faculty with at least 20 ratings. If
studentsaremorelikelytovisitthesite
to complain, faculty with 20 or more
ratingsmaynotberepresentativeofthe
population of faculty.A second limitationisthatFeltonetal.(2006)dismiss
the possibility that causality operates
in the opposite direction. It is possible
thathigh-qualityinstructorsareviewed
aseasierandmoreattractivebecauseof
their competence. For example, physical attractiveness is correlated with
intelligence,self-confidence,andoccupational success (Langlois, Kalakanis,
Rubenstein,Larson,Hallam,&Smoot,
2000). Thus, it would not be surprisingtodiscoverthatcompetentteachers
are(a)moreattractiveor(b)perceived
as more attractive than less competent
teachers. Likewise, competent faculty
members may (a) present material in
a more accessible manner or (b) be
perceived as easier than less competent instructors. The most compelling
evidence against Felton et al.’s (2004)
andFeltonetal.’s(2006)position(i.e.,
Marsh & Roche, 2000) was not mentioned in their studies. Specifically,
Marsh and Roche found no evidence
thathighSETratingswerecontaminatedfromgradingleniency.Instead,they
found better evidence that high SET
ratingswerecausedbyactuallearning.
One way to test this hypothesis is to
partialouttheeffectofstudentlearning
fromtherelationbetweenOverallquality and perceived Easiness. The data
availableinthisstudymakeitpossible
totestthefollowinghypothesis:
H5:TherelationbetweenRMPOverall quality and RMP Easiness is mediatedbystudentlearning.
METHOD
The data required for this study
included a group of RMP ratings and
corresponding SETs. RMP ratings are
easytoharvestfromtheRMPWebsite.
A few universities now publicly post
their SET results on Web sites. In the
firstanalysis,Icollectedcorresponding
data from the University of California
at San Diego (UCSD; http://www.cape
.ucsd.edu). UCSD was chosen for no
otherreasonthantheconveniencewith
which its data are presented. In the
second analysis, I collected RMP and
official SET data from four additional
Downloaded by [Universitas Maritim Raja Ali Haji] at 22:44 11 January 2016
universities.Thepurposeofthesecond
analysis was to determine if the relations in the first analysis were generalizable to business faculty at various
universities. The trade-off in the second analysis was that different settings
wereexploredbutthedifferentsettings
involved smaller sample sizes and differentquestionsintheirSETs.Thefour
additionaluniversitiesweretheUniversity of Tennessee, University of ColoradoatDenver,UniversityofColorado
atBoulder,andUniversityofColorado
atColoradoSprings.
Thelevelofanalysisinthisstudywas
theindividualfacultymember.Therewas
noattempttoexaminefaculty-by-course
level data or changes in ratings during
the period of available data. The focus
hereisonthesummarydataprovidedby
RMPandtheofficialSETs.AtRMP,studentsratefacultyonthree5-pointscales:
Easiness, Helpfulness, and Clarity. The
summary provided by RMP combines
theHelpfulnessandClarityscalesintoan
Overallqualitydimension.
AnalysisOne
The student evaluation system at
UCSDisapaper-and-pencilassessment
used in most classes with 15 or more
enrolled students. At the time of this
study,summarydatawereavailablefor
8,523 course sections offered from the
fallof2002tothewinterof2006.These
sections were offered by 1,684 identifiable professors. RMP ratings were
availablefor1,229UCSDfacultymembers.Mergingthesetwosourcesofdata
produced 1,002 professors with both
RMPratingsandUCSDratings.
TheUCSDevaluationsummarycontainstheclassaverageoffouritems.Students indicate their agreement with the
statement “I learned a great deal from
this course” on a 5-point scale ranging
from1(stronglydisagree)to5(strongly
agree). Students also indicate the number of hours they spend studying for
the course per week. Last, students are
asked whether they would recommend
thiscourseandwhethertheywouldrecommendthisinstructortoothers.
The University of Colorado system
uses standardized questions across its
campuses.The most relevant items are
presented in Table 2. In all, 22 faculty
membersfromtheBouldercampus,38
fromtheColoradoSpringscampus,and
44fromtheDenvercampuswerepresentinbothRMPandtheofficialSET.
RESULTS
AnalysisOne
Descriptive statistics and intercorrelations for the UCSD data are presented in Table 3. A few descriptive
statistics are worth pointing out. As
the unit of analysis is the individual
faculty member (with data aggregated
across classes), the average number of
UCSD ratings is 365.70 per professor.
