Instant Messaging Social Networks: Individual, Relational, and Cultural Characteristics Gustavo S. Mesch, Ilan Talmud (presenter) University of Haifa, and Anabel Quan-Haase University of Western Ontario,
Abstract Prevalent research on information and communication technologies (ICTs) tends to focus on either individual or group level characteristics, while neglecting to investigate the effect of relational variables on communication. We collected survey data in Israel (N = 492) and Canada (N = 293) to investigate the effect of individual, relational and cultural variables on frequency and topic multiplexity of Instant Messaging communication among dyads of university students.
Online communication and time spent online on socializing activities Seems to increase the closeness of existing social ties, supplementing face-to-face communication with close friends, distant friends, and family Links intimate friends who are temporarily geographically separated, as they take jobs or enroll in college Provides an alternative channel, filling intervals between face-to-face meetings and facilitating the coordination of social gatherings (QuanHaase, 2008).
Instant Messaging Most studies have focused on the effects of individual-level variables. The average IM session is longer for females than for males, because in part females use IM to establish, nurture, and develop personal relationships, while males use it for providing information (Baron, 2004).
IM use has a narrower range of purposes. It promotes rather than hinders intimacy, with frequent IM conversations encouraging the desire to meet face-toface with friends (Hu et al., 2004). Main uses of IM are for socializing, event planning, task accomplishment, and meeting new people (Flanagin, 2005; Grinter & Pallen, 2002; Walther, 1996).
IM for Relations Bringing Relations Back in Little relational or comparative research is available in Israel, IM is their most frequent Internet activity, with 77% of high school students reporting using IM on a daily basis. IM use was more popular with girls than boys, 70% and 64% respectively (Lemish, Ribak, & Alony, 2009).
IM differs from other digital textual communication channels such as email, SMS, and chat rooms. First, IM users generally engage in messaging with others whom they know. One-to-one and small group chats characterize IM's use in peer groups and the workplace, where IM is considered an essential communication tool (Grinter & Palen, 2002). IM communication was mostly restricted to one's "real space friends," people who first met face-to-face in physical settings such as schools or summer camps (Grinter & Palen, 2002). IM use was associated with the perceived intimacy between friends (Hu et al., 2004).\ The amount of IM use was also positively associated with verbal, affective, and social intimacy. Indeed, frequent conversation via IM actually encouraged the desire to meet face-to-face.
Studies have shown that because of its inherent positive network externalities, the rate of adoption among peers has grown exponentially by sheer in-group influence. Hence, IM is well suited for the study of social ties.
Two Cases Comparison: Students in Canada and Israel Immigration absorption Size Federated polity State involvement and cultural legacy Ethnic and immigrant segregation Civil Society State of War + conscription
Type of Relationship RQ1: Does the communicators' type of relationship have an effect on the frequency and content of IM communications? H1: A tie’s perceived closeness will be positively associated with more frequent communication and with the inclusion of more topics of conversation.
Tie Homophily H2: Same-sex relationships will display more frequent communication, and on a larger number of topics, than cross-sex relationships.
Propinquity RQ2: Does propinquity affect the frequency and multiplicity of topics of IM communications?
Tie Duration H3: As engagement duration between communicators increases, relational strength intensifies. Hence, the longer the relationship duration, the higher the frequency of IM communication and the greater the number of topics discussed.
Method The data collection took place in Canada and Israel between October 2005 and August 2006 . Size: Canada N=293. Israel N= 492; Mean age: Ca = 21; ISR = 25. Female: Ca=68%, ISR = 51% Majority Social sciences. Across grade levels
Name Generator for two ties Dependent Variables (dyadic): Communication frequency for each pair (i, j) Topic multiplexity
Independent Variables Relational: Tie duration Perceived Closeness Relationship Type Individual (ego): Age, gender, marital status.
Because we used slightly different model specifications, we used the BIC as a modelselection test to complement our analytic strategy, and to examine the extent to which we lost parsimony at the expense of specificity. The BIC indicates the relative likelihood that a model fits a given data set. The calculation of BIC for the Ordinary Least Squares models is: BIC(q) = N [ln(1 - R2(q)] + p(q) [ln (N)] ; where N is the number of cases, and p is the
Relationship type with IM contacts Israel and Canada (N=293) First IM Contact
Second IM Contact
Israel _______
Canada _______
Israel ______
Canada _______
Family
11%
16%
10%
20%
Close Friend
48%
42%
44%
49%
Distant Friends
23%
12%
35%
16%
Romantic Partner
14%
21%
5%
6%
Online Tie
4%
1%
6%
1%
Conclusion Individual level characteristics, such as: age, gender, and marital status proved not to be associated with communication frequency in either Canada or Israel. The extent of IM contacts was positively associated with communication frequency for both contacts in Israel and Canada. The extent of daily IM interaction was also positively associated with communication frequency for both contacts in both countries. The larger the IM network of respondents, and the more frequently they interact via IM, the more likely they are to communicate frequently with their first and second contacts.
