The Effectiveness of Short Message Service for Communication With Concerns of Privacy Protection and Conflict Avoidance

Authors


Abstract

Why would people use Short Message Service (SMS) to say something they would not say in person? There is a trend that SMS is becoming more and more popular because it facilitates more extended modes of communications. Using technology acceptance model, we hypothesize the attitude of SMS would be influenced by its perceived effectiveness for communications, perceived ease of use, and subjective norm. Besides the special aspects of communications, conflict avoidance and privacy protection will enforce the impact of perceived effectiveness of SMS for communication. We investigated 953 SMS users and the results support most of our hypotheses. Furthermore, our analyses also show there are differences between females and males on the influence mechanism behind their attitude towards SMS.

The study of communications through instant messaging based on modern platforms such as the Internet and cell phones has gained more and more ground in recent years (Ramirez et al., 2008). A theoretical review of communication through modern real-time devices showed that online deliberations via the Internet-based messaging services are not necessarily inferior to face-to-face or over-the-phone deliberations (Min, 2007). Empirical studies (e.g., Min, 2007; Rice, 1993) comparing face-to-face and Internet-based instant messaging suggested that both online and face-to-face deliberations can increase participants' knowledge, efficacy, and willingness to participate in politics as well as social activities.

There is a trend showing that mobile instant-messaging, usually in the form of short message service (SMS), is becoming very popular among cell phone users, especially young people (Ramirez et al., 2008). SMS is described as the ‘killer’ application of cell phones, based primarily on the fact that its usage has exceeded all expectations (Markett et al., 2006). According to the second annual AP-AOL Instant Messaging Trends Survey in 2007, 25% of respondents sent SMS from their cell phones, including one in three (32%) teens in America (Knowledge Networks, 2007). The situation of using SMS in Asian countries would be more popular. According to UNCTAD (2006), mobile phone subscribers in 2005 were over 122.6 per 100 inhabitants in Hong Kong, indicating that, on average, each person has more than one mobile phone. This observation signals a need for researchers and practitioners to understand how and why SMS, being a kind of instant messaging, attracts users.

In this study, we are interested to explore the attitude towards SMS, which is a learned predisposition or afterthought to respond to a person's positive or negative feelings after using SMS (Fishbein and Ajzen, 1975). Attitude towards SMS instead of behavioral intention to adopt is the focus of this study as we are interested in the postacceptance evaluation of SMS after the users have been using SMS regularly. Besides the perceived effectiveness of SMS for communications, we also include other favorable or unfavorable concerns, such as perceived ease of use and subjective norm, which is proxied by the number of person sending SMS to the respondent in a week, to investigate a person's attitude towards SMS.

By observing different modes of interventions in SMS communications, we also posit that both conflict avoidance and privacy protection would have a significant moderating effect on the relationship between perceived effectiveness of SMS for communications and attitude towards sending SMS via a mobile phone. It is induced by the facts that 43% of instant messaging users said they had used Internet-based instant messaging to deliver a personal message/information that they would not say to someone in person (Knowledge Networks, 2007). This phenomena would be related to the issues such as conflict management (avoiding direct conversation) and privacy protection (making sure that the communication is private and confidential) in an Internet-based platform. Hence, the study of SMS via mobile phone will broaden our knowledge of instant messaging in a mobile platform. In particular, how conflict avoidance and privacy protection will enhance the usage of SMS. The outcome of this study will also offer business practitioners a practical and comprehensive framework in explaining how people choose to use SMS.

Theoretical Framework

Attitude towards SMS

SMS is considered to be a form of new technology and a significant body of research supports the view that the use of any specific new technology is determined by behavioral intentions to use it, whereas behavioral intentions are determined by users' perceptions regarding the system. The Theory of Reasoned Action (TRA) (Fishbein and Ajzen, 1975), the Theory of Planned Behavior (TPB) (Ajzen, 1985), the Technology Acceptance Model (TAM) (Davis, 1989) and UTAUT (Venkatesh et al., 2003) would be the most commonly applied ways of analyzing the behavior and perceptions of SMS users. The validity and explanatory power of these models have been examined across many systems and contexts (King and He, 2006) and SMS appears to fit nicely into these categories (Tung, 2004; Kim et al., 2008).

