Behavior Research Methods, Instruments, & Computers
1999,31 (4), 572-577
Web-based administration
of a personality questionnaire:
Comparison with traditional methods
ROBERT N. DAVIS
University ofMichigan, Ann Arbor, Michigan
The World-Wide Webholds great promise as a meehanism for questionnaire-based research. But are
data from Web-based questionnaires eomparable to data from standard paper-and-peneil questionnaires?
This study assessed the equivalenee ofthe Ruminative Responses Seale in a Web-based format
and in a paper-and-peneil format among introductory psyehology, upper-level psyehology, and nonpsyehology
students. Internal eonsisteney eoefficients were eomparable aeross the groups. The participants
in the Web sample reported higher levels of self-foeused rumination than did the other groups.
Women in the Web sample reported more self-foeused rumination than did women in the other groups.
In the Websample, results did not eovary with aeeess loeation. These results suggest that findings from
Web-based questionnaire research are eomparable with results obtained using standard proeedures.
The eomputerized Web interface may also facilitate self-disclosure among research participants.
The use ofcomputers in psychological assessment has
increased steadily over the last 30 years (Burke, 1992).
The advent ofthe World-Wide Web otTers unprecedented
opportunities for researchers to benefit from the use of
computers in data collection (Schmidt, 1997; Smith &
Leigh, 1997). As a research medium, the Web offers researchers
at least four major benefits. First, the Web permits
individuals to send data to a researcher at their convenience,
in terms of time and location, which increases
the potential number of eligible research participants
(Schmidt, 1997; Smith & Leigh, 1997). Second, the Web
permits automatie transformation ofraw data into an analyzable
format, such as an SPSS data file, using procedures
such as CGI scripts (see Kieley, 1996; Schmidt,
Hoffman, & MacDonald, 1997). Third, Web-based research
is etTicient in terms oftime and the resources it requires
(Schmidt, 1997). Fourth, the Web provides a degree
of anonymity to research participants, which decreases
the influence of demand characteristics by facilitating
self-disclosure and eliminating observer bias (Esposito,
Agard, & Rosnow, 1984; Hewson, Laurent, & Vogel,
1996; Meszaros, Engelsmann, Meterissian, & Kusalic,
1995; Smith & Leigh, 1997).
Is research conducted over the Web equivalent to research
conducted in more traditional settings, such as a
psychology laboratory? Validity studies ofWeb-based research
have recently begun to examine this question and
This research was supported by U.S. Public Health Grant 51817 to
Susan Nolen-Hoeksema. I thank Michael Bishop, Jamie Polito, and Andrew
Ward for their assistance with data collection. I also thank Tom
Buchanan for his helpful comments on an earlier draft of this manuscript.
Correspondence concerning this article should be addressed to
R. N, Davis,who isnow at the University ofHouston, Department ofPsychology,
Houston, TX 77204-5341 (e-mail: robnd@uh.edu).
have reported favorable results (e.g., Krantz, Ballard, &
Scher, 1997). For example, the results from one experiment
that was conducted both on the Web and in a laboratory
were nearly perfectly correlated (Krantz et al.,
1997). Although only a few published studies have examined
the equivalence question with respect to Webbased
versus paper-and-pencil administration ofpersonality
questionnaires, nascent studies in this area have
reported comparable results across domains (Buchanan
& Smith, in press; Pasveer & Ellard, 1998; Smith & Leigh,
1997; Stanton, 1998). For example, Buchanan and Smith
administered the revised version ofthe Self-Monitoring
Scale (SMS-R; Gangestad & Snyder, 1985) to Web users
on line and also to a local student sampie in paper-andpencil
format. The factor structure ofthe SMS-R and coetTicient
alphas were found to be comparable across the
groups, with the Web version yielding some superior results
(e.g., coefficient alpha), as compared with previously
reported studies (Buchanan & Smith, 1999). Another
study that compared Web-based and paper-andpencil
vers ions ofan organizationaljustice scale found a
similar factor structure across the two domains, as weil
as more item variability and less missing data in the Webbased
condition (Stanton, 1998).
The advent ofWeb-based data collection is a direct extension
ofdata collection on stand-alone computers (Buchanan
& Smith, in press). Given the current dearth of
findings pertaining to Web-based data collection, studies
of computerized (non-Web-based) versus paper-andpencil
data collection procedures may provide useful information.