The average number of RMP ratings
is 11.15. The skewness of each rating
is provided to address the question of
whether students are more likely to
provide positive or negative ratings.
Consistent with previous research, the
percentageofstudentswhowouldrecommend the instructor is negatively
skewed. Contrary to critics’ concerns,
however, the RMP ratings of Overall quality are also negatively skewed.
AnalysisTwo
At the University of Tennessee, students are asked to rate the course and
instructoronavarietyof6-pointscales.
ThespecificitemsarepresentedinTable
1. In all, 61 business faculty members
from the University ofTennessee were
ratedinbothRMPandtheofficialSET.
TABLE1.IntercorrelationsfortheUniversityofTennessee(n=61)
Variable
RMP
Numberofratings
Overallquality
Easiness
SET
Numberofratings
Courseasawhole
Coursecontent
Instructor’scontributiontothecourse
Instructor’seffectivenessinteaching
Useofclasstime
Interestinwhetherstudentslearned
Amountyoulearnedinthecourse
Relevanceofcoursecontent
Evaluativeandgradingtechniques
Reasonablenessofassignedwork
Clarityofstudents’responsibilities
Correlations
M
SD
RMP
Numberof
ratings
RMP
Overall
quality
4.92
3.56
2.78
9.37
1.21
1.12
–.09
.04
.42**
583.89
3.45
3.47
3.72
3.53
3.48
3.60
3.38
3.53
3.31
3.51
3.52
633.20
0.57
0.49
0.63
0.70
0.64
0.56
0.57
0.52
0.65
0.54
0.62
.31*
–.06
–.06
–.04
–.02
–.03
–.07
–.05
–.03
.00
.06
.04
–.15
.61**
.60**
.65**
.67**
.49**
.59**
.63**
.60**
.43**
.52**
.41**
RMP
Easiness
–.02
.31*
.27*
.29*
.35**
.34**
.33*
.30*
.31*
.44**
.47**
.44**
Note.RMP=RateMyProfessors.com;SET=studentevaluationsofteaching.
*
p
ISSN: 0883-2323 (Print) 1940-3356 (Online) Journal homepage: http://www.tandfonline.com/loi/vjeb20
On the Validity of RateMyProfessors.com
Thomas Timmerman
To cite this article: Thomas Timmerman (2008) On the Validity of RateMyProfessors.com,
Journal of Education for Business, 84:1, 55-61, DOI: 10.3200/JOEB.84.1.55-61
To link to this article: http://dx.doi.org/10.3200/JOEB.84.1.55-61
Published online: 07 Aug 2010.
Submit your article to this journal
Article views: 68
View related articles
Citing articles: 18 View citing articles
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=vjeb20
Download by: [Universitas Maritim Raja Ali Haji]
Date: 11 January 2016, At: 22:44
OntheValidityofRateMyProfessors.com
Downloaded by [Universitas Maritim Raja Ali Haji] at 22:44 11 January 2016
THOMASTIMMERMAN
TENNESSEETECHNOLOGICALUNIVERSITY
COOKEVILLE,TENNESSEE
ABSTRACT.Thepurposeofthisstudy
wastoaddresssomeofthemostcommon
questionsandconcernsregardingRateMy
Professors.com(RMP).Datafrom5differentuniversitiesand1,167facultymembers
showedthat(a)theratingsarenotdominatedbygriping,(b)thesummaryevaluation
correlateshighlywithsummaryevaluations
fromanofficialuniversityevaluation,(c)
substantiverelationsaregenerallythesame
whenonlyasingleratinghasbeenprovided,(d)therelationbetweenRMPEasiness
andRMPQualityispartiallyexplained
bythefactthatlearningisassociatedwith
perceivedeasiness,and(e)thesubstantive
findingsgeneralizetobusinessfacultyin
differentuniversities.Theauthordiscusses
thepossiblevalueofRMPwithoutendorsingitsunlimiteduseforadministrative
purposes.
Keywords:RateMyProfessors.com,student
evaluationofteaching,validity
Copyright©2008HeldrefPublications
M
uch to the surprise (and chagrin)
of some educators, an enterprising college student took a routine
administrative university function (i.e.,
faculty evaluation) and turned it into a
moneymaking venture. John Swapceinski founded RateMyProfessors.com
(RMP) in 1999 after a particularly bad
experience with a faculty member. As
of March 2007, more than 1 million
professorsfrommorethan6,000schools
hadreceivedmorethan6.75millionratings. Swapcienski is now turning his
sights to other professional fields such
as medicine and law through a network
ofsimilarratings-basedWebsites(www.
ratingz.net). Not surprisingly, RMP has
generated a great deal of controversy
in the academic world and a great deal
of publicity in the nonacademic world.
