Pasek
6/4/14, 4:46 PM
First Monday, Volume 14, Number 5 - 4 May 2009
A recent draft manuscript suggested that Facebook use might be related to lower academic achievement in
college and graduate school (Karpinski, 2009). The report quickly became a media sensation and was picked
up by hundreds of news outlets in a matter of days. However, the results were based on correlational data
in a draft manuscript that had not been published, or even considered for publication. This paper attempts
to replicate the results reported in the press release using three data sets: one with a large sample of
undergraduate students from the University of Illinois at Chicago, another with a nationally representative
cross sectional sample of American 14– to 22–year–olds, as well as a longitudinal panel of American youth
aged 14–23. In none of the samples do we find a robust negative relationship between Facebook use and
grades. Indeed, if anything, Facebook use is more common among individuals with higher grades. We also
examined how changes in academic performance in the nationally representative sample related to
Facebook use and found that Facebook users were no different from non–users.
Contents
Introduction
The FG study
Prior research
The current examination
Methods
Results
Discussion
Limitations and future studies
Conclusions
Introduction
The proliferation of social networking sites has exploded in recent years, with an especially significant
increase in membership levels as seen in the case of Facebook. In the nationally representative sample used
in this study, Facebook users accounted for 16 percent of 14– to 22–year–olds in 2006 and 40 percent
among that same population in 2008. Indeed, with over 200 million unique users, Facebook use may have a
sizable real–world impact (Zuckerberg, 2009). Hence, as use of the social networking site has expanded, a
variety of studies have attempted to understand the correlates and potential effects of using the medium.
Researchers examining Facebook use from a media effects tradition have focused either on the social
implications of the medium or on the potential risks that users of social networking sites may experience.
http://firstmonday.org/ojs/index.php/fm/rt/printerFriendly/2498/2181
Page 1 of 15
Pasek
6/4/14, 4:46 PM
For instance, a variety of studies have noted that the use of Facebook is positively related to social capital
(Ellison, et al., 2007; Pasek, et al., in press; Valenzuela, et al., 2008). On the other hand, some research
suggests that Facebook users underestimate the potential privacy risks of sharing information on the site
(Acquisiti and Gross, 2006; Dwyer, et al., 2007).
Mass expansions of new technologies, especially among young people, have been ripe topics for hysteria.
Indeed, a large portion of the early research in mass communications was prompted by fears that the
motion picture industry might be clouding children’s minds (Blumer and Hauser, 1933). Similar concerns
were voiced by Putnam (2000) regarding an apparent negative relationship between television use and
social capital.
As with claims of dangers from older media, a recent study (and corresponding press release) indicating
that Facebook use and collegiate grade point averages (GPA) were negatively correlated generated a great
deal of media hype (hereafter “FG”; Karpinski, 2009). A Google News search for “Facebook” and “grades”
identified over 500 references to the FG study over a three day span [1]. In the immediate course of only a
few days, an unpublished and inadequately reviewed study that emphasized a simple correlation became an
established fact when disseminated through the news media.
The current examination serves as an attempt to restrain media hype with regard to the purported negative
relationship based on more rigorous investigation of more representative samples than the FG study
offered. We first examine the original FG study as well as the claims made in the press release to show that
the results were exaggerated. This misinterpretation was due both to the reporting and framing of the
study, which presented a raw correlation in a small convenience sample as strongly suggestive evidence, as
well as to failures on the part of mass media sources that reported the story. We also address the
relationship between Facebook use and grades using three studies with datasets much more closely attuned
to the question at hand. In contrast to the FG study, we find no evidence that Facebook use is related to
diminished academic achievement.
The FG study
The FG study should be regarded as problematic for a variety of reasons. We briefly address issues with its
sampling and analysis strategies that lead us to question the external validity of the results presented. We
then compare these results to claims made in the researcher’s press release that imply a causal and
directional influence of Facebook use on academic performance.
Sampling issues. Data from “102 undergraduate and 117 graduate students at a large, public Midwestern
university” were used in the FG study [2]. The draft manuscript explains that data were collected through a
“convenience sample” [3], of volunteers from summer and autumn classes at the university. While the
specifics of the classes and their potential attendees remain obscure, data reported in the paper make it
clear that the individuals sampled were unrepresentative of any large, public Midwestern university. Indeed,
a sample with 117 graduate students, 96 juniors and seniors, and only six freshman and sophomores is
unrepresentative of any university population at all.
Only 15 undergraduate students in the FG study reported that they did not use Facebook, making
conclusions about this group dubious at best. Further, the study included data from a large number of
graduate students, for whom grade point averages are often irrelevant unless they fall below a certain
threshold [4]. A quick look at the sample’s demographics reveals that individuals in the sample varied based
on their status as graduate or undergraduate students, full– vs. part–time students, age (spanning more
than a decade), gender, ethnicity, field of study, hours of external employment, Internet experience, and the
like. The author showed that many of these variables were related to Facebook use. Nevertheless, none of
these interrelated factors were controlled for in assessing Facebook “effects.”