Research comparing computerized and paperand-pencil
assessments has generally found that research
participants prefer computerized assessments and are
equally, ifnot more, willing to report sensitive information
to a computer than to a paper-and-pencil questionnaire
Copyright 1999 Psychonomic Society, Inc. 572
WEB-BASED ADMINISTRATION OF A PERSONALITY QUESTIONNAIRE 573
or to a human interviewer (Booth-Kewley, Edwards, &
Rosenfeld, 1992; Evan & Miller, 1969; Finegan & Allen,
1994; Hile & Adkins, 1997; King & Miles, 1995; Millstein,
1987; Petrie & Abell, 1994; Wilson, Genco, & Yager,
1986). In one study, recently admitted parasuicide patients
preferred a computerized interview to an interview
with a clinician (Petrie & Abell, 1994). In another study,
recent suicide attempters disclosed more information to
computerized depression and suicidal ideation questionnaires
than to a clinician interviewer (Levine, Ancill, &
Roberts, 1989). Moreover, the computerized measures
were better predictors of suicidality than was the clinician
interviewer (Levine et al., 1989). People also appear
to be more willing to report behaviors such as substance
use to a computer than to a clinician or a paper-and-pencil
questionnaire (Lucas, 1977; Lucas, Mullin, Luna, & McInroy,
1977; Skinner & Allen, 1983). Although these
findings are informative, the extent to which they generalize
to Web-based research is an open question.
Despite the initial promising evidence in support of
Web-based psychological research (e.g., Buchanan &
Srnith, in press; Krantz et al., 1997; Michalak, 1998;
Pasveer & Ellard, 1998; Smith & Leigh, 1997; Stanton,
1998), it is necessary to demonstrate the equivalence of
Web-based and traditional data collection methods empirically
before adopting Web-based methods extensively
(Krantz et al., 1997; Smith & Leigh, 1997). There
remain numerous potential threats to the reliability and
validity ofWeb-based questionnaires (see Hewson et al.,
1996; Schrnidt, 1997, for reviews). For example, participants
recruited from the Web are likely to be more heterogeneous
on demographie and other variables than participants
recruited by traditional means (Buchanan &
Smith, in press). Moreover, researchers lose considerable
control over the testing environment when conducting
Web-based research.Some participants may send
data from a quiet roorn, whereas others might send it
from a noisy computer laboratory (Reips, 1996). In addition,
Web-based research is open to exploitation by peopie
who participate numerous times and/or provide disingenuous
data (Schmidt, 1997; Smith & Leigh, 1997).
The primary concern underlying these possibilities is
that factors inherent in the Web interface, the Web sample,
and/or the testing environment might obfuscate the
extent to which an otherwise useful measure with good
psychometrie properties is able to measure the construct
of interest. In light of these concerns, several relevant
analyses will be performed in the present study (see
below).
The present study assessed the equivalence of a Webbased
versus a paper-and-pencil administration ofa personality
questionnaire commonly used as a screening instrument:
the Ruminative Responses Scale (RRS) from the
Response Styles Questionnaire (RSQ; Nolen-Hoeksema
& Morrow, 1991). The screening of participants for research
and their subsequent selection on the basis of inclusion
and exclusion criteria is a common practice among
researchers on college campuses and in medical settings.
Participants scoring at the upper and/or the lower end of
an overall distribution on variables of interest are frequently
selected for follow-up studies on the basis ofresults
from an initial screening session-for example, a
tendency to engage in self-focused rumination when experiencing
dysphoric affect (cf. Davis & Nolen-Hoeksema,
in press). These procedures facilitate research on interactions
between subject variables, such as personality
traits, and experimental treatments (cf. Lyubomirsky &
Nolen-Hoeksema, 1993, 1995), as weil as cross-sectional
studies of variables on which an experimenter wants to
select participants (e.g., age or gender).
The measure used in the present study assesses the extent
to which people tend to respond to feelings of sadness
or depression with self-focused rumination. Selffocused
rumination describes a tendency to engage in
passive thoughts and behaviors that focus one's attention
on one's depressed mood and on the implications ofthese
symptoms,rather than taking action to alleviate one's
symptoms (see Nolen-Hoeksema, 1990, 1991,for reviews).
