Conference PaperPDF Available

Don't be a phish: Steps in user education


Abstract and Figures

Phishing, e-mails sent out by hackers to lure unsuspecting victims into giving up confidential information, has been the cause of countless security breaches and has experienced in the last year an increase in frequency and diversity. While regular phishing attacks are easily thwarted, designing the attack to include user context information could potentially increase the user's vulnerability. To prevent this, phishing education needs to be considered. In this paper we provide an overview of phishing education, focusing on context aware attacks and introduce a new strategy for educating users by combining phishing IQ tests and class discussions. The technique encompasses displaying both legitimate and fraudulent e-mails to users and having them identify the phishing attempts from the authentic e-mails. Proper implementation of this system helps teach users what to look for in e-mails, and how to protect their confidential information from being caught in the nets of phishers. The strategy was applied in Introduction to Computing courses as part of the computer security component. Class assessment indicates an increased level of awareness and better recognition of attacks.
Content may be subject to copyright.
Don’t be a Phish: Steps in User Education
Stefan A. Robila
Montclair State University
RI 301, Computer Science
Montclair, NJ 07043
James W. Ragucci
Montclair State University
RI 301, Computer Science
Montclair, NJ 07043
Phishing, e-mails sent out by hackers to lure unsuspecting victims
into giving up confidential information, has been the cause of
countless security breaches and has experienced in the last year an
increase in frequency and diversity. While regular phishing
attacks are easily thwarted, designing the attack to include user
context information could potentially increase the user’s
vulnerability. To prevent this, phishing education needs to be
considered. In this paper we provide an overview of phishing
education, focusing on context aware attacks and introduce a new
strategy for educating users by combining phishing IQ tests and
class discussions. The technique encompasses displaying both
legitimate and fraudulent e-mails to users and having them
identify the phishing attempts from the authentic e-mails. Proper
implementation of this system helps teach users what to look for
in e-mails, and how to protect their confidential information from
being caught in the nets of phishers. The strategy was applied in
Introduction to Computing courses as part of the computer
security component. Class assessment indicates an increased level
of awareness and better recognition of attacks.
Categories and Subject Descriptors
K.3.2 [Computer and Education]: Computer & Information
Science Education Computer Science Education, Curriculum.
C.2.0 [Computer-Communication Networks]: General -
Security & Protection, K.6.5 [Management of Computing &
Information Systems]: Security & Protection Education
General Terms
Reliability, Experimentation, Security, Human Factors
Phishing, information security, computer education
Evolution does not only apply to plants and animals, it also
applies to human technology. Armor, for instance, was designed
to protect the wearer from being wounded. Technology eventually
evolved to find ways of piercing or breaking that armor. In
response, various types of armor have been developed to protect
the wearers from the new methods of destroying the armor,
creating an endless cycle of improvement within technology.
When one side improved its defenses, the other side would
improve its offences. Computer security and security threats have
evolved in the same manner. Throughout this ongoing race, one
thing remained constant as the weakest link, the human factor [1].
Spam, defined as the unsolicited sending of commercial e-mail
advertisements [2], which combined with online fraud has caused
losses of over 200 million dollars in 2003 alone [3]. A defense
against the overabundance of spam is the implementation of spam
filters currently used within most of the organizations or free mail
systems. Social engineering, defined by Kevin Mitnick as “using
manipulation, influence and deception to get a person, a trusted
insider within an organization to comply with a request, and the
request is usually to release information or perform some sort of
action item that benefits that attacker”[1] transforms spam to the
next level in the evolutionary chain. A combination of fraudulent
spam e-mail with social engineering creates a relatively new tactic
called Phishing. Phishing is using social engineering to send spam
e-mail(s) to unsuspecting victims, known as phishes [4]. The e-
mails are disguised (“spoofed”) as coming from legitimate
corporations and aim at directing users to copies of legitimate
websites. The “phishers” goal is to “fish” for confidential
information that the phishes have access to, such as bank account
numbers, usernames, passwords and social security numbers [2].
Phishing attacks come in a large variety of flavors. Some attacks
may masquerade as security upgrades or information verification
from the Bank of America. In more recent times, some phishing
attacks have appeared claiming to be a charity organization
collecting money to benefit the victims of Hurricane Katrina or
the 2004 Indian Ocean Tsunami [5] as seen in Fig 1 [6]. Damage
from these attacks is immense. Between May 2004 through May
2005, approximately 1.2 million U.S. computer users suffered
losses of nearly one billion dollars [7]. The United States is not
the only target. In the United Kingdom, it was reported in March
2005 that damage jumped 20% to cause 504 million pounds in
damage [8]. In addition, phishing attacks were also reported in
non-English speaking countries as early as 2004 [9].