Analysis strategy. The FG study used only one control variable to assess the relationship between Facebook
use and grade point averages — status as a graduate student. Use of only a single covariate leaves the
model underspecified given the large number of different populations included in the sample. A separate
finding of the study — namely that business and STEM [5] majors were much more likely to use Facebook
than individuals in the Humanities and Social Sciences (one category) — raises particular doubts. Engineers
generally have lower grade point averages than individuals in the humanities and social sciences, ceteris
paribus (Betts and Morell, 1999). Yet the author ignores this possibility when considering the Facebook–GPA
relationship.
http://firstmonday.org/ojs/index.php/fm/rt/printerFriendly/2498/2181
Page 2 of 15
Pasek
6/4/14, 4:46 PM
Descriptions of findings. In the FG study, the authors explained that “the suggested negative consequences
of use can alert administrators to find ways to limit access [to Facebook] … resulting in better academic
performance.” [6] The author, however, was clearly aware of the limitations of these findings. Indeed, the
draft paper later notes that directional relationships are not discernable using correlational data.
The press release published by the Ohio State University similarly juxtaposed recognition of the study’s
limitations with broad–sweeping claims about the implications of the findings. At one point, Karpinski
emphasizes that “we can’t say that use of Facebook leads to lower grades” but is also quoted as saying that
“there’s a disconnect between students’ claim that Facebook use doesn’t impact their studies, and our
finding showing that they had lower grades” (Grabmeier, 2009). Further, the only alternative explanation
proposed in either the press release or the draft manuscript involved the suggestion that Facebook users
might “still find other ways to avoid studying.” Yet it is quite possible that Facebook use is common for
individuals in disciplines where lower grades are the norm, that both Facebook use and low academic
performance are caused by some other untested factor, or that the relationship is entirely a function of a
small unrepresentative sample of students at a single university.
To what effect? While the FG study noted the need for further research, media sources were quick to
sensationalize the preliminary findings. On Time magazine’s Business and Tech blog the headline read:
“What Facebook Users Share: Lower Grades” (Hamilton, 2009). Other coverage took it a step further by
claiming a causal relationship: the Miami CBS affiliate proclaimed: “Study Finds Facebook Usage May Yield
Lower Grades” (CBS4, 2009) and MyFox Dallas/FortWorth declared: “Study: Facebook Hurts Grades” (MyFox
DFW, 2009). The study has also been widely reported globally. The Australian’s piece entitled “Facebook
Fixation Harms Students” reported: “Now academic research has validated the nagging suspicions of many
such students that Facebook is having a detrimental effect on their university results” (Wilson, 2009).
Similar reporting is found in the U.K.’s the Daily Telegraph story headlined “Facebook Students
Underachieve in Exams” (Khan, 2009). Indeed it should come as no surprise that media outlets have
sensationalized this story and ignored the researcher’s minor caveats.
Prior research
As Karpinski (2009) herself notes, she is not the first to examine the relationship between Facebook use and
grade point averages. She references two studies that lead to the hypothesis that academic performance
and use of the site might be negatively correlated. First, she cites a Master’s thesis by Matthew Boogart
(2006). While the thesis offers some suggestive evidence, Boogart only examined the relationship between
time spent on Facebook and GPA, and similarly failed to utilize control variables despite a diverse collection
of students from four universities. The second study Karpinski cites by Kubey, Lavin, and Barrows (2001)
does not mention social networking sites at all. The suggestion in the draft FG paper that social networking
sites represent a type of exclusively recreational use actually runs counter to other literature in the field
(e.g., Shah, et al., 2001; Pasek, et al., in press). It should be noted that an additional paper by Kolek and
Saunders (2008) found that there was no correlation between Facebook use and GPA in a representative
sample of students from a public Northeast research university. The draft FG manuscript cited the Kolek and
Saunders piece, but did not note its findings regarding the lack of a relationship between Facebook use and
grades.
The current examination
Due to the potential harm from misreporting these results and the important nature of the question raised
by the draft FG manuscript, we attempt to discern whether or not a relationship indeed exists between
Facebook use and grade point averages. Our investigation utilizes a multipronged approach by asking the
question in three different contexts. First, we look at a representative [7] cross–sectional sample of first–
year students from the University of Illinois at Chicago (UIC). Second, we examine the relationship in a
nationally representative sample of 14– to 22–year–olds. Additionally, we examine changes in grade point
averages from 2007 to 2008 among a longitudinal panel of nationally representative American youth aged
14–23.
http://firstmonday.org/ojs/index.php/fm/rt/printerFriendly/2498/2181
Page 3 of 15
Pasek
6/4/14, 4:46 PM
Methods
Data
Data for the UIC study come from 1,060 first–year students at the University of Illinois at Chicago in 2007.
Paper questionnaires were handed to students taking the mandatory first–year writing class. Because the
questionnaire was conducted in conjunction with the program, 85 of the 87 writing classes participated in
the exercise. The overall response rate for the study was 82 percent of students in the course [8].