By using cell phone networks as the platform, SMS is characterized as instant (almost real time in message delivery), cost-effective (very cheap in sending SMS locally and overseas), and ubiquitous communication tool (Markett et al., 2006). People will usually use SMS as a form of succinct and clear communication means (as it is more explicit to have something written to avoid misunderstanding while people are referring to important information such as bank account numbers, telephone numbers, or venue and schedule of a meeting). SMS is also ubiquitous as users can practically send SMS anywhere and anytime they want. Thus, it was found to be able to allow people to alter plans spontaneously (Ling and Yttri, 1999) and to coordinate other forms of communication, e.g., sending an SMS message to propose a subsequent telephone call or e-mail (Eldridge and Grinter, 2001).

In using SMS, young people have become very adept at inventing their own, often private languages for more effective and intimate communications (Segerstad, 2005; Hakkila and Chatfield, 2005). As a result, SMS is used to convey social or emotional content and is becoming more expressive and social in character (Kopomaa, 2000). With these properties, SMS would be an effective communication means. According to TPB, perceived effectiveness of SMS for communication refers to the belief that SMS would be an effective means for communication. This perception of the effectiveness of using SMS for private communications will very likely have a positive impact on the attitude of their users. Thus, we predict the following hypothesis.

H1: The perceived effectiveness of SMS for communications has a positive impact on users' attitude towards SMS.

The perceived ease of use of SMS

A good input design will make SMS usage in a handy manner, thus enhancing the attitude towards SMS. In this study we define perceived ease of use of SMS (PEOU) as user's subjective perception on how easy of using SMS in entering a message or mastering SMS with no effort (Davis, 1989). Right now we can send text messages in different languages such as Arabic, Finnish, Chinese, Japanese, and Thai, the Latin alphabet is widely used by texters around the world. This maximizes the number of characters that can be used, and minimizes the time and effort involved in entering text. Chinese characters, for example, may be difficult to input into a cell phone. Many Chinese texters use pinyin, phonetic Chinese in Latin script, to input text which appears on the screen in Chinese script, but even this may be a painstaking process. One way of circumventing this problem is to choose some readymade messages from menus presented on web sites: Once romantic messages, jokes, or short poems have been downloaded to a cell phone, it is then easy to pass them on. Although this means that many Chinese language messages are widely used and off-the-shelf, the text messages in circulation are often very creative. A good interface may provide different ways for users to input a SMS and this kind of interface would enhance the perceived ease of use of SMS (Branscomb and Thomas, 1985). Furthermore, if the instructions of using SMSs are clear and the management of SMSs is in good manner, users may find SMS an easy tool (Saade and Otrakji, 2007). This perception on how easy to send or manage SMS would influence the attitude towards SMS (Davis, 1989; Parikh and Verma, 2002). So we hypothesize that:

H2: PEOU has a positive effect on the attitude towards SMS.

Subjective Norm

According to TPB, subjective norm which is related to the influence of others towards oneself behavior, is valid in using SMS which is a kind of social communication tool. It is argued that the level of uncertainty declines as individuals move through the stages of the adoption process. Adopters could base their decisions on their own evaluation of a technology, rather than the subjective norm, in the postacceptance stage (Bajaj and Nidumolu, 1998; Malhotra and Galletta, 2005). Nevertheless, SMS being a communication means, is not standalone software and hence subjective norm would be an important concern. If a person receives a lot of SMS from his/her friends, he/she may intend to reply on the message using SMS. This is in agreement with the social influence as in e-learning tools used in Asian context (Kekkonen-Moneta and Moneta, 2002; McNaught and Lam, 2005). Moreover, in the cultural situation of Hong Kong, where people are more collectivist than people in the Western culture, we tend to predict that subjective norm would have a positive impact on the attitude towards SMS. Hence, we posit the following hypothesis.

H3: Subjective norm has a positive effect on the attitude towards SMS.