In Canada and Israel, if the residence of the contact was at the university dorms or in the same city, communication frequency was greater than in the cases of contacts residing in another city or in another country.
Distance clearly exerts a significant effect on communication frequency, despite IM’s negligible cost in comparison with land lines and cell phones. Relationship type influenced communication frequency: with a romantic partner it was greater than with a close friend. In contrast, IM communication with an online friend or family member was less frequent than with a close friend. IM communication with a distant friend was also less frequent than with a close friend. Results for relationship type showed a similar pattern for Israel and Canada. Surprise: gender homophily was not founded to have significant net effect in any model.
In both countries, we found that university students living in geographical proximity communicated more frequently and on more topics than students living at geographical distance did. Relationship type had an effect on patterns of IM communication: communication with close friends and romantic partners was more frequent than with distant friends and family ties. Perceived closeness was positively associated with comm. frequency and topic multiplexity. IM communication patterns reflect the type and nature of existing relationships between contacts rather than individual attributes and cultural characteristics of the
Future Research Same Relational patterns on two countries, but: Needs multi-cases, multi-level studies Needs real records rather then self reported evidence. Single source bias. From tie approach into whole network or partial network approach Needs comparison of various relational media Problem of media integration: multi-purpose
Table1. Relationship type with IM contacts Israel and Canada (N=293) First IM Contact
Second IM Contact
Israel _______
Canada _______
Israel ______
Canada _______
Family
11%
16%
10%
20%
Close Friend
48%
42%
44%
49%
Distant Friends
23%
12%
35%
16%
Romantic Partner
14%
21%
5%
6%
Online Tie
4%
1%
6%
1%
Table2. OLS regression predicting communication frequency with tie—Canada First Contact Model 1 ________________ -.03 -.066 (.026) -.079 -.032 (.122) .071 .031 (.124) -.171 -.075 (.140) .545** .237 (.116) .436** .319 (.067) .154* .135 (.060) .160 .017 (.470) .035 .012 (.162) -.918** -.286 (.173) .987** .354 (.195) -----------
Model 2 ____________ -.031 -.058 (.026) -.054 -.021 (.120) .087 .038 (.122) -.157 -.069 (.141) .555** .241 (.121) .418** .305 (.067) .148* .129 (.059) .220 .023 (.486) .162 .054 (.169) -.707** -.220 (.220) .876** .314 (.196) ---------
Constant
5.381** (.655)
-.106* (.054) .338* (.134) 4.106** (.950)
Adjusted R Square
.390
BIC F-Ratio
-82.261
Age Gender(1=male) Single Gender similarity Propinquity Frequency of IM use # of IM regular contacts Online Family Distant friend Partner Close friends (omitted category) Relationship duration Closeness to tie
**p < .01 *p < .05
Second Contact Model 3 ____________ -.014 -.028 (.028) -.125 -.052 (.129) -.038 -.017 (.120) -.026 -.011 (.138) .675** .305 (.124) .422** .320 (.072) .147* .134 (.064) .100 .011 (.496) .303 .111 (.163) -.488** -.175 (.169) .804** .173 (.279) ------------
Model 4 ____________ -.007 -.013 (.028) -.114 -.048 (.128) -.098 -.044 (.120) -.038 -.017 (.137) .706** .318 (.133) .418** .317 (.071) .138* .125 (.063) .407 .045 (.500) .320 .117 (.175) -.194 -.069 (.191) .676* .145 (.282) -------------
4.587** (.691)
-.033 (.061) .343** (.109) 3.015 (.860)
.407
.265
.287
-79.104 4.047
-27.47
-25.192 4.4
-.121 .169
-.038 .212
Table3. OLS regression predicting communication frequency with tie—Israel
Age Gender(1=male) Single Gender similarity Propinquity Frequency of IM use # of IM regular contacts Online Family Distant friend Partner Close friends (omitted category) Relationship duration Closeness to tie