The cross–sectional data as well as the longitudinal panel data were collected as part of the annual National
Annenberg Survey of Youth (NASY) conducted by The Adolescent Risk Communication Institute (ARCI) [9].
NASY respondents are initially recruited through random–digit dialing (RDD) telephone methods.
Interviewers asked how many individuals between the ages of 14 and 22 resided in the household and
requested parental permission to survey individuals under age 18. In 2008, 1,250 interviews were
successfully conducted. From that sample, 835 were new participants representing the cross–sectional
dataset with a response rate of 45 percent (AAPOR Formula 3) [10]. The additional 415 respondents were
volunteers among 900 from the 2007 NASY who elected to be re–contacted as part of a panel. These panel
participants were recruited in 2007 using the same RDD method as in other years [11]. The NASY response
rate in 2007 was 50 percent (AAPOR Formula 3), with a 2008 panelist re–contact rate of 74 percent.
GPA
Grade point averages were asked of respondents in both samples. In the UIC study, GPA was asked on an
eight–point scale ranging from “Mostly F’s” to “Mostly A’s” with categories in between such as “A’s and B’s.”
These were recoded on a zero–to–one scale (with 1 as “Mostly A’s”) to allow for comparison with the other
datasets in this paper (Mean = .76, SD = .19). Six (or less than one percent) of the 1,060 respondents did
not provide GPA information. In the NASY studies, GPA was coded on a four–point scale from “D or less” (0)
to “A” (1). Cross–sectional results (M = .75, SD = .24) were akin to panel results (M = .74, SD = .23).
Because individuals who were not in school could not report a current GPA, these individuals were dropped
from the sample (N = 145 for the cross–sectional analysis; N = 95 for the panel analysis) [12]. Additionally,
some individuals who were either homeschooled or in school but refused or did not know their GPAs were
dropped from the analysis (cross–sectional N = 12, panel N = 2). The approximate grade point average on
the traditional one to four scale for these results (M’s ≈ .75) would be equivalent to a 3.5, or a mix of both
A’s and B’s.
Facebook use
For both studies, Facebook use was assessed using a dichotomous measure. In the UIC study, respondents
were asked the following: “Have you ever used the following online sites and services?” regarding their
experiences with a number of sites. They were given the following response options: “no, have never used
it,” “tried it once, but have not used it since,” “yes, have tried it in the past, but do not use it nowadays,”
“yes, currently use it sometimes,” and “yes, currently use it often.” We consider Facebook users those who
chose one of the last two options: “yes, currently use it sometimes” (15.9 percent) or “yes, currently use it
often” (62.8 percent) for a total of 78.8 percent of the sample.
In the NASY samples, respondents were asked whether or not they had access to the Internet and how
frequently they used social networking sites before they were asked about the specific use of Facebook.
Only participants who reported both having access to the Internet and using social networking sites more
frequently than “never” were asked about their Facebook habits. The Internet screening question was
posed, “Do you have access to the Internet at home or somewhere else?” To determine whether individuals
used social networking sites, respondents were asked if they “Use online social network sites (SNS) like
MySpace or Facebook most days, once or twice a week, less often, or never.” For those who reported ever
using SNSs, a follow–up question asked, “Which, if any, of the following social networking sites do you use?”
Facebook was among the list of SNSs offered [13]. Respondents who did not have access to the Internet (N
= 29 cross–sectional; N = 9 panel), who “never” used SNSs (N = 88 cross–sectional, N = 45 panel), or who
did not report using Facebook were considered non–users, while those reporting Facebook use were
considered users (46.2 percent cross–sectional, 48.8 percent panel) [14].
Control variables
For all studies, we controlled for age, gender, race/ethnicity, and socio–economic status (SES). The two
http://firstmonday.org/ojs/index.php/fm/rt/printerFriendly/2498/2181
Page 4 of 15
Pasek
6/4/14, 4:46 PM
studies used different SES measures. In the UIC study, we used self–reported parental education levels
where we include the highest level of education for either parent. For NASY, individuals were assigned the
median household income for the zip code [15] in which they lived. In the case where individuals did not
provide a home zip code or an income match was not possible (N = 38), an imputed median income
(US$36,146) was inserted. Table 1 shows means for all variables, full question wordings are included in
Appendix 1.
Table 1: Means for Facebook users and non–users for all
variables.
Notes: Percentages by column where reported, because of rounding
numbers may not add to 100. 1: NASY samples only report
information for individuals who were in school at the time of the
survey. 2: Asian individuals included in other category for NASY
samples, as are responses for any non–white, non–black, non–
Hispanic group. 3: GPA is used as a 0–1 variable in the text.