People who ruminate may tend to sit alone and think
about how tired and unmotivated they feel or passively
review all the negative happenings in their lives, without
taking action to change their situation. The RSQ (NolenHoeksema
& Morrow, 1991) was developed to assess people's
habitual responses to depressed mood, and the RRS
consists of aseries of items on the RSQ that assess the
tendency to engage in self-focused rumination. A short
(IO-item) form of the RRS was recently developed and
was used in the present study (cf. Davis & Nolen-Hoeksema,
in press).
In the present study, the RRS was administered to four
sampies ofparticipants: (I) introductory psychology students;
(2) upper-Ievel psychology students; (3) nonpsychology
students; and (4) a Web-based studentsampie.
The nonpsychology and upper-Ievel psychology student
groups were included in order to determine whether they
differed from introductory psychology students even
when using the same methods (i.e., paper and pencil) for
filling out the questionnaire. The Web sampie was deliberately
drawn only from the local student population,
in order to ensure comparability with the other groups
on major demographie variables. This procedure enables
a comparison of group differences as a function of the
assessment medium (i.e., Web vs. paper and pencil), while
Iimiting the heterogeneity that might result from a general
Web sampie. Moreover, within the Web sampIe itself,
we analyzed possible differences as a function of
the location from which the participants accessed the
Web (e.g., at horne vs. a campus computer laboratory).
These analyses were performed to determine whether
concerns about loss ofexperimental control over the testing
setting were warranted-whether responses to the
questionnaire differed systematically as a function ofthe
participant's location.
It was predicted that the Web sampIe would report a
greater tendency to engage in self-focused rumination, as
compared with the other three groups. These predictions
574 DAVIS
were based on previous findings that people are generally
more willing to report sensitive information to a computer
than to a paper-and-pencil questionnaire. Moreover,
it was expected that decreased observer bias and increased
anonymity would facilitate self-disclosure in the Web
sampIe. Gender was also included as an independent
variable, in light of previous findings that women are
more likely to ruminate when depressed than are men (e.g.,
Schwartz & Koenig, 1996; Thayer, Newman, & McClain,
1994). Moreover, there are gender differences in Web
usage (GVU Center, 1998). Thus, the possible interaction
between sampIe group and gender was also examined.
METHOD
Participants
Four groups of individuals participated in the study: (I) 128 students
(56 males, 72 females) from the Web; (2) 118 students
(46 males, 72 females) from upper-Ievel psychology courses;
(3) 113students(50 males, 63 females) fromnonpsychology courses;
and (4) a total of 1,012 students (427 males, 585 females) from introductory
psychology courses. The participants ranged in age from
17 to 26 years and were drawn from the University of Michigan
campus. In the nonpsychology sampie, participants were recruited
from three lecture courses: History of the United States: 1865present;
Planets and Moons (Geology); and Introductory Biology.
The upper-Ievel psychology sampie was also recruited from three
lecture courses: Personality Psychology, Social Psychology, and
Abnormal Psychology. The Web participants were recruited from
flyers posted around the university campus that included the study's
URL (Internet address). A Pearson chi-square test indicated that the
groups did not differsignificantly in sex composition [X2(3) = .836,
p>.IO,N= 1,371].
Materials
Demographies questionnaire. The participants filled out a
one-page demographies questionnaire and provided information
about their age, sex, native language, and contact information.
Ruminative Responses Scale. A shortened (10-item) form of
the RRS ofthe RSQ was used in this study. This scale assesses how
participants tend to respond to their own symptoms of negative
emotion. The instructions for this questionnaire read:
People think and do many different things when they feel sad, blue, or
depressed. Please read each of the items below and indicate whether
you never, sometimes, often, or always think or do each one when you
feeI sad, down, or depressed. Please indicate what you generally do, not
what you think you should do.