Conflict Management

We propose that conflict management is an important incentive for people to use SMS for interpersonal communications. Although conflict can sometimes contribute positively to the relationship's creation and stability, unresolved conflict very often contributes to frustration, disaffection, and dissolution (Duck et al., 1991). The inability to manage conflict effectively to reach satisfactory resolution can damage relationships between people and can develop a generally negative view of the entire relationship with regards to its history, meaning, and dynamics, and may eventually result in separation and dissolution of relationships (Gottman, 1993; Leung, 1988). In this regard, conflict management can be viewed as a set of processes to deal with the possible frictions created by differences in planning, promise, competition, and cognitions of people (Canary, 2001; Leung, 1988; Tjosvold and Sun, 2002).

Conflict avoidance is a form of tactics that avoid disagreement or prevent resolution of conflict and to avoid a confrontation with the other party (Kozan, 1997; Ohbuchi et al., 1999). Some people may regard SMS a kind of communication tool for conflict management. When a person sends a SMS to one of his/her friends, it would minimize possible disturbances and in some cases, it is able to avoid direct confrontations. Moreover, the receivers can respond later or will not respond at all. This would prevent face-to-face direct conflicts. Ling and Yttri (1999) and Lasen (2002) studied the usage of cell phones in relation to conflict management. It has been attributable to its asynchronous nature of SMS such that users can wait and think before sending or replying to messages and thus have more control over the response of the receivers (Mitchell et al., 2002). The sender may also find that it is easier to say “no” via messaging service.

Today, SMS has become particularly popular with individuals and in cultures such as Hong Kong which tend to be reserved with other people especially face to face. Teenage boys and girls value SMS as a means to communicate without having to voice feelings and thoughts. The demands of brevity can also encourage sending SMS to be candid, frank, informal, or even cheeky: Ice can be broken for conflict avoidance without the risk of embarrassment (Plant, 2003). Thus, sending SMS for conflict avoidance would be a moderator, but not the independent variable, for SMS as an effective communication means, as it states a condition for SMS to be a favorable communication tool in the mind of a person. In other words, sending SMS for conflict avoidance, being a kind of communications' tactics, would reinforce the impact of communication effectiveness using SMS towards this attitude. Thus, we predict the following hypothesis.

H4: The concern for conflict avoidance has a positive moderation effect on the relationship between communications effectiveness on the attitude towards SMS.

Privacy protection

Due to the intimacy of SMS, it is used for communications in case of privacy issues. Privacy is defined as an individual's ability to determine when, how, and to what extent personal information is disseminated to others (Westin, 1967). According to Communication Privacy Management theory (CPM), the risk of privacy disclosure leads to the need to set up boundaries around public and private information (Clarke, 1999; Petronio, 2002). The theory also proposes that people feel they own their private information and if they should share any of this kind of information, they should still be able to control it within certain boundaries. This implies the kind of communication medium is an important consideration for private information dissemination.

SMS, being an intimate means and unlikely to be overheard by others, would be an ideal channel for communications with privacy concerns. Privacy concern is related to the anxiety on sensitive issues which cannot be overheard by others and just between the sender and the receiver. SMS is intimate because of the private access to a cell phone which is usually a personal property. Users can also easily protect their privacy by imposing a personal password in activating a phone. SMS also helps promote segregated bilateral relationships. Moreover, the impact of cell phone use on environments is much reduced when SMS instead of audio calls are used. A major advantage of SMS lies in the fact that messages can be sent and received in a highly unobtrusive way, even when bystanders are quite close as SMS is a textual message which can only be viewed through the screen of one's cell phone. Privacy protection, being a kind of contagion in communication, would enhance the perception on the effectiveness of SMS for communications. This is for a person who is concerned much about the privacy in communication and when he/she perceives SMS to be an effective communication tool, we postulate that his/her attitude towards SMS would be enhanced. In this regard, we suggest the concerns of privacy protection, being a moderator but an independent variable, would reinforce the relationship between communications effectiveness of using SMS and the attitude towards SMS.

H5: The concern for privacy protection has a positive moderation effect on the relationship between communications effectiveness on the attitude towards SMS.

Overall, Figure 1 depicts our theoretical model which hypothesizes how peoples' perceptions on the effectiveness of SMS in different aspects affect their attitude towards SMS.

Figure 1.