University of
Illinois at
Chicago
2008 NASY
Cross–Section1
Facebook Non– Facebook Non–
user
user user
user
2007–2008
NASY Panel1
Facebook Non–
user
user
N
831
224
319
381
156
164
Age
18.4
18.5
17.9
16.2
18.2
16.2
Female
(%)
56.3
53.6
47.0
43.4
54.5
55.5
White,
non–
Hispanic
(%)
44.9
33.8
77.7
62.3
74.4
54.9
Black,
non–
Hispanic
(%)
7.9
7.3
6.6
9.7
6.4
18.3
Asian,
non–
Hispanic
(%)
31.6
22.4
8.2
7.3
10.9
6.1
6.6
20.5
7.7
20.7
Other,
non–
Hispanic
(%)2
Hispanic
(%)
15.4
34.8
Parental
Educ
(Less
than HS)
(%)
6.0
12.5
Parental
Educ
(High
School)
(%)
17.6
23.7
Parental
http://firstmonday.org/ojs/index.php/fm/rt/printerFriendly/2498/2181
Page 5 of 15
Pasek
6/4/14, 4:46 PM
Educ
(Some
College)
(%)
18.8
25.0
Parental
Educ
(College
Grad)
(%)
37.4
23.2
Parental
Educ
(Post–
College)
(%)
20.0
15.6
Income
(US$,
Based on
zip code)
52,412
44,162
53,399
43,673
Education
— High
School or
Less (%)
47.0
83.6
44.2
85.4
Education
— In
College
or More
(%)
53.0
15.9
55.8
14.6
3.34
3.17
3.34
3.13
3.42
3.11
GPA
(1–4)3
3.29
3.27
Lagged
GPA
(1–4)3
Results
None of the three studies detect a robust negative relationship between grade point averages and use of
Facebook. Instead, the three studies find a mixed bag of results, which reiterates the fact that corollary
relationships should be considered in their environmental and methodological context rather than used to
generalize trends (See Table 2). For example, in the UIC sample, even without controls there is no
statistically significant relationship between the two (Pearson’s r = .010, ρ = .746). However, the NASY
studies do indeed have statistically significant relationships when controls are absent, albeit in opposite
directions. The cross–sectional data in this nationally representative sample report a positive correlation
without controls (Pearson’s r = .122, ρ = .002). In contrast, we find that 2008 Facebook use is negatively
related to changes in grades from 2007 to 2008 within the NASY panel (Pearson’s r = -.148, ρ = .010).
Table 2: Correlations between Facebook use and GPA.
Note: 1: Correlation for the NASY Panel is between 2008
Facebook use
http://firstmonday.org/ojs/index.php/fm/rt/printerFriendly/2498/2181
Page 6 of 15
Pasek
6/4/14, 4:46 PM
and change in grade point average from 2007–2008. ** p <
.01.
Pearson’s Correlation
Study
Coef.
sig.
N
UIC study
.010
.746
1049
NASY Cross–Section
.122**
.002
660
NASY Panel1
-.148**
.010
303
These initial results highlight the precarious situation for researchers trying to decipher correlations. At first
glance, these unsophisticated outcomes offer mixed conclusions in regards to the relationship between
Facebook use and academic performance. Nonetheless, we wanted to ensure that these results were not
spurious. For instance, we could imagine that the preponderance of Facebook use among college students
might lead college–bound high school students to adopt the site at a higher rate than others. This might
lead us to conclude mistakenly that the relationship is more positive than would actually be the case. By
controlling for demographics, we partially mitigate this possibility, among other potential confounds.
With the control variables inserted, however, the results are nearly identical across our data sets. In both
the UIC study and the NASY panel, Facebook use is completely unrelated to students’ grades (Beta = -.003
and -.004, s.e. = .015 and .025, respectively; Table 3). In the NASY cross–section, the relationship remains
significantly positive after controls (Beta = .051, s.e. = .021), but represents a distinction of only one–fifth
of a letter grade and is no longer significant if the sample is limited to college students (Beta = .022, s.e. =
.036). All relationships remain consistent when we control for hours of Internet use (not shown).
Table 3: Regressions predicting grade point averages.
* p < .05 | ** p < .01 | *** p < .001.
NASY Cross– NASY
UIC study
Section1
Panel1
Beta
s.e Beta
s.e Beta
s.e
Intercept
.492*** .140 .734*** .025 .306*** .039
Age
.014 .007
.012
.033 -.051 .045
Female
.033** .012 .074*** .018
.006 .021
Black, non–
-.114*** .024 -.166*** .028 -.099*** .029
Hispanic
Asian, non–
.002 .014
Hispanic
Other, non–
-.065
.037 -.083* .042
Hispanic
Hispanic
-.032 .018 -.036
.025 -.019 .031
Parental Educ (Less
-.002 .027
than HS)
Parental Educ
-.012 .019
(High School)
Parental Educ
.026 .017
(College Grad)
Parental Educ
.034 .019
(Post–College)
Income by zip code
-.066
.073
.043 .089
Education level
-.018
.035
.016 .045
Lagged GPA
.608*** .044
Facebook
-.003
N
http://firstmonday.org/ojs/index.php/fm/rt/printerFriendly/2498/2181
1022
.015
.051*
658
.021
-.004
.025
303
Page 7 of 15
Pasek
6/4/14, 4:46 PM
R–Squared
.044
.089
.456
Discussion
This paper is intended both to get at the heart of the Facebook–GPA connection and to set the record
straight. As researchers, we have long known the importance of replication and peer review. Without these
safeguards, an intriguing preliminary finding can enter the popular discourse as if it represents established
fact. Indeed, it is often difficult to convey to those outside our fields the proper heft to bestow upon a
particular study. The FG study did not suggest or attempt to provide a definitive understanding of the
Facebook–GPA relationship. Yet easily sensationalized results and a widely distributed press release
positioned the findings on a path bound to spiral out of control.