The original RRS includes 22 items describing responses to depressed
mood that are self-focused (e.g., I think "Why do I react
this way?"), symptom-focused (e.g., I think about how hard it is to
concentrate), and focused on the possible causes and consequences
of one's mood (e.g., I think "I won't be able to do my job if I don't
snap out of this"). Scores on this scale show high test-retest reliability
(r = .80 at 6 months, r = .66 at I year; Nolen-Hoeksema &
Larson, 1998; Nolen-Hoeksema, Parker, & Larson, 1994), internal
consistency (Cronbach's alpha = .89; Nolen-Hoeksema & Morrow,
1991), and acceptable convergent and predictive validity (Butler &
Nolen-Hoeksema, 1994;Nolen-Hoeksema & Morrow, 1991; NolenHoeksema
et al., 1994). The 10 items on this shortened scale were
based on item analyses conducted with a community sampie of
1,122 adults (Nolen-Hoeksema, unpublished data). The 10 items
from the RRS that correlated most strongly with total scores on the
longer scale, and on which at least 15% ofthe sample endorsed an
answer other than never, were chosen for inclusion on the shortened
scale. The internal consistency (Cronbach's alpha) ofthis 10-item
scale in the community sampie of 1,122 was .87, and the correlation
between this IO-item scale and the total 22-item scale was .93.
Procedure
Paper administration. The investigators received permission
from instructorsin large lecturecoursesto visit their respectiveclasses
(listed above) in order to solicit student participation in a research
study. Both the course lecturers and the participants were blind as
to the purpose of the study. In all cases, all but one to three students
in each lecture course participated in the study. All the participants
provided informed consent and received adebriefing after completing
the questionnaire. The participants in the introductory psychology
class received course credit for completing the forms, and those
in the other classes were paid $5 for completing the forms. In addition,
the participants were instructed to print and sign their names
on a research recruitment sheet at the end of the questionnaire
packet ifthey were interested in possibly being selected to participate
in other studies for pay.
Web administration. The participants responded to flyers on
campus advertising a research study that required visiting a Web
site and filling out a questionnaire on line. Interested participants
tore off a piece of the flyer that included the study's URL and then
used a Web browser to visit the Web site constructed for questionnaire
administration. The Web site included a copy of the consent
form, as weil as the demographies questionnaire and RRS. The Web
page also informed participants that they would have a chance of
being selected to participate in another study for pay by filling out
the questionnaires. Thus, the participants in each ofthe four groups
had some incentive for completing the forms. The participants
typed the information requested by the demographie questionnaire
onto the Web page and used a mouse to click on their responses to
the RRS. Only one response per questionnaire item was permitted,
and the participants had the ability to scroll up or down on the Web
page. After filling out the questionnaires, the participants used a
mouse to click on one of two text boxes, "Send Form" or "Clear
Form."Clicking on "Send Form" sent the data to the experimenter
by means of e-mail. When this occurred, the participants then saw
a screen indicating that their message had been sent successfully.
Clicking on "Clear Form" erased the data completely. This option
was provided in the event that participants opted not to submit their
data or wished to clear the page and start over. The Web page forms
were converted to e-mail message formatusing the program Htmail
(Disser, 1996).
In light of concerns that the participants from the lecture classes
could also have participated in the Web portion of the study, and
given that it was possible for some people to complete the questionnaire
more than once on line, research assistants checked the data
before the analyses to ensure that no data were entered for the same
participant twice. These data-cleaning procedures resulted in the
deletion of data from 2 participants who completed the questionnaire
on line more than once (entries following the first instance of
participation were viewed as spurious). Data were also deleted from
I participant who completed the questionnaires both on line and in
two of her classes. These procedures resulted in four independent
samples.
RESULTS
Interna) Consistency
To examine the psychometrie properties of the Webbased
RRS, as compared with the paper-and-pencil version,
the internal consistency (Cronbach's alpha) of the
RRS was assessed separately for the four sampIes. Cronbach
's alphas for the RRS were as folIows: .82 (Web sam-
WEB-BASED ADMINISTRATION OF A PERSONALITY QUESTIONNAIRE 575
pie), .88 (upper-Ievel psychology students), .88 (nonpsychology
students), and .83 (introductory psychology
students).
Group and Gender Differences
Sampie group and gender differences were assessed,
using a 2 (gender) X 4 (group) analysis of variance
(ANOVA), with sampie group (Web sampie, upper-Ievel
psychology students, nonpsychology students, and introductory
psychology students) and gender as independent
variables and RRS total scores as dependent variables.