A Research Model of Attitude Towards SMS

Research Methodology

Sampling Procedure

The questionnaire survey was taken over three different periods (a pilot study in April 2008 in Hong Kong, an extensive survey in August 2008 in Hong Kong and a third extensive survey in October 2008 in Shanghai). The target population consists of those people who have experience in sending SMSs at least three times a week in their current practice. This is to ensure the respondents are experienced users of SMS. In this regard, our research assistant will ask the potential respondents about their frequency of using SMS in order to ensure they belong to our target group before putting forward the questionnaire. The response rate was around 10.0% in Hong Kong and 8.0% in Shanghai. That is 1 person in 10 was willing to answer the questionnaire when an interviewer approached to them in Hong Kong and 1 person in 8 answered the questionnaire in the Shanghai survey. The questionnaires were distributed randomly on the street. In most of the cases, the survey was conducted by face-to-face interview. We adopted this approach for three reasons. First, a street survey can reach different kinds of respondents with different ages, job natures and education level. This will make our result more representative. Second, a street survey by interviewing is also a way to keep the response rate at a higher level as people will regard our survey to be a sincerer one. Third, we can conduct a face-to-face interview on the questionnaire items and the respective responses are reliable.

In order to gather more returns and to show our sincerity in conducting the survey, a small gift of around HK$5 or RMB$4.5 (e.g., a pen, a calendar, etc.) was given for every completed survey. We also explained to the respondents the purposes of the survey so as to motivate them to fill in the questionnaires on the spot. The confidentiality of the results was stressed. To minimize data entry errors, all the collected data were checked for consistency. All the duplicate responses and all the responses that had too many missing values were removed. In total, 31 questionnaires were void. As a result, 953 valid responses were collected on the three surveys. The numbers of responses on the first pilot survey in Hong Kong, the second extensive survey in Hong Kong and the third survey in Shanghai were 183, 558, and 210, respectively. The profile of the respondents in the surveys is shown in Table 1, which fits nicely with the general populations in the Asian region that send at least three SMSs in a week. From Table 1, the mode statistics show that the regular users of SMS are young in age 18–25 (634, 66.7%), graduate (447, 47.0%), and receive SMSs from four to six persons in a week (268, 28.1%).

Table 1.  Descriptive Statistics of the Respondents
Gender:  Male (456, 47.9%), Female (475, 49.9%)Age:  18–25 (634, 66.7%), 26–30 (79, 8.3%), 31–40 (176, 18.5%), 41–50 (19, 2.0%), 51 or above (5, .5%)
Education: Secondary School (140, 14.7%), Diploma/Higher diploma (100,10.5%), Graduate (447, 47.0%), Post graduate (250, 26.3%)
Average number of persons sending SMS to the respondent in a week: 0 (35, 3.7%), 1–3 (240, 25.2.0%), 4–6 (268, 28.1%), 7–9 (147, 15.4%), 10–12 (98, 10.3%), 13–15 (39, 4.1%), 16 or above (39, 4.1%)

Instrument Development

As shown in Table 2, for the construct measuring the extent of perceived effectiveness of using SMS for communications (PEffSMS4Comm), we adopt the study from Zack (1994). It consists of items like, “I think communicating via messaging service is very effective.” The construct perceived ease of use (PEOU) is modified from Davis (1989) and items such as “I think learning to input a message is easy.” For the subjective norm, we suggest its proxy variable—the average number of persons sending SMS to the individual in a week (PeopleTxSMS2Resp) to reflect the influence of subjective norm on the attitude towards SMS. It is measured using a rating of 1 to 7 with 1 standing for zero, 2 standing for one to three persons, 3 standing for four to six persons, 4 standing for seven to nine persons, 5 standing for 10 to 12 persons, 6 standing for 13 to 15 persons, and 7 standing for 16 or more persons. This scale is adopted from our interview with those Chinese users who use SMS as their daily practice. The attitude towards SMS (AttSMS) is measured from Ling and Yttri (1999) which consists of items such as “I think messaging service is important to me”

Table 2.  Descriptive Statistics and Reliability
VariablesMean*Std. Dev.
  1. *1—strongly disagree and 7—strongly agree.