The results presented in this response paper suggest that there is no negative relationship between
Facebook use and academic performance. Two of our analyses suggest that Facebook users were no more or
less likely to get good grades than non–users. The third study found evidence that Facebook use was
slightly more common among individuals with higher grades. Indeed, our findings are in direct contradiction
to those presented in the original FG study as well as the flurry of sensational media that ensued.
A few distinctions between the NASY cross–section and UIC studies may account for the slight difference in
results. Because Facebook use began in relatively privileged environments (first at Harvard University, then
at select U.S. colleges; boyd and Ellison, 2007), individuals more likely to enter those elite environments
may have been prone to both higher grades and Facebook use. By accounting for prior grades in the NASY
panel, we show that Facebook users neither seem to be deteriorating nor improving relative to non–users.
Nonetheless, it is still possible that Facebook use could be having an effect on these individuals. Of
particular note, since we do not have a reporting category higher than an “A” for GPA, GPA could be
experiencing a ceiling effect whereby individuals who had an “A” in 2007 had no possibility for improvement.
Hence, with higher initial grades, Facebook users had less room to improve.
Limitations and future studies
As with all research on new and evolving media, the changing nature of Facebook use may itself lead to
changing media effects. We should not be content to assume that this study — or any other for that matter
— provides a definitive answer on the implications of a medium. This is especially true given that Facebook
only emerged in 2004 and that 79 percent of our UIC sample and 43 percent of our nationally
representative sample were using the site by 2007 and 2008 respectively [16]. Indeed, in another few years
it is hard to imagine what could happen in our constantly evolving media environment.
We also do not intend this study to suggest that Facebook use, writ large, cannot exhibit a negative
relationship with academic performance. Individuals spending more than 30 hours a week on the site will
likely suffer from some sort of extreme time replacement effect (Pasek, et al., 2006). In that vein, we do
not suggest that Facebook use is some unmitigated good. As with most engaging hobbies and community
activities, Facebook use can be an effective means of participating in society as well as a means of
withdrawing from it. The question is not whether individuals are using a particular medium, but how. While
common media uses can induce broad effects, this did not seem to be the case with regard to Facebook and
academic disengagement.
Conclusions
http://firstmonday.org/ojs/index.php/fm/rt/printerFriendly/2498/2181
Page 8 of 15
Pasek
6/4/14, 4:46 PM
In this study we examined the relationship between Facebook use and academic achievement. In contrast to
recent sensational and unsubstantiated news reports that Facebook use lowers academic performance,
results from three studies indicate that the two variables are likely unrelated. We found no relationship in a
representative study of first–year undergraduate students at the University of Illinois at Chicago and a slight
positive relationship in a nationally representative survey of youth. Further, we used national longitudinal
data to assess changes in academic achievement from 2007 to 2008 in a nationally representative panel
study of young people. Changes in academic achievement did not vary with Facebook use when
demographic controls were considered. Facebook simply does not seem to have a generalizable impact on
grades.
About the authors
Josh Pasek (joshpasek.com) is a Ph.D. candidate studying political communication at Stanford University.
Josh’s papers have been published in Communication Research, Political Communication, the American
Journal of Education, and the Journal of Applied Developmental Science and have been presented at
meetings for the American Political Science Association, the International Communication Association, the
American Association for Public Opinion Research, and the Association of Consumer Research among other
locales. He recently finished serving as the Assistant Editor for Political Communication and is co–director of
the Methods of Analysis Program in the Social Sciences at Stanford University. His research interests include
political socialization, civic education, the role of media as a democratic institution, survey design, public
opinion, and civic engagement.
eian more is Senior Research Coordinator for the Adolescent Risk Communication Institute within The
Annenberg Public Policy Center of the Annenberg School for Communication at the University of
Pennsylvania. He received both his BA and MA from the University of Pennsylvania. His previous publications
focus on youth behaviors related to new media, online civic engagement, health behavioral trends depicted
in media across time, and risky behaviors among adolescents. Eian’s research interests include survey
methodology, the relationship between youth behavior and their media consumption, globalization, and
identity performance in virtual communities.