Although the frequency of phishing attacks has recently stabilized
[10], context aware phishing is of particular concern for the
future. A context aware attack consists of the phisher gaining
knowledge of what sites and services the phish uses and
customizing an attack that appears to be from the target’s service
[11]. While currently, a phishing attack success rate is under 1%
[1] a context aware attack would result in much higher rates [11].
Permission to make digital or hard copies of all or part of this work for
personal or classroom use is granted without fee provided that copies are
not made or distributed for profit or commercial advantage and that
copies bear this notice and the full citation on the first page. To copy
otherwise, or republish, to post on servers or to redistribute to lists,
requires prior specific permission and/or a fee.
ITiCSE’06, June Month 1–2, 2004, City, State, Country.
Copyright 2005 ACM 1-58113-000-0/00/0004…$5.00.
Faced with these prospects, one must see what tools are available
to fight phishing. On one hand, application development will
continue to improve spam and phishing detection. On the other
hand, human factor risks can be reduced by spam and phishing
education. In this paper we analyze various efforts in educating
students related to phishing and we discuss our experience with
teaching based on phishing IQ tests and context information. The
paper is organized as follows. In Section 2 we briefly look at
previous phishing education efforts. In Section 3 we discuss
context aware phishing and the design of context aware attacks
and tests. In Section 4 we discuss the pedagogical setting for
phishing education. Section 5 presents a summary of the IQ tests
as well as results of the educational assessment survey. We end
the paper with Conclusions and Future Work (Section 6) and
Anti-phishing action is supported by a wide group of interesting
parties including most of the financial institutions. Private and
government institutions have developed phishing awareness
websites including [12], [13], [14].
According to [3], there are five individual issues that have to be
addressed in order to combat phishing: education, preparation,
avoidance, intervention and treatment. Within these groupings,
education is given the least attention. The paper only states that
users need to be educated in how to recognize suspicious requests
in their e-mail. There is minimal instruction of how users should
learn how to identify phishing attacks. In fact, both [3] and [15]
briefly list what should be taught to users, but they do not list a
vehicle for this education. The websites listed above, while
relatively up to date, mainly provide a description of phishing
attacks and some good sense advice.
An alternative approach is described in [16]. Researchers at
Indiana University conducted a study of 1,700 students in which
they collected websites frequently visited by students and either
sent them phishing messages or spoofed their e-mail addresses.
The respondents were provided with a discussion forum and were
informed of their participation in the study. Nevertheless, while
the authors indicate that public awareness was of concern, the
only clear outcome was an increased effort on the university IT to
protect against phishing. The large success rate of the attacks
(over 50% in some experiments) was mainly due to the use of
social context information.
One tool that has been developed by [17] to educate users is a
phishing IQ test. The test provides users with a combination of
actual phishing attacks and legitimate e-mails and users are asked
to distinguish between them. Once finished the user is given a
score of how well he or she was able to identify the e-mail. As a
result of this IQ test, users’ awareness of phishing increased. In a
period of a year, the test has seen an average score increase of 14
points, with this year the average score being a 75% [18].
Although this increase is a good sign, the average score shows
that 25% of the e-mails are wrongly identified. Results show that
82% of the test takers identify phishing e-mails correctly, but
legitimate e-mails are only identified correctly worldwide 52% of
the time [18]. This alarmingly low number compared to the
correct number of detected phishing attempts could probably be
attributed to users classifying all of the e-mails on the test as phish
While benign in nature, the MailFrontier’s IQ tests have limited
usability in a classroom. Given its generality, it is likely that most
of the students will have never dealt with the companies listed in
the tests. As such, the corresponding messages would have end up
being classified as SPAM from the beginning, would have been
placed in a bulk folders, or deleted without being opened [15].
Alternatively, the Social Phishing experiments described in [16]
would have a stronger impact since they addressed entities
familiar to the users. However, a direct phishing attack targeting
the students, while interesting from the point of research, would
have limited educational impact.