WiththegrowingpopularityofRMPand
similar sites, business faculty members
maybenefitfromempiricalevidencethat
addressesthemostcontroversialaspects
ofsuchratings.
Althoughotherstudieshaveaddressed
theintercorrelationsofRMPdimensions
(e.g.,Felton,Koper,Mitchell,&Stinson,
2006;Felton,Mitchell,&Stinson,2004;
Riniolo, Johnson, Sherman, & Misso,
2006), this study examined three new
questionsthatsurroundRMP.First,some
suggest that RMP ratings are biased by
a disproportionate number of negative
ratings.The most prevalent argument is
that students are not motivated to visit
the Web site and make a rating unless
they have had a bad experience with a
professor. If this is true, RMP ratings
should be positively skewed (i.e., more
negative ratings), and there should be a
negative correlation between the num-
ber of ratings received and the reported
qualityofaprofessor.Asecondquestion
concernstheoverallvalidityoftheRMP
ratings. How well do the RMP ratings
correlate with official student evaluations of teaching (SET) conducted by
universities?A final concern with RMP
isthemeaningofratingsderivedfroma
small sample of students. Critics argue
(andpsychometrictheoryconfirms)that
smaller samples are less reliable and
therefore less valid. Evidence about the
validityofRMPmayhelpbusinessfacultymembersdeterminehowtointerpret
their individual ratings. Such evidence
mayalsohelpbusinessprofessorsinform
the public about the value (if any) of
suchsites.
BACKGROUNDAND
HYPOTHESES
LongbeforeRMPcamealong,SETs
were controversial (Theall & Franklin,
2001).Arestudentsqualifiedtoevaluate
collegefaculty?Dostudentevaluations
simplyreflectexpectedgrades?Arestudent evaluations influenced by faculty
personality? With so many sources of
contamination, should student ratings
be used for any decision-making purposes? These are only a few of the
most common questions regarding student ratings. Despite the fact that an
abundance of research has addressed
these questions (e.g., Marsh & Roche,
2000), supporters and critics are not
easilyswayedbydatafromonesideto
theother.
The first question addressed in this
studyisthedistributionoftheratings.An
Associated Press article in 2003 quoted
September/October2008
55
Downloaded by [Universitas Maritim Raja Ali Haji] at 22:44 11 January 2016
AmericanAssociationofUniversityProfessorsspokespersonJonathonKnightas
claiming,“Thesekindof[RMP]postings
willinevitablyfocusonstudentgripesand
have no credibility.” If RMP ratings are
dominated by gripes, it could mean that
thesampleofstudentswhopostmaynot
berepresentativeofthelargerpopulation
ofstudents.Thesamplemayhaveagreaterproportionofinferior,unmotivated,or
mean students. Such a sample would be
easy to spot because the distribution of
ratingswouldexhibitpositiveskew.Traditional SETs typically exhibit negative
skew(Tagomori&Bishop,1995;Wolfer
&Johnson,2003).Thus,positiveskewin
the RMP data may be a good indicator
thatthesampleofratersatRMPisdifferentthanthesampleofraterswhocompletetraditionalSETs.Thusthefollowing
hypothesesareoffered:
Hypothesis1a(H1a):Ratingsofoverall quality from RateMyProfessor.com
arepositivelyskewed.
H1b: Ratings of overall quality from
officialSETsarenegativelyskewed.
Aside from these predictions, the
possibilitythatstudentsaremorelikely
to visit RMP to complain would also
be evidenced by a negative relation
between the number of ratings and the
average level of the ratings. In other
words,accordingtotheaforementioned
logic, faculty members with many ratings should have more gripers and
therefore lower average ratings. The
secondhypothesisfollows:
H2: There is a negative relation
between the number of RMP ratings
andthemeanlevelofoverallquality.
The second hypothesis concerns the
concurrent validity of RMP ratings. Do
RMPratingsofOverallqualitycorrelate
withsimilartypesofsummarymeasures
from official university SETs? If RMP
ratingsarepositivelyskewedandprimarily capture complaints, whereas official
ratings are negatively skewed and capture something else, there is no reason
to expect a positive relation between
RMPratingsandofficialSETs.Thethird
hypothesisisthenullhypothesis:
H3:Thereisnosignificantcorrelation
betweenRMPratingsofOverallquality
andsummarymeasuresinSETs.