Eszter Hargittai (eszter.com) is Associate Professor of Communication Studies and Faculty Associate of the
Institute for Policy Research at Northwestern University where she heads the Web Use Project. In 2008/09,
she is also Fellow at the Berkman Center for Internet & Society at Harvard University. Her research focuses
on the social and policy implications of information technologies with a particular interest in how IT may
contribute to or alleviate social inequalities. Her research projects have looked at differences in people’s
Web–use skills, the evolution of search engines and the organization and presentation of online content,
political uses of information technologies, and how IT are influencing the types of cultural products people
consume. Her papers are available at www.webuse.org/papers.
Acknowledgements
Hargittai’s study at the University of Illinois at Chicago was made possible by a generous grant from the
John D. and Catherine T. MacArthur Foundation. Hargittai expresses her thanks to Ann Feldman and Tom
Moss of the UIC First–Year Writing Program for supporting this project. She is also grateful to Waleeta
Canon, Gina Walejko and the 2006–07 group of research assistants in the Web Use Project group for their
assistance with data collection and data entry.
The Adolescent Risk Communication Institute (ARCI) was created by the Annenberg Foundation in 2002 as
part of the Annenberg Public Policy Center at the University of Pennsylvania. ARCI aims to inform
researchers, policymakers, and the public regarding strategies to prevent risks to healthy adolescent
development and to enhance the well–being of youth. It conducts the annual National Annenberg Survey of
Youth, the Annenberg Media Health Coding Project and reviews of research by panels of experts. ARCI
would like to thank Shulman, Ronca, and Bucuvalas, Inc. for their help in collecting responses and preparing
the data set.
http://firstmonday.org/ojs/index.php/fm/rt/printerFriendly/2498/2181
Page 9 of 15
Pasek
6/4/14, 4:46 PM
Notes
1. This search on Google News covered 12 to 15 April 2009.
2. Karpinski, 2009, p. 7. While the conference presentation and press release refer to Adam Duberstein as a
second author, only Karpinski’s information was included on the draft paper circulated to the news media.
For this reason, we cite the paper as single–authored throughout the text.
3. Karpinski, 2009, p. 7.
4. In one study, attrition from Ph.D. programs was slightly more likely for individuals with higher GPAs
(Lovitts and Nelson, 2000).
5. The authors define STEM as statistics, technology, engineering, math, and medical majors.
6. Karpinski, 2009, p. 12.
7. Throughout this paper, we use the term “representative” to refer to data derived from probability
samples.
8. Further information about the sample can be found in Hargittai, 2007.
9. ARCI is a research department under the Annenberg Public Policy Center of the Annenberg School for
Communication within the University of Pennsylvania. Details about NASY can be found online at
http://www.annenbergpublicpolicycenter.org/ProjectDetails.aspx?myId=10.
10. NASY response rates are comparable to those obtained by the CDC in its national telephone surveys of
behavioral risk factors in adults.
11. 2007 NASY respondents were randomly assigned the opportunity to take the survey online on the
condition that they had access to the Internet, if they opted to take it online but failed to do so, a phone
interview was conducted instead. This opportunity was not granted to new 2008 participants. The 415
panelists were contacted via e–mail in 2008 — or by letter if no e–mail address had been obtained — and
given the opportunity to take the survey online once again. A phone interview was conducted for those who
did not complete the Web survey.
12. Only respondents enrolled during both years were kept in the panel sample.
13. The full list included MySpace, Facebook, Friendster, LinkedIn, Flickr, and an “Other” option that allowed
them to provide their own response.
14. Descriptive statistics are unweighted and reported for individuals who were in school at the time of the
study.
15. From 2000 Census Summary File 3 (http://factfinder.census.gov/servlet/DownloadDatasetServlet?
_lang=en).
16. The lower rate of Facebook membership in the NASY data is likely a result of sample composition. NASY
recruits younger participants of whom a slight majority are in high school. Facebook membership, however,
is more common for college–aged youth. Among only college students in the NASY cross–sectional sample
(N=228), frequency of Facebook use was comparable to that of the UIC sample (74.1 percent).
References
A. Acquisiti and R. Gross, 2006. “Imagined communities: Awareness, information sharing, and privacy on
the Facebook,” Lecture Notes in Computer Science, number 4258, pp. 36–58; version at
http://privacy.cs.cmu.edu/dataprivacy/projects/facebook/facebook2.pdf.
J.R. Betts and D. Morell, 1999. “The determinants of undergraduate grade point average: The relative
importance of family background, high school resources, and peer group effects,” Journal of Human
Resources, volume 34, number 2, pp. 268–293.http://dx.doi.org/10.2307/146346
http://firstmonday.org/ojs/index.php/fm/rt/printerFriendly/2498/2181
Page 10 of 15
Pasek
6/4/14, 4:46 PM
H. Blumer and P.M. Hauser, 1933. Movies, delinquency, and crime. New York: Macmillan.
M.R.V. Boogart, 2006. “Uncovering the social impacts of Facebook on a college campus,” Master’s thesis,
Department of Counseling and Educational Psychology, Kansas State Uniyersity.
d. boyd and N.B. Ellison, 2007. “Social network sites: Definition, history, scholarship,” Journal of Computer–
Mediated Communication, volume 13, number 1, at http://jcmc.indiana.edu/vol13/issue1/boyd.ellison.html.