We suggest a hybrid approach where we employ IQ tests enriched
by social contexts. Borrowing from the idea of context aware
phishing attacks, described in [11], if an IQ test is developed
using known services that a group of users have a high probability
of using, then the element of inexperience with the
company/service is eliminated. Instead, users would be familiar
with the source that they are being tested on and should be better
able to identify real e-mails from fraudulent phishing attempts.
Once this is accomplished, then a different set of reasons can be
assumed for a user to falsely identify an e-mail. Reasons for false
identification might include any of the following:
User is unaware of what phishing is.
User is aware of phishing, unaware how to search for it.
User does not suspect a phishing attempt from that
There are several approaches that would allow collection of social
context information. These include collection of browser data
(such as history collection, auto-fill properties, etc. - for some
excellent examples, visit [20]), exploit of the browser tabs [21],
access to user’s labs, and crawling social network websites [16].
Fig 1. Phishing Example
We pursued to design a customized phishing IQ test focused on
Montclair State University students. Picking companies and
online services that a large portion of the student body uses per
the specifications addressed above in section three required
different strategies of data collection. Most of the data was
collected from the computer laboratory where the test was
administered and the campus computer centers. All of the
computers examined in these locations underwent browser
analysis from tools found at [20] followed by analysis of the
browsers cache and local history to determine if people used the
machines to visit any additional websites of interest. Other
websites that were used for this project were included after
looking at the resources available to students on campus. For
instance, the ATM machines on the MSU campus are run by
BankX. Therefore, it is logical to assume that many, if not most of
the students who attend the university have an account at BankX
to avoid service charges when using the ATM machines. We also
note that while the use of social network websites is very
attractive, generating user personalized IQ tests was beyond the
scope of our work.
Once the data of what MSU students use was gathered, phishing
and legitimate e-mails were generated and mixed with others from
a variety of sources including [10] and [14] and then grouped in
an IQ survey that contained 12 questions, with six answers that
are legitimate and six that are phishing attacks. At the end, the
survey was formed with messages from: University registrar (P), (L), BankX (L), University Parking (L),
(P), BankY (L), (L), (P),
(P), (P), (P), (L)
where P and L indicate a phishing or a legitimate e-mail
respectively. Fig 2 shows a sample question related to To implement the survey and record the results we
used an in-house mechanism based on javascript and an Apache
Tomcat server running on a Sun Solaris box. A survey taker
would be presented with images of an email at a time and asked to
indicate if it is a phishing attempt or not. Upon selecting and
confirming an answer, the next e-mail image was displayed. At
the end, the application displayed the number of correct answers.
4.1 Steps
Fig. 3 describes the steps taken in discussing and testing phishing
knowledge. We started by presenting what phishing is and
discussing its implications. Following, we discussed the nature of
the phishing emails and we discussed the methods that can be
employed by attackers to generate better targeted messages as well
as identify many browser and web vulnerabilities. Next, we
discussed ways to increase public awareness on phishing and
debated on aggressive (such as the phishing emails used in [12])
versus passive (such as the IQ tests [20]) education.
Steps six and seven were used to assess the quality of the
instruction. First, an IQ survey was administered followed by a
class assessment survey. The session was timed at approximately
30 minutes. A control number was presented to the survey taker.
The number could be used to show that the student has taken the
survey; however, it could not be related to the answers, in order to
ensure anonymity.
4.2 Course Description
We have applied the phishing education module to two sections of
our Introduction to Computing course. Approved to satisfy the
General Education requirements for computer fluency in a liberal
arts college, the course includes one lecture and one lab session
every week for a semester. The lectures are dedicated to various
computing issues such as computer organization, structured
programming, networks, privacy and ethics, while the labs are
focused on acquiring skills on various productivity packages as
well as learning basic concepts in programming, web development
and netiquette. The course is attended by large numbers of non-
science majors and it is organized in sections of up to 24 students
(due to lab space limitations).
As per the arguments in [22], the Introduction to Computing
course was a natural choice for piloting the phishing education
module since the vast majority of students will have computing
experience mainly from the user’s point of view. In addition, one
component of the course is teaching students about basic
computer and network security issues, a broad topic that includes
phishing. Finally, the lab component of the course allowed us to
have the students take the surveys in a set timeframe and to
discuss the general results immediately.