Table I presents the means and standard deviations for
all scores by sampIe group and gender. There were significant
group differences on RRS scores [F(3, 1363) =
4.23, p < .01]. Post hoc follow-up tests, using Tukey's
honestly significant differences (HSD) tests, revealed
that students in the Web sampIe reported levels ofrumination
that were significantly higher than those reported
by students in both introductory psychology (p < .0I) and
nonpsychology (p < .0I) courses, but not those reported
by students in upper-Ievel psychology courses (p > .10).
Significant gender differences were also obtained on RRS
scores [F( I, 1363) = 8.47,p< .01], indicating that women
reported significantly greater rumination than did men.
In addition, there was a nonsignificant trend toward a
group X gender interaction on RRS scores [F(3, 1363) =
2.22,p< .09]. Follow-up one-way ANOVAs within each
gender separately revealed no significant group differences
among males [F(3,575) = 0.53, p > .10], but did
reveal significant group differences among females
[F(3,788) = 6.70,p < .0001]. Post hoc follow-up tests,
using Tukey's HSD, revealed that women in the Web
sampIe reported significantly greater rumination than
did women in introductory psychology (p < .0001) and
nonpsychology (p < .0001) courses. Women in the Web
sampIe also reported greater rumination than did women
in upper-Ievel psychology courses, but the difference
was not statistically significant (p < .08).
Analysis of Web Data by Access Location
The 128participants who filled out the RRS on the Web
opted to do so from a variety oflocations across campus.
The server program used in this study (Htmail) also automatically
detected and reported the terminal location
from which the data originated. In the Web sampIe, 13%
(n = 17) used a modem to access the Web page from
horne, 49% (n = 63) accessed the Web page from a computing
site on campus, and 38% (n = 48) accessed the
Table I
Means and Standard Deviations of Scores
on the Ruminative Responses Scale
Total Men Wornen
Group M SD M SD M SD
Web 11.31 4.76 9.64 4.47 12.61 4.60
Psychology 10.41 5.03 10.02 4.30 10.65 5.45
Introductory psychology 9.83 4.91 9.45 4.97 10.11 4.85
Nonpsychology 9.04 4.97 8.80 4.52 9.24 5.32
Web page from a computer on campus that was not part
of a public computing site (e.g., from a science laboratory
or an office). Follow-up analyses using a one-way
ANOVA, with access location (horne, campus computing
site, or other campus computer) as the independent
variable and RRS total scores as dependent variables,
found no significant differences across access locations
[F(2,125) = 1.34,p > .10].
DISCUSSION
In this study, students who completed a personality
questionnaire on the Web reported higher levels of selffocused
rumination than did students who completed the
same questionnaire in paper-and-pencil format. The internal
consistency of the Web version was comparable
with the paper-and-pencil versions. Although students in
the Web sampIe participated in the study from varying
locations, this difference did not affect their responses
in any systematic manner. Taken together, these findings
support the use of the RRS on the Web and provide
tentative evidence that the Web is a viable mechanism
for questionnaire-based data collection and/or subject
screemng.
The participants in the Web sampIe reported more selffocused
rumination than did the participants who filled
out paper-and-pencil questionnaires, although the Web
sampIe did not differ from the students in upper-Ievel
psychology courses. Although women reported higher
levels ofrumination than did men in the general sampie
(collapsed across group), women in the Web sampie also
reported more rumination than did women who filled out
paper-and-pencil questionnaires. This finding might reflect
increased self-disclosure among women in the context
of a Web-based questionnaire. Some women may
feel more open to disclose tendencies such as ruminative
coping in a context that is arguably less evaluative than
a typical data collection context. By allowing people to
have more control over their environment when they disclose
information about sensitive issues, such as how one
responds to depressed mood, Web-based questionnaires
may encourage increased frankness ofresponse and selfdisclosure
(Hewson et al., 1996). A further possibility is
that some people may gain greater access to their feelings
and personality when cornpleting a Web-based questionnaire,
as compared with a questionnaire in a large lecture
class. Given that women are more likely than men to
ruminate when depressed, this tendency may be facilitated
in the context of a Web-based questionnaire.