CBS4, 2009. “Study finds Facebook usage may yield lower grades,” CBS4.com, at
http://cbs4.com/local/facebook.college.grades.2.984408.html.
C. Dwyer, S.R. Hiltz, and K. Passerini, 2007. “Trust and privacy concern within social networking sites: A
comparison of Facebook and MySpace,” paper presented at the Thirteenth Americas Conference on
Information Systems, Keystone, Colo. (9–12 August); version at
http://csis.pace.edu/~dwyer/research/DwyerAMCIS2007.pdf.
N.B. Ellison, C. Steinfeld, and C. Lampe, 2007. “The benefits of Facebook ‘friends’: Social capital and college
students’ use of online social network sites,” Journal of Computer–Mediated Communication, volume 12,
number 4, at http://jcmc.indiana.edu/vol12/issue4/ellison.html.http://dx.doi.org/10.1111/j.10836101.2007.00367.x
J. Grabmeier, 2009. “Study finds link between Facebook use, lower grades in college,” Ohio State University,
at http://researchnews.osu.edu/archive/facebookusers.htm.
A. Hamilton, 2009. “What Facebook users share: Lower grades,” Time Magazine: Business and Tech (14
April), at http://www.time.com/time/business/article/0,8599,1891111,00.html.
E. Hargittai, 2007. “Whose space? Differences among users and non–users of social network sites,” Journal
of Computer–Mediated Communication, volume 13, number 1, at
http://jcmc.indiana.edu/vol13/issue1/hargittai.html.http://dx.doi.org/10.1111/j.1083-6101.2007.00396.x
A.C. Karpinski, 2009. “A description of Facebook use and academic performance among undergraduate and
graduate students,” paper presented at the Annual Meeting of the American Educational Research
Association, San Diego, Calif.
U. Khan, 2009. “Facebook students underachieve in exams,” Daily Telegraph, at
http://www.telegraph.co.uk/education/educationnews/5145243/Facebook-students-underachieve-inexams.html.
E.A. Kolek and D. Saunders, 2008. “Online disclosure: An empirical examiniation of undergraduate Facebook
profiles,” NAPSA Journal, volume 45, number 1, pp. 1–25.
R.W. Kubey, M.J. Lavin, and J.R. Barrows, 2001. “Internet use and collegiate academic performance
decrements: Early findings,” Journal of Communication, volume 51, number 2, pp. 366–
382.http://dx.doi.org/10.1111/j.1460-2466.2001.tb02885.x
B.E. Lovitts and C. Nelson, 2000. “The hidden crisis in graduate education: Attrition from Ph.D. programs,”
Academe Online (November–December), at
http://www.aaup.org/AAUP/pubsres/academe/2000/ND/Feat/lovi.htm.
MyFox DFW, 2009. “Study: Facebook hurts grades,” MyFox Dallas Fort Worth, at
http://www.myfoxdfw.com/dpp/news/tech/Study_Facebook_Hurts_Grades.
J. Pasek, E. More, and D. Romer, in press. “Realizing the social Internet? Offline social networking meets
online civic engagement,” Journal of Information Technology and Politics.
J. Pasek, K. Kenski, D. Romer, and K.H. Jamieson, 2006. “America’s youth and community engagement:
How use of mass media is related to civic activity and political awareness among 14 to 22 year olds,”
Communication Research, volume 33, number 3, pp. 115–
135.http://dx.doi.org/10.1177/0093650206287073
R.D. Putnam, 2000. Bowling alone: The collapse and revival of American community. New York: Simon &
Schuster.
H. Schuman, E. Walsh, C. Olson, and B. Etheridge, 1985. “Effort and reward: The assumption that college
grades are affected by quantity of study,” Social Forces, volume 63, number 4, pp. 945–
966.http://dx.doi.org/10.1093/sf/63.4.945
http://firstmonday.org/ojs/index.php/fm/rt/printerFriendly/2498/2181
Page 11 of 15
Pasek
6/4/14, 4:46 PM
D.V. Shah, N. Kwak, and R.L. Holbert, 2001. “‘Connecting’ and ‘disconnecting’ with civic life: Patterns of
Internet use and the production of social capital,” Political Communication, volume 18, pp. 141–
162.http://dx.doi.org/10.1080/105846001750322952
S. Valenzuela, N. Park, and K.F. Kee, 2008. “Lessons from Facebook: The effect of social network sites on
college students’ social capital,” paper presented at the 9th International Symposium on Online Journalism,
Austin, Tex. (4–5 April); version at http://online.journalism.utexas.edu/2008/papers/Valenzuela.pdf.
L. Wilson, 2009. “Facebook fixation harms student grades,” The Australian, at
http://www.theaustralian.news.com.au/story/0,24897,25325762-12332,00.html.
