Kruger 2017
Kruger 2017
a r t i c l e i n f o a b s t r a c t
Article history: Phantom cell phone experiences are instances when mobile cell phone users perceive ringing, vibrations,
Received 8 September 2015 and/or visual stimuli indicating an incoming call or message, yet no call or message actually registers on
Received in revised form the phone. As a human signal-detection issue, these experiences may be influenced by psychological
5 January 2017
factors. We hypothesize that phantom cell phone experiences are a symptom of psychological de-
Accepted 10 January 2017
Available online 13 January 2017
pendency on cell phone communications and the resulting hypersensitivity to stimuli in the same class
as cell phone notification methods. We find support for this hypothesis with path model analyzing re-
sponses from a large undergraduate sample. Self-reported symptoms of cell phone dependency strongly
Keywords:
Phantom ringing
predicted the frequency of phantom cell phone experiences and mediated the relationships between
Phantom vibrations personality factors, age, and participant sex with the frequency of phantom cell phone experiences.
Cell phones Women, younger individuals, and those with lower conscientiousness and emotional stability (i.e. higher
Dependency neuroticism) had higher symptoms of cell phone dependency.
Addiction © 2017 Elsevier Ltd. All rights reserved.
Personality
http://dx.doi.org/10.1016/j.chb.2017.01.017
0747-5632/© 2017 Elsevier Ltd. All rights reserved.
D.J. Kruger, J.M. Djerf / Computers in Human Behavior 70 (2017) 360e364 361
phone experiences (Parisi, 2013). Those higher in extraversion for interpersonal communication (Drouin et al., 2012). Those who
enjoy being around other people more than being alone, are used phones in the classroom and when eating, a possible sign of
attention-seeking, assertiveness, sociable, and enjoy being the dependency, were more likely to have phantom experiences (Subba
center of attention. Extraverts have stronger desires to create and et al., 2013). Drouin et al. (2012) found that those who checked for
maintain social relationships, and the social support functions of text messages more often and were more anxious when they did
the Internet appear to be a basis for Internet addiction (Young, not receive immediate replies were more bothered by phantom
1998). Both extraversion and anxiety predict mobile phone addic- vibrations. The relationship between the amount of cell phone use
tion (Bianchi & Phillips, 2005; Hong, Chiu, & Huang, 2012). The and the likelihood of phantom experiences varies by study
larger social network and greater level of interactions with others (Catchings, Bush, Copes, Schaefer, & Wixom, 2010). Mobile phone
associated with extraversion may lead to greater expectations for, dependency does predict levels of mobile phone usage (Billieux,
as well as actual, social messaging through one's phone. Linden, & Rochat, 2008, 2007; Hong et al., 2012), however it is
Those higher in conscientiousness tend to be organized and important to distinguish between frequency of use and psycho-
dependable, show self-discipline, are motivated for achievement, logical aspects of phone use, such as dependency (Walsh, White, &
and prefer planned over spontaneous behavior. Lin, Chen, Li, and Young, 2010). High usage levels of social technologies are not al-
Lin (2013) found that novelty seeking, but not harm avoidance or ways predictive of addiction symptoms (Charlton & Danforth,
reward dependence, predicted phantom vibration and ringing. As 2007). Psychological interpretations of messaging, rather than the
conscientiousness is related to self-monitoring, those higher in actual amount of messaging, may be the proximate cause of
conscientiousness are better at self-monitoring and may experi- symptoms such as phantom experiences (Drouin et al., 2012).
ence phantom phone events less frequently (as they are not actu-
ally occurring). Recognizing that individuals higher in
1.1. Overview and hypotheses
conscientiousness are less prone to addictive tendencies, including
Internet addictions (Kuss & Griffiths, 2011; Wilson, Fornasier, &
Psychological dependency on cell phones, based on the rein-
White, 2010), Drouin et al. (2012) found that those higher in
forcing aspects of interpersonal communication they provide, may
conscientiousness were less likely to experience phantom
increase the likelihood of phantom cell phone communication ex-
vibrations.