Of the 48 IQ surveys completed, we recorded an average
correctness rate of approximately 6.87 corresponding to an
Fig 2. Phishing e-mail created from a confirmation e-mail
sent by to a dummy account
Fig. 3 Phishing Education Steps
1. Learn what phishing is [4]
2. Understand the implications of phishing [12]
3. Discuss the ways phishing can be detected [12], [13]
4. Discuss the ways phishing information can be collected.[20]
5. Discuss ways to evaluate phishing education [12], [20]
6. Fill Phishing IQ
7. Discuss general results and evaluate the session.
average IQ of 57.29%. This means that approximately one out of
two e-mails, the students would erroneously identify a legitimate
message as a phishing attack, or a phishing attack, as a legitimate
Next, we analyzed the success rate for the legitimate and phishing
e-mails respectively. In case of the legitimate e-mails, the students
we able to correctly identify on average 3.64 out of 6 (60%) while
the phishing e-mails were recognized on average 3.22 out of 6
(53%). Interestingly, the messages easiest recognized as legitimate
was one from (81% correct match rate) while the
lowest legitimate recognition was received by a message from
BankX (50% correct match rate). For phishing, the lowest correct
identification was achieved by a message purporting to be from
the university’s registrar office (35% correct match rate) while the
highest was built from scratch claiming to be from an instructor
and referring students to (75%). We note that the
question with the lowest overall correct match rate, the phishing
attempt based on the university’s registrar’s office, was requesting
the student to login and verify personal information. At the time
of the survey administration, the university was revealing that a
large number of student data was inadvertently disclosed publicly
and that identity theft risk for the students was increased. We
believe that the students were aware of the incident and that this
played a major role in their choice for an answer. If such a
phishing attack would have really occurred, a 65% success rate
would have been disastrous. Also note that a month after the
survey was administered, the university sent a legitimate version
of the e-mail to the student body.
A second survey was administered to students to evaluate the
educational value of the session. The surveys were anonymous
and no points were awarded based on its completion. 78% of the
respondents indicated that they were not aware of phishing prior
to the class. At the same time, 93% acknowledged receiving
possible phishing e-mails in the past and 28% revealed that they
probably answered to phishing attacks in the past.
Fig. 4 and 5 are examples of the questions intended to test
understanding of social context factors for phishing. We note that,
as previously reported, most of the students still consider
messages from a friend as legitimate, although a large minority
(40%) now has second thoughts. In addition, most of the
participating students have become aware of the vulnerabilities
existing in browser data (see Fig 5). The students have positively
perceived both the IQ surveys as well as the overall phishing
session, with 94% agreeing that the IQ survey was helpful (Fig 6)
and all agreeing that the session was helpful (Fig. 7).
Finally, class discussions have revealed that most of the students
place significant trust in the educational institutions as well as the
social network websites. As such, while many have easily
declared that they would ignore banking or auction/payment site
emails, they were surprised to discover that attacks could occur by
spoofing the university or social network website communication.
We believe that this is explainable by the current focus of the anti-
phishing and anti ID-theft campaigns that put emphasis on
protection of financial and health information and ignore popular
web destinations. Unfortunately, such destinations also handle
private information (such as names, addresses, date of birth)
which can be used for identity theft. Successful phishing attacks
based on them would be equally disastrous.
Fig 4. Receiving a message from a friend makes me less likely
to assume it is a phishing attempt.
Fig. 5 Information stored by the Internet browser can be used
to refine a phishing attack.
Fig. 6 The phishing IQ survey was helpful in understanding
the topic.
Fig. 7 Attending this session made me more informed about
phishing and less likely to fall prey to such an attack.
Phishing has become a significant problem for internet users.
While most of its effects are noticeable in the United States, it is
expected that phishing will continue to expand all over the world.
Recent reports such the ones produced by the Antipishing
Working Group [23] and the Korea Internet Security Center [24]
reflect these trends and predict a continuous increase in attacks
and diversity. While technology advances continue to fight the
problem, user education continues to constitute a significant
In this paper we have discussed an approach to user education that
involves quantitative testing and social context aware examples.
The strategy allowed us to include phishing topics in an
Introduction to Computing course aimed at students pursuing a
non-computer science education. The phishing IQ test and the
session evaluation survey reveal that the current student body is
mostly oblivious to phishing threats. Upon being exposed to the
topics and shown how to analyze a message for phishing
characteristics, students are able to correctly identify most of the
threats. The students have positively appreciated the session and
its format and have acknowledged its usefulness.