Attitude towards SMS (Cronbach's alpha = 0.880)5.231.386
 I think messaging service is important to me5.161.576
 I would accept using messaging service5.611.388
 Messaging service is a part of my lifestyle4.911.672
Perceived Effectiveness of SMS for Communications (Cronbach's alpha = 0.894)4.871.125
 I think communicating via messaging service is very beneficial4.981.269
 I think communicating via messaging service is very effective4.791.328
 I think communicating via messaging service is useful5.001.253
 I get what I want out of the communication via messaging service4.691.321
PEOU (Cronbach's alpha = 0.891)5.561.265
 I think learning to input a message is easy5.681.407
 The input system of messaging service is easy to use5.491.428
 I think using messaging service is easy5.501.352
Conflict Avoidance (Cronbach's alpha = 0.772)4.921.326
 I generally avoid an argument5.111.429
 I usually avoid open discussion of my differences with the other person4.721.509
Privacy Protection (Cronbach's alpha = 0.814)5.291.193
 I am concerned about threats to my personal privacy5.261.412
 I want to be left alone with my thoughts5.331.358
 I am concerned about improper invasion of privacy5.281.422

For the moderators, the construct measuring the concern for conflict avoidance (ConflictAvd) is adopted from Morris et al. (1998), Ling and Yttri (1999), and Lasen (2002). It includes items such as “I generally avoid an argument.” For the construct measuring the concern for privacy protection (PrivacyProt), it is referred from Milberg et al. (1995), Ling and Yttri (1999), and Vidmar and Flaherty (2006). The item “I am concerned about threats to my personal privacy.” is adopted. All of the above constructs are assessed by a 7-point scale (1 = highly disagree, to 7 = highly agree).

Prior studies have reported that individual IS perceptions and behaviors may differ across such personal factors as gender, education, and age (e.g., Agarwal and Prasad, 1999; Frankel, 1990; Gefen and Straub, 1997; Venkatesh et al., 2003; Venkatesh and Morris, 2000). These factors are therefore controlled in the study. Gender is coded with 0 for “male” and 1 for “female.” Age is coded with 1 for “18 to 25,” and 5 for “51 or above.” Similarly, education is coded with 1 for secondary school and 4 for postgraduate.

Analyses, Findings, and Discussions

Data Analyses

We obtained the means, standard deviations, and bivariate correlations for all data used to analyze predictions of all variables. We have checked for multivariate normality of the data which are fairly normally distributed.

Validity of Data

We have tested sample bias by comparing key constructs from earlier (the first survey in Hong Kong) and later respondents (the later extensive surveys in Hong Kong and Shanghai) using the Kolmogorov-Smirnov two-sample test (Siegel and Castellan, 1988). This test assesses whether significant differences exist in the distribution of respondents and nonrespondents for a given variable, including differences in central tendency, dispersion, skewness, and so forth. The results showed that the data of the pilot survey and the later extensive surveys in Hong Kong and Shanghai are fairly equal. We also tested for the difference between Hong Kong and Shanghai on the means and distribution of the measured constructs. The results showed that there was not much difference. Although in practice, Shanghainese use SMS much more than Hong Kongers, their perception on how the effectiveness of SMS towards communications, privacy protection, and conflict management is not much different from that of Hong Kongers. Thus, we aggregated the three surveys into a data set for subsequent analyses.

Instrument Validity

Reliability refers to the extent to which a construct is free from errors and yields consistent results. Cronbach's alpha was used to measure the internal consistency of the multi-item scales used in this study. As the Cronbach's alpha values of all of the constructs were over 0.7, it can be claimed that they are all reliable. Moreover, as all of the measures of the constructs have been used in past studies, and the questionnaire was validated by experts in the fields of IT and behavioral science before it was administered, the content validity of all the constructs can be deemed to be acceptable.

Convergent validity of the measurement scales was evaluated using the two criteria suggested by Fornell and Larcker (1981), namely (1) all the indicator factor loadings should be significant and exceed 0.70, and (2) the average variance extracted (AVE) for each construct should exceed the variance due to measurement errors for that construct (i.e., should exceed 0.5). The factor loadings of the 15 items are shown in Table 3. All items exhibited a loading value higher than 0.7 on their respective constructs. Thus, acceptable item convergence on the intended constructs was achieved. The AVEs of the constructs, as shown in Table 3, were all greater than the variances due to measurement errors. Hence, both conditions for convergent validity were amply fulfilled.

Table 3.  Factor Analysis
 12345
  1. Extraction Method: Principal Component Analysis.