56
JournalofEducationforBusiness
A positive correlation between RMP
ratings and SETs could imply that both
measure the same source of contamination (Pike, 1999). If all ratings reflect
theperceivedeasinessofteachersorthe
perceivedattractivenessofteachers(Feltonetal.,2006),apositiverelationmay
emerge that does not imply any sort of
construct validity. In this study, I start
withtheassumptionthatvalidatedSETs
are meaningful measures of teaching
effectiveness. This assumption is based
onthelargeamountofevidenceconnecting objectively assessed student learning with student ratings (Cohen, 1981;
Marsh,1987;Theall&Franklin,2001).
The third hypothesis, introduced earlier,concernstheconclusionsthatcanbe
drawnfromRMPratingsbasedonvery
few ratings. A central tenet in psychometrictheoryisthatreliabilityincreases
withthenumberofobservations.Therefore, the relation between RMP ratings
andSETsdiscussedpreviouslymayexist
butmaybeattenuatedbytheunreliability of the ratings from faculty members
withonlyafewRMPratings.Thus,the
finalhypothesisisoffered:
H4: Relations between RMP ratings
and SETs increase as the number of
RMPratingsincrease.
Previous research on RMP ratings is
limited to a few studies that explore the
intercorrelations among RMP dimensions.Forexample,Feltonandcolleagues
(Felton et al., 2004; Felton et al., 2006)
foundthatRMPratingsofOverallquality
(created by combining Helpfulness and
Clarity ratings) were strongly correlated
withRMPratingsofEasinessandPhysicalattractiveness(Hotness).Theauthors’
interpretationofthesefindingsisthatstudent ratings of quality are contaminated
by factors unrelated to the actual effectivenessoftheinstructor.Inotherwords,
quality ratings are caused by perceived
easiness and attractiveness. Felton et al.
(2006)thenconcluded“theseself-selected evaluations from RateMyProfessors.
com cast considerable doubt on the usefulnessofin-classstudentopinionsurveys
for purposes of examining quality and
effectivenessofteaching”(p.13).
Itisimportanttonotetwolimitations
of those studies. First, the studies by
Felton et al. (2004; 2006) only examined faculty with at least 20 ratings. If
studentsaremorelikelytovisitthesite
to complain, faculty with 20 or more
ratingsmaynotberepresentativeofthe
population of faculty.A second limitationisthatFeltonetal.(2006)dismiss
the possibility that causality operates
in the opposite direction. It is possible
thathigh-qualityinstructorsareviewed
aseasierandmoreattractivebecauseof
their competence. For example, physical attractiveness is correlated with
intelligence,self-confidence,andoccupational success (Langlois, Kalakanis,
Rubenstein,Larson,Hallam,&Smoot,
2000). Thus, it would not be surprisingtodiscoverthatcompetentteachers
are(a)moreattractiveor(b)perceived
as more attractive than less competent
teachers. Likewise, competent faculty
members may (a) present material in
a more accessible manner or (b) be
perceived as easier than less competent instructors. The most compelling
evidence against Felton et al.’s (2004)
andFeltonetal.’s(2006)position(i.e.,
Marsh & Roche, 2000) was not mentioned in their studies. Specifically,
Marsh and Roche found no evidence
thathighSETratingswerecontaminatedfromgradingleniency.Instead,they
found better evidence that high SET
ratingswerecausedbyactuallearning.
One way to test this hypothesis is to
partialouttheeffectofstudentlearning
fromtherelationbetweenOverallquality and perceived Easiness. The data
availableinthisstudymakeitpossible
totestthefollowinghypothesis:
H5:TherelationbetweenRMPOverall quality and RMP Easiness is mediatedbystudentlearning.
METHOD
The data required for this study
included a group of RMP ratings and
corresponding SETs. RMP ratings are
easytoharvestfromtheRMPWebsite.