The internal consistency ofthe RRS was similar across
the four groups. This finding indicates that these questionnaires
consistently measure a single construct and
that results from the Web version are comparable with
those obtained in the other sampies. The Web versions of
the RRS also were unaffected by differences in the location
from which the participants accessed the Web. This
finding suggests that the data people provide to a Web
page do not differ significantly as a function ofwhether
576 DAVIS
they choose to fill out questionnaires at horne, a public
computing site, or another computer on campus. Although
permitting access from multiple locations decreases experimental
control, it does not appear to affect questionnaire
results adversely.
A highly publicized study recently suggested that Internet
usage was associated with increases in depression
and loneliness and with declines in family communication
and the size of participants' social circles (Kraut et al.,
1998).This finding raises the question ofwhether the present
results reflect a sampling artifact, rather than differences
that are a function ofthe assessment medium. AIthough
such a possibility cannot be mied out definitively,
it is unlikely to have occurred for several reasons. First,
the present study sampled from college students with a
working knowledge of Web browsing, rather than from
first-time Internet users. Second, the findings of Kraut
et al. demonstrate only small relationships at best (e.g.,
the largest correlation between Internet usage and adjustment
measures is .19). Third, although Kraut et al.'s
data were interpreted as showing that participants used
the Internet as their social and emotional functioning declined,
Internet usage may haveprovided users with a newfound
hobby or a constructive activity to engage in given
that social and emotional functioning had declined, rather
than having caused the social and emotional difficulties.
Thus, there remains no conclusive or causal evidence that
Internet users have more social and/or emotional problems
than non-Internet users, nor has it been established
definitively that Internet use causes social and/or emotional
problems. The lack of such evidence and the dissimilarity
ofthe sampies used in this study and in that of
Kraut et al. suggest that the present results are not simply
a sampling artifact.
It might be argued that the Web sampie reported spuriously
higher levels ofself-focused rumination than did
the other groups owing to computer anxiety (cf. Tseng,
Tiplady, MacLeod, & Wright, 1998). This suggestion is
unlikely for three reasons. First, the participants in the
Websampie volunteered to participate in this study, which
required filling out a questionnaire on line. It is not likely
that computer-anxious people would volunteer for such
a study. Second, previous research has found that greater
levels of education are associated with lower computer
anxiety (Igbaria & Parasuraman, 1989). All the participants
in the present study were college students attending
a highly competitive university. Third, the RRS is not
a symptom questionnaire and has no apriori relationship
to computer anxiety. Rather, it measures the relatively
stable tendency to respond to depressed mood with passive,
emotion-focused thoughts and behaviors.
Alternatively,it could be argued that differences among
the groups other than the assessment medium might account
partly or fully for the observed results. Specifically,
the Web sampie replied to a general advertisement,
whereas the other groups were solicited at a lecture. Thus,
the Websampie may have been more motivated and more
interested in participating. It should be noted, however, that
flyers advertising the study were strategically placed near
computer laboratories and other computer facilities to
minimize inconvenience to potential participants. Many
people were observed simply writing down the URL as
they went to check their e-mail or to perform other routine
computing tasks. Thus, both the Web sampie and the
paper-and-pencil sampies were solicited at comparable,
although not identical, moments of convenience.
This study provides some initial data that tend to support
the use ofthe Web for questionnaire-based research.
However, more validation of Web-basedresearch is needed
(Krantz et al., 1997). The results ofthe present study increase
confidence in the use of the Web as a screening
device and as a method for collecting questionnaire data.
However, the participants were sampled from a college
campus and were restricted in terms of age range. More
data, using multiple measures and a broader sampie, would
be desirable in future studies. In addition, experimental
studies ofthe mechanisms underlying participants' higher
scores in the Web group would be useful. Identification
of such mechanisms and the continued validation of the
Web as a research tool appear to be growing areas of research
(Schmidt, 1997; Smith & Leigh, 1997).
REFERENCES
BOOTH-KEWLEY, S., EDWARDS, J. E., & ROSEN FELD, P. (1992). Impression
management, social desirability, and computer administration of
attitude questionnaires: Does the computer make a difference? Journal
ofApplied Psychology, 77, 562-566.
BUCHANAN, T., & SMITH,J. L. (1999). Using the Internet for psychological
research: Persona1ity testing on the World-Wide Web. British
JournalofPsychology,90,125-144.
BURKE, M. J. (1992). Computerized psychological testing: Impacts on
measuring predictor constructs and future job behavior. In N. Schmitt
& W.C. Borman (Eds.), Personnel selection in organizations (pp, 203-
238). San Francisco: Jossey-Bass.