periences. The perception of incoming messages is a function of
Higher neuroticism is the tendency to experience unpleasant
human signal-detection, and sensitization to rewarding stimuli is a
emotions easily, such as anger, anxiety, depression, and vulnera-
documented feature of psychological addiction, including hyper-
bility. The inverse of neuroticism is emotional stability, with higher
sensitivity to the entire class of stimuli related to the one that is
impulse control and lower reactivity. Those high in emotional
rewarding (Olsen, 2011). Thus, sensations similar to the ringing,
stability have a stable and calm personality and are less excitable.
vibration, and visual displays of phones may be interpreted as
Those high in neuroticism respond poorly to stressors and often
messages. Those who are more psychologically dependent on cell
interpret common situations as threatening. Extraversion and
phone communication may be more likely to interpret ambiguous
neuroticism predict the strength of emotional responses to text
sensations (whether sonic, tactile, or visual) as incoming calls and
messages, extraversion predicts excessive phone use whereas
messages. Personality factors related to the likelihood of phantom
neuroticism predicts use for relationship maintenance (Igarashi,
cell phone experiences are likely mediated through their relation-
Motoyoshi, Takai, & Yoshida, 2008). Those higher in neuroticism
ship to symptoms of cell phone dependency (See Fig. 1). The Mobile
may experience phantom phone events more frequently that those
Phone Problem Use Scale (MPPUS; Bianchi & Phillips, 2005) con-
higher in emotional stability, as they are more reactive, excitable,
tains several items on negative reinforcement (removal of an
and more likely to misinterpret stimuli.
adverse stimulus, such as negative affect) and positive reinforce-
The associations between personality factors and experiences
ment (a rewarding stimulus) via cell phone use. Thus, this subset of
phantom phone events may be mediated by relationships with
items may be used to assess cell phone dependency.
psychological dependency on cell phones. An addiction is a
recognized pathological medical condition characterized by
compulsive engagement in rewarding stimuli, despite adverse 2. Method
consequences (Malenka, Nestler, & Hyman, 2009). Addictive stim-
uli are reinforcing and intrinsically rewarding, that is they increase 2.1. Participants
the likelihood that individuals will seek repeated experience and
they are perceived as positive or desirable, respectively (Malenka Participants were 766 undergraduate students (384 women and
et al., 2009). 382 men) recruited from the Psychology Subject Pool of a large
An attempt by researchers to add Mobile Phone/Technology public university in the Midwestern USA. Participants were un-
addiction to the Diagnostic and Statistical Manual of Mental Dis- aware of the topic of the study before signing up for the study,
orders (DSM-5) was unsuccessful (Deb, 2015). Thus, although which was listed by its alphanumeric IRB code. The average age of
compulsive or excessive cell phone use may appear to have the participants was 19 years (SD ¼ 1, range 16e28). Participants
addictive properties, it may be more accurately described in terms could identify with multiple ethnicities: 45% Western European,
of psychological dependence than addiction. Psychological depen-
dence involves withdrawal symptoms upon cessation of drug use
or engagement in certain behaviors that develop through consis-
tent and frequent exposure to a stimulus (Malenka et al., 2009).
Although clinicians distinguish addiction and psychological
dependence as respectively mediated by positive reinforcement (a
rewarding stimulus) and negative reinforcement (removal of an
adverse stimulus, such as negative affect), we consider both types
of reinforcement to be related to psychological cell phone
dependency.
Dependency on cell phone messaging may be based on a need Fig. 1. Psychological process of phantom cell phone experiences.
362 D.J. Kruger, J.M. Djerf / Computers in Human Behavior 70 (2017) 360e364
36% Eastern European, 13% East Asian, 9% Latino/a/Hispanic, 8% experiences, similar to models in other studies (e.g., Drouin et al.,
South Asian, 8% African American, 7% Arab/Middle-Eastern, 3% 2012; Hong et al., 2012; Igarashi et al., 2008). We ran iterated an-
Native American/Alaskan native, 1% Pacific Islander, and 6% Other; alyses until the parsimonious model was reached, including all
20% were not of European ancestry. All participants reported significant relationships hypothesized in the theoretical model (See
regularly carrying and using a mobile phone. Fig. 2). The first model included all possible predicted relationships;
non-significant relationships were deleted one at a time in suc-
2.2. Procedure cessive models based on the highest p-value until all relationships
in the model were statistically significant. Modification indices
Participants were recruited from four terms of the introductory were then used to free prospective relationships among personality
psychology subject pool (Winter 2015, Fall 2015, Winter 2016, Fall variables. Although freeing relationships among personality vari-
2016). All participants received a research credit for participation, ables substantially improved model fit, it did not affect the pattern
completed online consent forms, and were then given access to an of other significant relationships.