A few universities now publicly post
their SET results on Web sites. In the
firstanalysis,Icollectedcorresponding
data from the University of California
at San Diego (UCSD; http://www.cape
.ucsd.edu). UCSD was chosen for no
otherreasonthantheconveniencewith
which its data are presented. In the
second analysis, I collected RMP and
official SET data from four additional
Downloaded by [Universitas Maritim Raja Ali Haji] at 22:44 11 January 2016
universities.Thepurposeofthesecond
analysis was to determine if the relations in the first analysis were generalizable to business faculty at various
universities. The trade-off in the second analysis was that different settings
wereexploredbutthedifferentsettings
involved smaller sample sizes and differentquestionsintheirSETs.Thefour
additionaluniversitiesweretheUniversity of Tennessee, University of ColoradoatDenver,UniversityofColorado
atBoulder,andUniversityofColorado
atColoradoSprings.
Thelevelofanalysisinthisstudywas
theindividualfacultymember.Therewas
noattempttoexaminefaculty-by-course
level data or changes in ratings during
the period of available data. The focus
hereisonthesummarydataprovidedby
RMPandtheofficialSETs.AtRMP,studentsratefacultyonthree5-pointscales:
Easiness, Helpfulness, and Clarity. The
summary provided by RMP combines
theHelpfulnessandClarityscalesintoan
Overallqualitydimension.
AnalysisOne
The student evaluation system at
UCSDisapaper-and-pencilassessment
used in most classes with 15 or more
enrolled students. At the time of this
study,summarydatawereavailablefor
8,523 course sections offered from the
fallof2002tothewinterof2006.These
sections were offered by 1,684 identifiable professors. RMP ratings were
availablefor1,229UCSDfacultymembers.Mergingthesetwosourcesofdata
produced 1,002 professors with both
RMPratingsandUCSDratings.
TheUCSDevaluationsummarycontainstheclassaverageoffouritems.Students indicate their agreement with the
statement “I learned a great deal from
this course” on a 5-point scale ranging
from1(stronglydisagree)to5(strongly
agree). Students also indicate the number of hours they spend studying for
the course per week. Last, students are
asked whether they would recommend
thiscourseandwhethertheywouldrecommendthisinstructortoothers.
The University of Colorado system
uses standardized questions across its
campuses.The most relevant items are
presented in Table 2. In all, 22 faculty
membersfromtheBouldercampus,38
fromtheColoradoSpringscampus,and
44fromtheDenvercampuswerepresentinbothRMPandtheofficialSET.
RESULTS
AnalysisOne
Descriptive statistics and intercorrelations for the UCSD data are presented in Table 3. A few descriptive
statistics are worth pointing out. As
the unit of analysis is the individual
faculty member (with data aggregated
across classes), the average number of
UCSD ratings is 365.70 per professor.
The average number of RMP ratings
is 11.15. The skewness of each rating
is provided to address the question of
whether students are more likely to
provide positive or negative ratings.
Consistent with previous research, the
percentageofstudentswhowouldrecommend the instructor is negatively
skewed. Contrary to critics’ concerns,
however, the RMP ratings of Overall quality are also negatively skewed.
AnalysisTwo
At the University of Tennessee, students are asked to rate the course and
instructoronavarietyof6-pointscales.
ThespecificitemsarepresentedinTable
1. In all, 61 business faculty members
from the University ofTennessee were
ratedinbothRMPandtheofficialSET.
TABLE1.IntercorrelationsfortheUniversityofTennessee(n=61)
Variable
RMP
Numberofratings
Overallquality
Easiness
SET
Numberofratings
Courseasawhole
Coursecontent
Instructor’scontributiontothecourse
Instructor’seffectivenessinteaching
Useofclasstime
Interestinwhetherstudentslearned
Amountyoulearnedinthecourse
Relevanceofcoursecontent
Evaluativeandgradingtechniques
Reasonablenessofassignedwork
Clarityofstudents’responsibilities
Correlations
M
SD
RMP
Numberof
ratings
RMP
Overall
quality
4.92
3.56
2.78
9.37
1.21
1.12
–.09
.04
.42**
583.89
3.45
3.47
3.72
3.53
3.48
3.60
3.38
3.53
3.31
3.51
3.52
633.20
0.57
0.49
0.63
0.70
0.64
0.56
0.57
0.52
0.65
0.54
0.62
.31*
–.06
–.06
–.04
–.02
–.03
–.07
–.05
–.03
.00
.06
.04
–.15
.61**
.60**
.65**
.67**
.49**
.59**
.63**
.60**
.43**
.52**
.41**
RMP
Easiness
–.02
.31*
.27*
.29*
.35**
.34**
.33*
.30*
.31*
.44**
.47**
.44**
Note.RMP=RateMyProfessors.com;SET=studentevaluationsofteaching.
*
p