First Contact Model 1 Model 2 ____________ ______________ -.011 -.028 -.011 -.028 (.020) (.020) -.075 -.030 -.069 -.027 (.123) (.121) .084 .033 .054 .021 (.123) (.121) .197 .078 .221 .088 (.135) (.133) .498** .196 .500** .197 (.122) (.120) .225** .185 .222** .183 (.062) (.061) .340** .263 .330** .256 (.066) (.065) -.421 -.058 -.126 -.017 (.364) (.369) -.316 -.077 -.325 -.079 (.223) (.220) -.631** -.213 -.434** -.146 (.154) (.162) .519** .147 .365* .103 (.195) (.198) -------------------
Constant
5.382** (.586)
-.118* (.060) .380** (.113) 3.818 (.741)
Adjusted R Square
.296
BIC F-Ratio
-104.02
**p < .01; *p < .05
Second Contact Model 3 Model 4 _____________ ____________ -.026 -.060 -.031 -.071 (.022) (.022) .001 .001 .021 .009 (.127) (.123) .177 .070 .156 .062 (.127) (.124) .011 .004 -.026 -.010 (.123) (.120) .507** .302 .505** .203 (.125) (.122) .196** .166 ..202** .170 (.064) (.062) .341** .261 .314** .240 (.069) (.068) -.456 -.070 -.032 -.005 (.344) (.349) -.101 -.025 -.126 -.031 (.216) (.210) -.577** -.221 -.269 -.103 (.142) (.156) .220 .044 .139 .028 (.258) (.252) ----------------------
-.110
5.744 (.640)
-.168* (.059) .444** (.104) 4.045 (.739)
.318
.258
.297
-107.364 7.857
-78.433
-92.604 13.57
.190
-.157 .254
Table 4. OLS regression predicting topic multiplexity of IM communication—Canada
Age Gender(1=male) Single Gender similarity Propinquity Frequency of IM use # of contacts Online Family Distant friend Partner Close friends (omitted category) Duration
First Contact Model 1 Model 2 _____________ ____________ -.071 -.057 -.082 -.066 (.072) (.70) -.096 -.560 -.599 -.102 (.334) (.326) .287 .053 .355 .066 (.340) (.332) -.118 -.633 -.338 -.063 (.385) (.382) -.078 -.421 .031 .006 (.319) (.329) .410* .128 .426* .133 (.185) (.181) .060 .023 .098 .037 (.164) (.160) -1.515 -.068 .076 .003 (1.293) (1.318) -.732 -.104 -1.138* -.161 (.446) (.459) -1.686** -.225 -.369 -.049 (.477) (.596) 1.799** .276 1.925** .295 (.537) (.532) ----------------------
Second Model 3 _____________ -.007 -.006 (.070) -.626* -.114 (.318) -.110 -.022 (.295) -.413 -.079 (.338) -.493 -.097 (.304) .716** .236 (.178) .088 .035 (.158) -2.835* -.135 (1.223) -1.32** -.210 (.402) -1.59** -.248 (.417) -.026 -.002 (.687) -----------
Contact Model 4 _____________ .005 .004 (.069) -.634* -.115 (.314) -.227 -.045 (.294) -.383 -.073 (.336) -.247 -.048 (.327) .701** .231 (.175) .055 .022 (.155) -1.948 -.093 (1.229) -1.49** -.238 (.430) -.831 -.129 (.470) -.182 -.017 (.691) --------------
4.303* (1.700)
.109 (.150) .829** (.267) -.088 (2.113)
Constant
6.996** (1.803)
.461** (.146) .841* (.364) .832 (2.579)
Adjusted R Square
.157
.202
.153
.183
BIC F-Ratio
12.671
7.917 7.857
13.843
14.656 5.172
Closeness to tie
**p < .01; *p < .05
.226 .180
.055 .222
Table 5. OLS regression predicting topic multiplexity of IM communication—Israel
Age Gender(1=male) Single Gender similarity Propinquity Frequency of IM use # of contacts Online Family Distant friend Partner Duration Closeness to tie Constant Adjusted R Square BIC F-Ratio
First Contact Model 1 Model 2 _______________ ___________ -.099 -.111 -.098* -.111 (.049) (.049) -.468 -.082 -.453 -.079 (.304) (.300) .151 .026 .085 .015 (.304) (.301) -.006 -.001 .046 .008 (.334) (.330) -.122 -.021 -.120 -.021 (.302) (.298) .521** .188 .514** .186 (.153) (.151) .319 .109 .294 .100 (.162) (.160) -1.976* -.118 -1.325 -.079 (.903) (.919) -1.276* -.136 -1.296* -.138 (.553) (.546) -1.065* -.158 -.627 -.093 (.378) (.402) 1.090* .135 .752 .093 (.484) (.491) .006 .003 (.148) .838** .184 (.281) 6.828** (1.454) .166 .186
4.738** (1.666) .206
-20.869
-44.977
**p < .01; *p < .05
-20.28 5.882
Second Model 3 ____________ -.061 -.064 (.052) -.836** -.152 (.290) -.099 -.018 (.292) .072 .013 (.289) .072 .013 (.289) .295* .112 (.149) .406* .140 (.184) -2.06** -.144 (.787) -1.84** -.207 (.494) -1.59** -.276 (.333) -.602 -.054 (.604)
Contact Model 4 _____________ -.072 -.075 (.051) -.802** -.146 (.285) -.123 -.022 (.287) .083 .015 (.279) .068 .012 (.284) .324* .123 (.146) .375* .130 (.181) -1.307 -.091 (.803) -1.90** -.214 (.485) -1.059 -.183 (.362) -.726 -.065 (.594) .227 .095 (.139) .858** .223 (.248)
.234 -50.292 8.75