More work remains to be done. Given a predicted increase in
tools available to fight phishing, it is expected that future attacks
will continue to be more and more refined in user and event
specificity. Previous work together with our experiments show
that such attacks have an extremely high success rate since they
most likely appeal to the user’s emotions. Accordingly, phishing
education will continue to be improved by use of user specific
tools. A social aware IQ test, that would be personalized on each
user, could be such a tool.
Our work can be expanded for use in other courses as well for
general public training. Coupled with context information, one
can always design specific tests, targeted at specific user groups.
Overall, our approach to phishing education was shown to be an
attractive tool.
We would like to thank Dr. Carl Bredlau for his vital part in the
construction of the creation of the survey using jsp files, as well as
hosting the files on his server.
A sample of the phishing test together with other materials is
available at
[1] CNN. com, “A convicted hacker debunks some myths.”
k.cnna/index.html 2005, accessed 01/06/06
[2] Duntemann J., Degunking Your Email, Spam, And Viruses.
Scottsdale, Arizona: Paraglyph Press, 2004
[3] Merwe A, Loock M., and Dabrowski M.. “Characteristics
and responsibilities involved in a Phishing attack.” Proc.
ACM WISCT 05, 92, 249-254, 2005
[4], accessed 30 Nov 2005
[5] Roberts, Paul F. “Cyber-looters Capitalize on Katrina.”
eWeek. 12 Sept. 2005: 11-12
[6] MailFrontier Phishing IQ, “Paypal Tsunami” example,
gif, accessed 3 Nov. 2005.
[7] Kerstein P.L., “How Can We Stop Phishing and Pharming
accessed 27 Nov 2005
[8] Richardson T., “Brits Fall Prey to Phishing.” The Register.,
accessed 27 Nov 2005
[9] Sunday Morning Herald, “Phishing Spreads in Europe”,
.html, accessed 5 Jan 2006
[10] Anti-Phishing Working Group, October 2005 Report,
5.pdf, accessed 27 Nov 2005
[11] Jakobsson M., Modeling and Preventing Phishing Attacks.
Phishing Panel in Financial Cryptography ’05.
[12] Anti-Phishing Working Group,,
accessed 27 Nov 2005
[13] Better Business Bureau,,
accessed 4 Jan 2006
[14] Microsoft, Consumer Awareness Page on Phishing
spx, accessed 6 Jan 2006
[15] Emigh A., Online Identity Theft: Phishing Technology,
Chokepoints, and Countermeasures. Radix Labs. 3 Oct,
[16] Jagatic T., Johnson N., Jakobsson M., and Menczer F.,
“Social Phishing”, Communications of ACM, to appear,
experiment/phishing-preprint.pdf, accessed 3 Jan 2006
[17] Mail Frontier. Phishing IQ,,
accessed 3 Nov 2005
[18] Horgan D.,.“The Phishing Phleet”
he_phishing_ph.html, accessed 2 Dec 2005
[19] Brandt A., “Phishing Anxiety May Make You Miss
Messages” PCWORLD. October 2005: 34
[20] IU Phishing Research,,
accessed 6 Jan 2006
[21], “Browser Phishing Flaw Could Hook
accessed 15 Dec 2005
[22] Werner, Laurie. “Redefining Computer Literacy in the Age of
Ubiquitous Computing.” Proc. ACM SIGITE 05, 95-99, 2005
[23] Anti-Phishing Working Group, “Phishing Activity Trends
apwg_report_DEC2005_FINAL.pdf, accessed 20 March
[24] Korea Internet Security Center, “Korea Phishing Activity
Trends Report”,
200601_KoreaPhishingReport_Jan2006.pdf, accessed 20
March 2006
... This approach was partially experimented by [73] in the introduction to computing courses taken by students not pursuing a computer science education. The authors concluded in the class assessment that students had an increased level of awareness and were better able to recognise phishing scams. ...
Full-text available
According to the international Anti-Phishing Work Group (APWG), phishing activities have abruptly risen over the last few years, and users are becoming more susceptible to online and mobile fraud. Machine Learning techniques have potential for building technical anti-phishing models, with a handful already implemented in the real time environment. However, majority of them have yet to be applied in a real time environment and require domain experts to interpret the results. This gives conventional techniques a vital role as supportive tools for a wider audience, especially novice users. This paper reviews in-depth, common, phishing countermeasures including legislation, law enforcement, hands-on training, and education among others. A complete prevention layer based on the aforementioned approaches is suggested to increase awareness and report phishing to different stakeholders, including organizations, novice users, researchers, and computer security experts. Therefore, these stakeholders can understand the upsides and downsides of the current conventional approaches and the ways forward for improving them.