online anonymous survey that they could complete at their
convenience. 3. Results
2.3. Measures Almost all respondents (99%) had a touch screen phone
(including iPhone, Android, or Windows based operating systems),
Participants first completed the Ten Item Personality Inventory four respondents had a flip phone without a touch screen, two
(TIPI; Gosling, Rentfrow, & Swann, 2003) assessing the Big Five respondents had a slide phone with a full hard keyboard, one
personality dimensions (Openness, Conscientiousness, Extraver- respondent had an older style phone (Palm, Blackberry, etc.) with a
sion, Agreeableness, and Emotional Stability/Neuroticism). Partici- full hard keyboard, and one respondent did not know what kind of
pants then saw the item, “Some people have reported phantom phone he had. The most common phantom phone experience was
vibrations or phantom ringing from their cell phones, when it phantom vibrations (77%), followed by phantom notification (47%)
seems like they get a call or a message but do not see anything and phantom ringing (37%). The phone dependency scale had
when they check their phones. Has this every happened to you?” excellent inter-item reliability (Cronbach's alpha ¼ 0.816). At the
Participants separately indicated whether or not they experienced bivariate level, phone dependency symptoms predicted the fre-
“Phantom ringing,” “Phantom vibration,” and “Phantom notifica- quency of phantom cell phone experiences (See Table 1). Women
tion (image on the screen).” Those who responded “Yes” were reported higher phone dependency symptoms than did men; those
shown additional items on the next survey page, asking “How often higher in conscientiousness and emotional stability (i.e., lower in
do you experience phantom ringing?” (Several times a day, About neuroticism) had lower phone dependency symptoms.
once a day, Several times a week, About once a week, Several times As expected, the parsimonious path model had an excellent fit
a month, About once a month, Less than once a month). Data for to the data, c2(19, N ¼ 766) ¼ 15.00, p ¼ 0.723, GFI ¼ 0.996,
these items were recoded into one variable for each type of AGFI ¼ 0.990, NFI ¼ 0.968, CFI ¼ 1.00, RMSEA ¼ 0.000. Unique
phantom sensation, so that “No” in the screening question (i.e., relationships in the path model indicated that older participants
Never) ¼ 0 and Several times a day ¼ 7. These scores were summed were higher in openness and had lower phone dependency (See
to create a phantom experiences scale. Fig. 2). Women were higher in conscientiousness, extraversion, and
Participants responded to six items from the Mobile Phone agreeableness, and were lower in emotional stability than men.
Problem Use Scale (MPPUS; Bianchi & Phillips, 2005) that appeared Women reported greater cell phone dependency than men, and
to assess symptoms of cell phone dependency (7 options from men reported more phantom phone experiences, independent of
Disagree Strongly to Agree Strongly); “I have used my mobile phone other factors.
to make myself feel better when I was feeling down; I have used my Sobel tests indicated that cell phone dependency significantly
mobile phone to talk to others when I was feeling isolated; I feel mediated the relationship between sex and the frequency of
anxious if I have not checked for messages or switched on my phantom phone experiences, Z ¼ 4.67, p < 0.001, the relationship
mobile phone for some time; I become irritable if I have to switch between age and phantom phone experiences, Z ¼ 2.69, p ¼ 0.007,
off my mobile phone for meetings, dinner engagements, or at the the relationship between conscientiousness and phantom phone
movies; I feel anxious if I have not received a call or message in experiences, Z ¼ 2.77, p ¼ 0.006, and the relationship between
some time; When I am not using my cell phone, I am thinking about emotional stability (i.e. neuroticism) and phantom phone experi-
using it or planning the next time I can use it.” Scores from these ences, Z ¼ 4.05, p < 0.001. Both conscientiousness, Z ¼ 2.92,
times items were averaged to create a phone dependency scale. p ¼ 0.003, and emotional stability (i.e. neuroticism), Z ¼ 3.75,
Other (unused) items on the MPPUS, such as “I have received mo- p < 0.001, significantly mediated the relationship between sex and
bile phone bills I could not afford to pay,” “All my friends own a cell phone dependency. Sex and phone dependency predicted
mobile phone,” assess other constructs. phantom phone experiences, in combination accounting for 9.1% of
All participants indicated the type of mobile phone that they the variance.