BUTLER, L. D., & NOLEN-HoEKSEMA, S. (1994). Gender differences in
responses to a depressed mood in a college sampIe. Sex Roles, 30,
331-346.
DAVIS, R. N., & NOLEN-HoEKSEMA, S. (in press). Cognitive inflexibility
among ruminators and nonruminators. Cognitive Therapy & Re-
search.
DISSER, D. (1996). Htmail: A genericforms-to-email gateway [Computer
program]. Ann Arbor: University of Michigan Information Techno1ogy
Division.
ESPOSITO, J.• AGARD, E., & ROSNOW, R. (1984). Can confidentiality of
data pay off? Personality & Individual Differences, 5, 744-745.
EVAN, W. M., & MILLER,J. R. (I 969). Differential effects on response
bias of computer vs. conventional administration of a social science
questionnaire: An exploratory methodological experiment. Behavioral
Science, 14,2[6-227.
FINEGAN, J. E., & ALLEN,N. J. (1994). Computerized and written questionnaires:
Are they equivalent? Computers in Human Behavior, 10,
483-496.
GANGESTAD, S. W., & SNYDER, M. (1985). "To carve nature at itsjoints":
On the existence ofdiscrete classes in personality. Psychological Review,
92, 317-340.
GRAPHICS, VlSUALIZATION, AND USABILITY CENTER(1998). G VU:~ 9th
WWW survey results [On-line]. Atlanta: Georgia Institute of Technology,
College of Computing. Available URL: http://www.cc.gat-
ech.edu/gvu/user_surveys/
HEWSON, C. M., LAURENT, D., & VOGEL, C. M. (1996). Proper method-
WEB-BASED ADMINISTRATION OF A PERSONALITY QUESTIONNAIRE 577
ologies for psychological and sociological studies conducted via the
Internet. Behavior Research Methods, Instruments. & Computers, 28,
186-191.
HILE, M. G., & ADKINS, R. E. (1997). 00 substance abuse and mental
healthc1ients prefer automatedassessments?Behavior Research Methods.Tnstruments,
& Computers, 29,146-150.
IGBARIA, M..& PARASURAMAN. S. (1989). A path analytic study ofindividual
characteristics, computer anxiety, and attitudes toward microcomputers.
Journal ofManagement, 15,373-388.
KIELEY, J. M. (1996). CGI scripts: Gatewaysto World-Wide Webpower.
Behavior Research Methods.Instruments. & Computers, 28,165-169.
KING, W.C., & MILES, E. W. (1995). A quasi-experimental assessment
ofthe effectofcomputerizing noncognitivepaper-and-pencil measurements:
A test of measurement equivalence. Journal 01'Applied Psvchology,
80, 643-651.
KRANTZ, 1. H., BALLARD, 1..& SCHER, J. (1997). Comparing the results
of laboratory and World-Wide Web sampies on the determinants of
female attractiveness. Behavior Research Methods, Instruments. &
Computers, 29, 264-269.
KRAUT, R., PATTERSON, M., LUNOMARK, V.. KIESLER, S., MUKOPHAOHYAY,
T., & SCHERLIS, W. (1998). Internet paradox: A social technologythat
reduees soeial involvementand psychological well-being?
American Psychologist, 53, 1017-1031.
LEVINE, S., ANClLL, R. J., & ROBERTS, A. P.(1989). Assessment ofsuicide
risk by computer-delivered self-rating questionnaire: Preliminary
findings. Acta Psychiatrica Scandinavica, 80, 216-220.
LUCAS, R. W.(1977). A study of patients' attitudes to computer interrogation.
International Journal ofMan-Machine Studies, 9, 69-86.
LUCAS, R. w., MULLIN, P.1..LUNA, C. B. X., & MclNROY, D.C. (1977).
Psychiatristsand a computer as interrogatars of patients with alcoholrelated
illnesses: A camparison. British Journal ofPsychiutry, 131,
160-167.
LYUBOMIRSKY, S.. & NOI.EN-HoEKSEMA, S. (1993). Self-perpetuating
properties of dysphoric rumination. Journal ofPersonalitv & Social
Psvchology, 65, 339-349.