... Tseng et al. (2011) have also created a game to educate phishing users on the basis of web content. The effect of user education was investigated in Robila and Ragucci (2006) phishing and identifying good e-mails. Tseng et al. (2011) developed a game to alert users about phishing using the contents of the website. ...
E-Mails are commonly used as a medium of communication for personal and pro- fessional purposes. Information shared via mail is also sensitive and private, such as financial information, credit records, login data, and so on. This makes them useful to cyber attackers who can exploit this knowledge for illegal purposes. Phishing is a technique used by fraudsters to gain confidential data from users by alleging to be from known sources. In a phished e-mail, the sender can encourage users to provide private in- formation under false premises. Phishing e-mails have also been frequently used in financial institutions and consumer fraud. This paper discusses an overview of dif- ferent techniques for phishing e-mail detection and protection currently used in e-mail filtering. A comparative analysis and review of these techniques are carried out. This survey provides an understanding of the phishing detection issue, its present space for a solution, and its future directions for study.
... Users must have their own understanding of spam in order to combat the issue. Robila and Ragucci (2006) focussed on spam and phishing education as they taught students about illegitimate email characteristics. Their research indicates teaching undergraduate students to recognise illegitimate emails improves the ability to correctly identify spam and phishing in an ecologically valid setting. ...
This study explored distinct perceptual and decisional contributions to spam email mental construal. Participants classified spam emails according to pairings of three stimulus features – presence or absence of awkward prose, abnormal message structure, and implausible premise. We examined dimensional interactions within general recognition theory (GRT; a multidimensional extension of signal detection theory). Classification accuracy was highest for categories containing either two non-normal dimension levels (e.g., awkward prose and implausible premise) or two normal dimension levels (e.g., normal prose and plausible premise). Modelling indicated both perceptual and decisional contributions to classification responding. In most cases perceptual discriminability was higher along one dimension when stimuli contained a non-normal level of the paired dimension (e.g., prose discriminability was higher with abnormal structure). Similarly, decision criteria along one dimension were biased in favour of the non-normal response when stimuli contained a non-normal level of the paired dimension. Potential applications for training are discussed. Practitioner Summary: We applied general recognition theory (i.e., multivariate signal detection theory) to spam email classification at low or high levels of three stimulus dimensions: premise plausibility, prose quality, and email structure. Relevant to training, this approach helped identify perceptual and decisional biases that could be leveraged to individualise training.
... Educating people about phishing and online safety can be done in different ways, for example using cartoons [27]. Robila & Ragucci [21] tested user education consisting of quantitative testing and social context aware examples. The students that followed this education showed better phishing identification skills and gave a positive evaluation of the education they received. ...
Doorstep scams are scams, often happening at the front door, in which a con artist has a convincing, but fraudulent, story with the purpose of coming into your house and/or stealing money. Various campaigns to educate people exist, but they do not focus on the verbal skills people can use to prevent themselves from becoming a victim. This paper describes the conceptual design of a proposed training application. This application will provide an agent-based learning environment for high-risk doorstep scam victims. In order to create a training application, field research has been done to the content and progress of doorstep scams, which is used to create interactive scenarios.
... The concept of testing the ability to detect phishing in an educational setting is challenging [32]. Getting the attention of children aged 8-13 to focus on cybersecurity is no less of a challenge. ...
Conference Paper
Full-text available
User training is a commonly used method for preventing victimization from phishing attacks. In this study, we focus on training children, since they are active online but often overlooked in interventions. We present an experiment in which children at Dutch primary schools received an antiphishing training. The subjects were subsequently tested for their ability to distinguish phishing from non-phishing. A control group was used to control for external effects. Furthermore, the subjects received a re-test after several weeks to measure how well the children retained the training. The training improved the children’s overall score by 14%. The improvement was mostly caused by an increased score on the questions where they had to detect phishing. The score on recognizing legitimate emails was not affected by the training. We found that the improved phishing score returned to pre-training levels after four weeks. Conversely, the score of recognition of legitimate emails increased over time. After four weeks, trained pupils scored significantly better in recognizing legitimate emails than their untrained counterparts. Age had a positive effect on the score (i.e., older children scored higher than younger ones); but sex had no significant influence. In conclusion, educating children to improve their ability to detect phishing works in the short term only. However, children go to school regularly, making it easier to educate them than adults. An increased focus on the cybersecurity of children is essential to improve overall cybersecurity in the future.