M. Zuckerberg, 2009. “200 million strong,” Facebook Blog, at http://www.blog.facebook.com/blog.php?
post=72353897130.
Appendix 1: Full question wordings.
Notes: Both datasets also recorded if a respondent offered “Don’t Know” or refused to
answer a question. These numbers were universally small and represented a total loss in N
of 38 in the UIC sample, 42 in the NASY cross–sectional study, and 17 in the panel.
1: Question was preceded by a filter question.
Variable
UIC
question
Age
In what
year were
you born?
“Just to
confirm, what
is your age?”
Are you:
“Is the
[‘person’] a
male or
female?”
Male, Female
“Are you
yourself of
Hispanic or
Latino origin
or descent,
such as
Mexican,
Puerto Rican,
or Cuban?”
Yes, No
“How would
you classify
your race? Do
you consider
yourself to be
White, Black,
Asian,
Alaskan/Pacific
Islander,
Native
American, or
some
combination of
races?”
White/Caucasian,
Black/African
American, Asian,
Alaskan/Pacific
Islander, Native
American, Other
Gender
Hispanic
origin
Race
Response options
Male, Female
Are you of
Hispanic
Yes, No
or Latino
origin?
What is
your race?
Check all
that apply.
White/Anglo/Caucasian/Middle
Eastern; Black/African
American; Asian, please
specify; American Indian or
Alaskan Native; Other, please
specify.
NASY
question
Response
options
Response coded
by interviewer
as: Grade 8 or
lower, High
school freshman,
High school
sophomore, High
school junior,
High school
http://firstmonday.org/ojs/index.php/fm/rt/printerFriendly/2498/2181
Page 12 of 15
Pasek
6/4/14, 4:46 PM
senior, First year
of technical or
vocational school
after high school,
Second or higher
year of technical
“What grade
or vocational
or level of
school are you school after high
school, First year
in?”1
of college,
Second year of
college, Third
year of college,
Fourth year of
college, First
year of graduate
or professional
school after
college, Second
or higher year of
graduate or
professional
school after
college
Education
Matched to zip
code using
Census 2000
decennial
long–form
data “What is
the postal zip
code of your
permanent
home
address?”
Income
Parental
education
Grades
What is
the
highest
level of
education
your
father
obtained?
[repeated
for
mother]
How
would you
describe
the grades
you
received
last
semester?
Less than high school degree,
High school degree, Some
college, College graduate (for
example: B.A., B.S., B.S.E),
Advanced graduate (for
example: master’s,
professional, Ph.D., M.D.,
Ed.D.)
Mostly A’s, A’s and B’s, Mostly
B’s, B’s and C’s, Mostly C’s,
C’s and D’s, Mostly D’s, Mostly
F’s
“What is your
approximate
letter grade
average in the
school you
currently
attend?” If a
A, B, C, D, Other
numerical
grade is given,
ask “Would
that be an A,
B, C, D, or
something
else?”1
Have you
ever used
the
http://firstmonday.org/ojs/index.php/fm/rt/printerFriendly/2498/2181
Page 13 of 15
Pasek
6/4/14, 4:46 PM
Facebook
use
Hours of
Internet
use
following
sites and
services?
For each
site or
service,
let us
know if
you have
never
used it,
used it
once but
have not
used it
since,
used to
use it but
no longer
do,
currently
use it
sometimes
or
currently
use it
often.
On an
average
weekday,
not
counting
e–mail,
chat and
phone
use, about
how many
hours do
you spend
visiting
Web sites?
[repeated
for
“Saturday
or
Sunday”]
No, have never used it; Tried
it once, but have not used it
since; Yes, have used it in the
past, but do not use it
nowadays; Yes, currently use
it sometimes; Yes, currently
use it often
“Which, if any,
of the
following
social network
sites do you
use?”1
MySpace,
Facebook,
Friendster, Flickr,
Other
None; More than zero, but
less than one hour per day; 1
hour; 2 hours; 3 hours; 4
hours; 5 hours; 6 or more
hours
“About how
many hours do
you spend
using the
Internet on a
typical
weekday?”1
Less than 1 hour,
1 to 2 hours, 3
to 5 hours, 5 to
8 hours, More
than 8 hours
Editorial history
Paper received 19 April 2009; accepted 23 April 2009.
Copyright © 2009, First Monday.
Copyright © 2009, Josh Pasek, eian more, and Eszter Hargittai.
Facebook and academic performance: Reconciling a media sensation with data
by Josh Pasek, eian more, and Eszter Hargittai
First Monday, Volume 14, Number 5 - 4 May 2009
http://firstmonday.org/ojs/index.php/fm/rt/printerFriendly/2498/2181
Page 14 of 15
Pasek
6/4/14, 4:46 PM
http://firstmonday.org/ojs/index.php/fm/rt/printerFriendly/2498/2181
http://firstmonday.org/ojs/index.php/fm/rt/printerFriendly/2498/2181
Page 15 of 15