have: Full touch screen (iPhone, Android, Windows); Flip phone
(without touch screen); Hybrid phone with both a touchscreen and 4. Discussion
a full hard keyboard; Older style Palm/Blackberry/etc. with a hard
keyboard; and Don't know. The survey ended with demographic We supported our hypothesis that phantom cell phone experi-
questions (e.g. sex, race, age) after the measures outlined above. ences are related to psychological dependency on cell phone
communications. These experiences may in fact be a component of
2.4. Analyses psychological dependency, as one of the psychological features of
addiction is hypersensitivity to stimuli similar to the ones that are
We ran descriptive statistics and bivariate correlations with rewarding (Olsen, 2011). Self-reported symptoms of cell phone
scale values (see Table 1). We created a structural path model with dependency included negative reinforcement where cell phone use
phone dependency and personality factors mediating the rela- alleviates adverse emotions, adverse emotions created by barriers
tionship between demographic indicators and phantom to cell phone use, and obsessive thoughts regarding cell phone use
D.J. Kruger, J.M. Djerf / Computers in Human Behavior 70 (2017) 360e364 363
Table 1
Descriptive statistics for and correlations among variables, N ¼ 766.
1 2 3 4 5 6 7 8
1. Phantom experiences 1
2. Phone dependency 0.268*** 1
3. Openness 0.057 0.017 1
4. Conscientiousness 0.020 0.116*** 0.048 1
5. Extraversion 0.005 0.007 0.329*** 0.094** 1
6. Agreeableness 0.066 0.010 0.108** 0.113** 0.033 1
7. Emotional stability 0.038 0.233*** 0.061 0.243*** 0.148*** 1
8. Sex 0.024 0.228*** 0.074* 0.135*** 0.079* 0.185*** 0.181*** 1
9. Age in years 0.013 0.130*** 0.091* 0.042 0.031 0.020 0.037 0.146***
M 2.93 4.23 5.07 5.32 4.52 4.80 4.52 1.50
SD 1.63 1.23 1.12 1.20 1.52 1.14 1.34 0.50
* ** ***
p < 0.05, p < 0.05, p < 0.001; Age in years: M ¼ 18.65, SD ¼ 1.04.
Mobile Phone Problem Use Scale (MPPUS; Bianchi & Phillips, 2005).
For example, “I feel anxious when people don't immediately reply
to my text message,” (ER) and “I feel anxious if I have not received a
call or message in some time,” (MPPUS). However, the items in the
Emotional Reaction subscale focus more narrowly on checking for
text messages, whereas the items we used from the MPPUS are a
broader assessment of the psychological aspects of cell phone
dependency.
We were surprised at the lack of direct relationships between
personality and the frequency of phantom phone experiences (see
Table 1). Personality factors may have a real but weak relationship
to phantom cell phone experiences that may not have been iden-
tified in the current study. For comparison, the relationship be-
tween conscientiousness and phantom vibrations identified by
Drouin et al. (2012) corresponds to an r-value of 0.150. Identifying
Fig. 2. Results of the path analysis. N ¼ 766, c2(19) ¼ 15.00, p ¼ 0.723, GFI ¼ 0.996, relationships with the same effect size with statistical power of
NFI ¼ 0.968, CFI ¼ 1.00, RMSEA ¼ 0.000. For Sex, Female ¼ 1, Male ¼ 2; Coefficients for
relationships between personality factors depicted in inset table: O ¼ Openness, C ¼
0.80 (1 - b) would require a sample size of 347. The current sample
Conscientiousness, E ¼ Extraversion, A ¼ Agreeableness, ES ¼ Emotional Stability. size of 766 would be able to identify a correlation of 0.101, ¼ 0.101,
Coefficients are p < 0.001 unless indicated; * indicates p < 0.05, ** indicates p < 0.01. marginally above 1% shared variance, with a statistical power of
0.80 (1 - b). This represents the typical value for a small effect, an
effect size (d) of 0.20 (See Cohen, 1988). Thus our analyses are not
when not using one's phone. In our study, participants who expe- underpowered, and future research will need to address this
rienced these aspects of dependency more frequently over- discrepancy to provide more insight.
perceived the experience of the stimuli that they craved. We
assessed multiple sensory modes of phantom cell phone experi-
ences, auditory, tactile, and visual. Phantom vibration was the most 4.1. Limitations and conclusion
common, experienced by over three-quarters of participants.