Conference Paper
Abstract—Phishing attacks are prevalent and humans are central to this online identity theft attack, which aims to steal victims’ sensitive and personal information such as username, password, and online banking details. There are many antiphishing tools developed to thwart against phishing attacks. Since humans are the weakest link in phishing, it is important to educate them to detect and avoid phishing attacks. One can argue self-efficacy is one of the most important determinants of individual’s motivation in phishing threat avoidance behaviour, which has co-relation with knowledge. The proposed research endeavours on the user’s self-efficacy in order to enhance the individual’s phishing threat avoidance behaviour through their motivation. Using social cognitive theory, we explored that various knowledge attributes such as observational (vicarious) knowledge, heuristic knowledge and structural knowledge contributes immensely towards the individual’s self-efficacy to enhance phishing threat prevention behaviour. A theoretical framework is then developed depicting the mechanism that links knowledge attributes, self-efficacy, threat avoidance motivation that leads to users’ threat avoidance behaviour. Finally, a gaming prototype is designed incorporating the knowledge elements identified in this research that aimed to enhance individual’s self-efficacy in phishing threat avoidance behaviour
Full-text available
Phishing is one of the most common cyber threats in the world today. It is a type of social engineering attack where the attacker lures unsuspecting victims into carrying out certain tasks mostly to steal personal and sensitive information. These stolen information are exploited to commit further crimes e.g. blackmails, data theft, financial theft, malware installation etc. This study was carried out to tackle this problem by designing an anti-phishing learning algorithm to detect phishing emails and also to study the accuracies of human phishing prediction to machine prediction. A graphical user interface was designed to emulate an email-client system that popped-up a warning on detecting a phishing mail successfully and collection of predictions made by expert and non-expert users on anti-phishing techniques. These predictions were compared to the predictions made by the machine learning algorithm to compare the efficiencies of all predictions considered in this research. The performance of the classifier used was measured with metrics such as confusion matrix, accuracy, receiver operating characteristic curve and area under graph
Full-text available
Phishing is one of the many types of cybercrime targeting internet users. A phishing message is sent with the aim to obtain information from a potential victim. One of the reasons phishing is popular has to do with the connectivity that the internet provides. A message can be spread to thousands of recipients with little effort and at negligible cost. A successful phishing attack can lead to identity theft and loss of money for the victims.When an organisation is targeted, phishing can lead to, among other things, compromised network security and stolen intellectual property. Phishing is highly scalable. On the other side of the scalability spectrum are less scalable modus operandi. We categorise less scalable methods as “fishing for information”. In this thesis, we aim to explore the spectrum of scalability. This thesis uses a socio-technical approach by describing both experiments and technical perspectives to “fishing” and phishing. This thesis starts by exploring definitions of phishing in literature and analysing their concepts. This provides us with a foundation of what constitutes phishing. Following on the definition, we explore two modus operandi that are less scalable than phishing, using USB keys and QR codes.We focus on measuring attack effectiveness on the boundary between the physical (i. e., objects on the floor) and digital world (i. e., getting a computer virus.) By quantifying the effectiveness of an attack using experiments, we investigate the feasibility of less scalable attacks. Then, we investigate the thought patterns that potential victims use in order to assess a phishing email. The thought patterns, or heuristics, determine whether a recipient of phishing becomes a victim or not. Knowledge on people’s thought patterns can be used to improve user training. Subsequently, we created a anti-phishing training to be provided to children.We show that training children is feasible and increases their ability to detect phishing on the short term. Finally, we performed a large-scale analysis of phishing emails in the Netherlands.We discuss patterns in terms of both attacker behavior as well as recipient behaviour. Our results demonstrate the effectiveness of phishing with different degrees of scalability. Less scalable methods of attack require more effort on the part of the attacker, but provide higher effectiveness. More scalable attacks provide lower success rates, but require less effort than scalable attacks. The contributions in this thesis allow researchers and security professionals to better understand the dynamic nature of phishing.