LYUBOMIRSKY, S., & NOLEN-HoEKSEMA, S. (1995). Effects of selffocused
rumination on negative thinking and interpersonal problem
solving. Journal 01' Personulitv & Social Psvchology, 69, 176-190.
MESZAROS, A., ENGELSMANN, F., METERISSIAN, G.. & KUSALIC, M.
(1995).Computerized assessment of depression and suicidal ideation.
Journal ofNervous & Mental Diseuse, 183,487-488.
MICHALAK, E. E. (1998). The use ofthe Internet as a research taoI: The
nature and characteristics of seasonal affective disorder (SAD)
amangst a population ofusers.lntemcting with Computers, 9, 349-365.
MII.LSTEIN, S. G. (1987). Acceptability and reliability of sensitive information
collected via computer interview. Educational & Psvchological
Measurement, 47, 523-533.
NOLEN-HoEKSEMA, S. (1990).Sex differences in depression. Palo Alto,
CA: Stanford University Press.
NOI.EN-HoEKSEMA, S. (1991). Responses to depression and their effects
on the duration ofdepressive episodes. Journal ofAbnormal Psvchologv,
100,569-582.
NOLEN-HoEKSEMA, S.. & LARSON, J. (1998). A dynamic model ofthe
gender differences in depressive symptoms. Manuscript submitted for
publication.
NOLEN-HoEKSEMA. S.. & MORROW, J. (1991). A prospective study of
depression and posttraumatic stress symptoms after a natural disaster:
The 1989 Loma Prieta earthquake. Journal ofPersonality & Social
Psychology, 61, 115-121.
NOLEN-HoEKSEMA, S.. PARKER, L., & LARSON, J. (1994). Ruminative
coping with depressed mood following lass. Journal ofPersonality
& Social Psychology, 67, 92-104.
PASVEER, K. A., & ELLARD, J. H. (1998). The making of a personality
inventory: Help from the Www. Behavior Research Methods, Instruments.
& Computers, 30, 309-313.
PETRIE, K., & ASELL, W. (1994). Responses ofparasuicides to a computerized
interview. Computers in Human Behavior, 10,415-418.
REIPS, U. (1996, October). Experimenting in the World Wide Weh. Paper
presented at the Society for Computers in Psychology conference,
Chicago.
SCHMIOT, W. C. (1997). World-Wide Web survey research: Benefits,
potential problerns,and solutions.Behavior Research Methods, Instruments,
& Computers, 29, 274-279.
SCHMIOT, W.C; HOFFMAN, R..& MACDoNALD, J. (1997). Operate your
own World-Wide Web server. Behavior Research Methods, Instruments,
& Computers, 29, 189-193.
SCHWARTZ, J. A. J., & KOENIG, L. J. (1996). Response styles and negative
affect among adolescents. Cognitive Therapy & Research, 20,
13-36.
SKINNER, H. A..& ALLEN, B. A. (1983). Does the computer make a difference?
Computerized versus self-report assessment of alcohol,
drug, and tobacco use. Journal ofConsulting & Clinical Psychology,
51,267-275.
SMITH, M. A.. & LEIGH, B. (1997). Virtual subjects: Using the Internet
as an alternative source of subjects and research environment. Behavior
Research Methods, Instruments. & Computers, 29, 496-505.
STANTON, J.M. (1998). An empirical assessment ofdata collection using
the Internet. PersonnelPsychology, SI, 709-725.
THAYER, R. E., NEWMAN, J. R., & MCCLAIN, T. M. (1994). Selfregulation
ofmood: Strategies for changing a bad mood, raising energy,
and reducing tension. Journal 01'Personality & Social Psychology,
67, 910-925.
TSENG, H.. TIPLADY, B.. MACLEOD, H. A., & WRIGHT, P.(1998). Computer
anxiety: A comparison ofpen-based personal digital assistants,
conventional computer and paper assessment of mood and performance.
British Journal ofPsychology, 89, 599-610.
WILSON, F. R., GENCO, K. T.. & YAGER, G. G. (1986). Assessing the
equivalence of paper-and-pencil vs, computerized tests: Demonstration
ofa promising methodology. Computers in Human Behavior, I,
265-275.
(Manuscript received November 16, 1998;
revision aecepted far publieation April 27, 1999.)