Self-reported symptoms of cell phone dependency mediated the The use of university undergraduates as participants creates the
relationships between personality factors, age, and participant sex risk that results will not generalize across the population, though
with the frequency of phantom cell phone experiences. Women, there may not be any specific reason why they are not suitable for a
younger individuals, and those with lower conscientiousness and test of the hypotheses. Cell phones are part of their social fabric
emotional stability (i.e. higher neuroticism) had higher symptoms contemporary university undergraduates in the United States
of cell phone dependency. As there were no overall sex differences (Chen & Katz, 2009). Our brief personality measures may reduce
in the frequency of phantom phone experiences (see Table 1), the predictive power, though the TIPI has demonstrated adequate
significant unique relationship between participant sex and the levels of convergence with widely used Big-Five measures in self,
frequency of phantom phone experiences (See Fig. 2) indicates that observer, and peer reports; test-retest reliability; patterns of pre-
a greater portion of the variance in frequency of phantom phone dicted external correlates; and convergence between self and
experiences is explained by model factors for women than for men. observer ratings (Gosling et al., 2003).
Drouin et al. (2012) found a relationship between dependency We rely on retrospective self-report data; inaccuracies in self-
on text messages and how bothered people were by phantom vi- reporting phantom phone experiences would contribute noise to
brations, but not between dependency on text messages and fre- our analyses and thus make it more difficult to identify the pre-
quency of phantom vibrations. Compulsive checking of text dicted relationship. Thus, this weakness is a hurdle for identifying
messages and negative emotional reactions to lack of incoming significant relationships rather than a confounding factor offering
messages are intuitively related to finding phantom text messages an alternative explanation for identified relationships.
bothersome, and may even load on the same factor psychometri- It may be possible to conduct a prospective study with experi-
cally. What could explain this discrepancy in results? There is some ence sampling methods for replication and extension of this work.
overlap between Igarashi, Motoyoshi, Takai, and Yoshida's (2008) However, using a smartphone to record phantom phone experi-
Emotional Reaction subscale, utilized by Drouin et al. (2012) to ences may be problematic, especially if the interactive use of the
indicate text message dependency, and the items we used from the smartphone is an aspect of the rewarding stimuli that participants
may crave.
364 D.J. Kruger, J.M. Djerf / Computers in Human Behavior 70 (2017) 360e364
Overall, our study provides additional insights on human in- Deb, A. (2015). Phantom vibration and phantom ringing among mobile phone
users: A systematic review of literature. Asia-Pacific Psychiatry, 7, 231e239.
teractions with social technology, in an area where scientific liter-
http://dx.doi.org/10.1111/appy.12164.
ature is scarce. We provide evidence that those with greater Drouin, M., Kaiser, D., & Miller, D. (2012). Phantom vibrations in young adults:
psychological dependency on cell phone communications experi- Prevalence and underlying psychological characteristics. Computers in Human
ence hyper-sensitization to stimuli similar to those indicating cell Behavior, 28, 1490e1496.
Gosling, S. D., Rentfrow, P. J., & Swann, W. B., Jr. (2003). A very brief measure of the
phone communications, and may even actively hallucinate these Big Five personality domains. Journal of Research in Personality, 37, 504e528.
sensations. These results may help advance the discussion Hong, F.-Y., Chiu, S.-I., & Huang, D.-H. (2012). A model of the relationship between
regarding the possibility of extending the definitions of addiction psychological characteristics, mobile phone addiction and use of mobile phones
by Taiwanese university female students. Computers in Human Behavior, 28,
into the technological domain. 2152e2159.