Conference Paper
Full-text available
A first contribution of this paper is a theoretical yet practically applicable model covering a large set of phishing attacks, aimed towards developing an understanding of threats relating to phishing. We model an attack by a phishing graph in which nodes correspond to knowledge or access rights, and (directed) edges correspond to means of obtaining information or access rights from already possessed information or access rights – whether this involves interaction with the victim or not. Edges may also be associated with probabilities, costs, or other measures of the hardness of traversing the graph. This allows us to quantify the effort of traversing a graph from some starting node (corresponding to publicly available information) to a target node that corresponds to access to a resource of the attacker’s choice. We discuss how to perform economic analysis on the viability of attacks. A quantification of the economical viability of various attacks allows a pinpointing of weak links for which improved security mechanisms would improve overall system security.
Acknowledgments The author acknowledges,sponsorship,from,the U.S. Department,of Homeland Security, Science and Technology Directorate (DHS S&T). Points of view in this document,are those of the author and do not necessarily,represent,the official position of the U.S. Department,of Homeland Security or the Science and Technology,Directorate. The content of this report was shaped,by the members of the Identity Theft Technology Council, a public-private partnership between DHS S&T, SRI International, the Anti-Phishing Working Group (APWG), and private industry. Particular thanks are due to Dan Boneh, Drew Dean, Louie Gasparini, Ulf Lindqvist, John Mitchell, Peter Neumann, Robert Rodriguez, Jim Roskind and Don Wilborn for their contributions. Intended Audience This report is intended,for technically sophisticated readers,such as security practitioners, executives, researchers, and others who wish to understand methods,employed,by online identity thieves and countermeasures,that can prevent such crimes. Executive Summary Phishing is online identity theft in which confidential information is obtained,from an individual. Phishing includes deceptive attacks, in which users are tricked by fraudulent messages into giving out information; malware attacks, in which malicious software causes data compromises; and DNS-based attacks, in which the lookup of host names is altered to send users to a fraudulent server. The Gartner group estimates,that the direct phishing-related loss to US banks and credit card issuers in 2003 was $1.2 billion. Indirect losses are much higher, including customer service expenses, account replacement costs, and higher expenses,due to decreased,use of online services in the face of widespread,fear about the security of online financial transactions. Phishing also causes
'Phishing' is a fraudulent activity defined as the creation of a replica of an existing Web page to fool a user into submitting personal, financial, or password data. There are security service guidelines for both software security and web site security development environments. Developers use these guidelines when planning new systems (or during re-engineering of existing systems) to ensure a secure environment. The purpose of this paper is two-fold: firstly to consider the characteristics of a phishing attack and to identify a list of issues relevant to it; and secondly, to compare the nature of a phishing attack with the security services guidelines provided and to pinpoint the weakness(es) of phishing attacks if these guidelines are adhered to.
Conference Paper
Most computer literacy courses encountered by college students in a non-technical major encompass a foundation set of computing skills including efficient use of word processing, spreadsheet, database, and presentation software. Yet current college graduates are facing fresh challenges as end-users in a work force transformed by legislation that is revolutionizing digital data communication, by nearly boundary-less computer systems that include mobile and static devices, and by employer expectations for safeguarding critical data resources. For example, data privacy legislation affects all end-users of computer systems in the workplace. As employees, new graduates will have access to critical data to perform their jobs, yet they could be the weakest link in an otherwise effectively secure computer system, primarily because of inadequate education, negligence, and inexperience. Technical and mathematical computer security has progressed substantially in the last few years, but new graduates are typically lacking in the knowledge of computer security as a fundamental component of their workplace roles. This paper proposes a computer literacy course content and structure that incorporates substantial practice in end-user computer security.
How Can We Stop Phishing and Pharming Scams
  • P L Kerstein
Kerstein P.L., "How Can We Stop Phishing and Pharming Scams?", accessed 27 Nov 2005
Degunking Your Email , Spam, And Viruses
  • J Duntemann
  • Duntemann J.
Duntemann J., Degunking Your Email, Spam, And Viruses. Scottsdale, Arizona: Paraglyph Press, 2004
A convicted hacker debunks some myths
  • Cnn
  • Com
  • CNN
CNN. com, "A convicted hacker debunks some myths." k.cnna/index.html 2005, accessed 01/06/06
Cyber-looters Capitalize on Katrina
  • Paul F Roberts
Roberts, Paul F. "Cyber-looters Capitalize on Katrina." eWeek. 12 Sept. 2005: 11-12