Igarashi, T., Motoyoshi, T., Takai, J., & Yoshida, T. (2008). No mobile, no life: Self-
Author disclosure statement perception and text-message dependency among Japanese high school stu-
dents. Computers in Human Behavior, 24, 2311e2324.
Kuss, D. J., & Griffiths, M. D. (2011). Online social networking and addiction e a
No competing financial interests exist. review of the psychological literature. International Journal of Environmental
Research and Public Health, 8, 3528e3552.
Lin, Y., Chen, C., Li, P., & Lin, S. (2013). A dimensional approach to the phantom
Acknowledgments vibration and ringing syndrome during medical internship. Journal of Psychi-
atric Research, 49, 1254e1258.
We thank the University of Michigan's Undergraduate Research Malenka, R. C., Nestler, E. J., & Hyman, S. E. (2009). Reinforcement and addictive
disorders. In A. Sydor, & R. Y. Brown (Eds.), Molecular neuropharmacology: A
Opportunities Program for partial funding of this project. We also
foundation for clinical neuroscience (2nd ed., pp. 364e375). New York: McGraw-
thank the participants for their time and effort in responding to our Hill Medical.
survey. McCrae, R. R., & Terracciano, A. (2005). Personality profiles of cultures: Aggregate
personality traits. Journal of Personality and Social Psychology, 89, 407e425.
Olsen, C. M. (2011). Natural rewards, neuroplasticity, and non-drug addiction.
References Neuropharmacology, 61, 1109e1122.
Parisi, D. (2013, January 22). Banishing phantoms from the skin: “vibranxiety” and
Bianchi, A., & Phillips, J. G. (2005). Psychological predictors of problem mobile the pathologization of interfacing. Flow, 17 [Cited 21 Feb 2014.] Available from
phone use. Cyberpsychology, Behavior, & Social Networking, 8, 39e49. URL http://flowtv.org/2013/01/banishing-phantoms-from-the-skin/.
Billieux, J., Linden, M. V. D., D’acremont, M., Ceschi, G., & Zermatten, A. (2007). Does Rothberg, M., Arora, A., Hermann, J., Kleppel, R., St. Marie, P., & Visintainer, P. (2010).
impulsivity relate to perceived dependence on and actual use of the mobile Phantom vibration syndrome among medical staff: A cross sectional survey.
phone? Applied Cognitive Psychology, 21, 527e537. British Medical Journal, 341, c6914.
Billieux, J., Linden, M. V. D., & Rochat, L. (2008). The role of impulsivity in actual and Subba, S., Mandelia, C., Pathak, V., Reddy, D., Goel, A., Tayal, A., et al. (2013). Ring-
problematic use of the mobile phone. Applied Cognitive Psychology, 22, xiety and the mobile phone usage pattern among the students of a medical
1195e1210. college in South India. Journal of Clinical and Diagnostic Research, 7, 205e209.
Catchings, D., Bush, L., Copes, B., Schaefer, A., & Wixom, E. (2010). Cell phone use Tanner, W. P., Jr., & Swets, J. A. (1954). A decision-making theory of visual detection.
and psychological correlates: An explorative study. In Paper presented at the Psychological Review, 61, 401e409.
2010 undergraduate research and scholarship conference, Idaho, Unites States Walsh, S. P., White, K. M., & Young, R. M. (2010). Needing to connect: The impact of
[Cited 21 Feb 2014.] Available from URL http://scholarworks.boisestate.edu/hs_ self and others on young people's involvement with their mobile phone.
10/12/. Australian Journal of Psychology, 62, 194e203.
Charlton, J. P., & Danforth, I. D. W. (2007). Distinguishing addiction and high Wilson, K., Fornasier, S., & White, K. M. (2010). Psychological predictors of young
engagement in the context of online game playing. Computers in Human adults' use of social networking sites. Cyberpsychology, Behavior, & Social
Behavior, 23, 1531e1548. Networking, 13, 173e177.
Chen, & Katz. (2009). Extending family to school life: College students' use of the Young, K. S. (1998). Internet addiction: The emergence of a new clinical disorder.
mobile phone. International Journal of Human-Computer Studies, 67, 179e191. Cyberpsychology, Behavior, & Social Networking, 1, 237e244.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). New
York: Academic Press.