  2. Rotation Method: Varimax with Kaiser Normalization.

  3. a Rotation converged in 5 iterations.

AttSMS1.831.290.247.088.082
AttSMS2.755.154.392.078.137
AttSMS3.833.283.244.112.065
PEffSMS4Com1.186.829.141.154.082
PEffSMS4Com2.136.884.132.114.088
PEffSMS4Com3.188.851.140.059.046
PEffSMS4Com4.179.799.099.126.055
PEOU1.231.147.865.080.099
PEOU2.212.130.882.074.100
PEOU3.300.191.795.094.067
ConflictAvd1.099.173.141.840.219
ConflictAvd2.098.166.058.885.126
PrivacyProt1.095.063.005.110.853
PrivacyProt2.080.066.121.094.821
PrivacyProt3.034.075.104.133.817

The correlation matrix of the data set is shown in Table 4. This enables us to examine all potentially overlapping constructs. If the items comprising a construct do not overlap much with other constructs (i.e., the AVE of a construct is larger than its squared intercorrelations with other constructs), then discriminant validity of the construct is assured (Fornell and Larcker, 1981). Table 4 shows that the diagonal elements (reporting the square root of the variance shared between a construct and its measures) are all higher than the correlations between target constructs without exceptions, which suggest discriminant validity of all the constructs in this study.

Table 4.  Correlation Matrix
 AVE12345
  1. **Correlation is significant at the 0.01 level (2-tailed).

1. AttSMS.651.807    
2. PEffSMS4Comm.708.507**.841   
3. PEOU.719.607**.369**.848  
4. ConflictAvd.744.277**.243**.212**.863 
5. PrivacyProt.690.223**.179**.236**.336**.830

Judging from the standard deviations of all the items, the sampled data had enough variations to represent the population. The means of the items, as shown in Table 2, suggest that the respondents had positive attitude towards SMS as a whole (mean = 5.23). They perceive SMS is effective for communications (mean = 4.86) and easy to use (mean = 5.56). Their concerns for conflict avoidance and privacy protection are high with means of 4.92 and 5.29 respectively.

To check the existence of common method bias, we conducted the Harmon one-factor analysis suggested by Podsakoff and Organ (1986). A factor analysis combining every variable in the research framework did not detect a single factor explaining the majority of covariance. In addition, the results of the regression analysis showed different degrees of significance for the regression coefficients. The above evidence collectively suggested that common method bias is not a serious concern in this study.

Test of hypotheses

The results of the regression analysis are presented in Table 5. The R2 of attitude towards SMS for step one (just the control variables), step two (including the main effect) and step three (including the main effect and two way interactions) are .125, .538 and (.546 with interaction of PEffSMS4Comm*ConflictAvd or .551 with interaction of PEffSMS4Comm*ConflictAvd), respectively. For the interaction effect, the incremental R2 is a significant increment according to the F-test between step 2 with the direct effect and steps 3b or 3b with the interaction effects.

Table 5.  Regression Analyses on the Attitude Towards SMS
 Step 1 Control VariablesStep 2 Main EffectsStep 3a Two-way interactionsStep 3b Two-way interactions
Control Variables
 Gender.196***.122***.126***.123***
 Age−.276***−.085**−.084**−.089***
 Education.184***.086***.085**.090***
Main Effects
 PEffSMS4Comm .270***.254***.268***
 PEOU .444***.434***.419***
 NumPersonTxSMS2Resp .170***.167***.176***
Interaction Effects
 PEffSMS4Comm*ConflictAvd  .102*** 
 PEffSMS4Comm*PrivacyProt   .066*
Model Information
 R2.125.538.551.546
 ΔR2 from previous step .413.013.008

Of the control variables, they have significant impacts on the AttSMS. Our findings showed that SMS are especially favorable for females (β = .122***), for young people (β = −.085**), and for highly educated people (β = .086***). It is reasonable that young people are more acceptable to information and communication technologies (ICT) than the elderly. On the other hand, more educated people are expressive in their written communications using SMS.

Our findings, as indicated in Table 5, showed that the PEffSMS4Comm is a significant factor to predict AttSMS. According to the first-order effects shown in Step 2, PEffSMS4Comm was positively associated with AttSMS and thus hypothesis 1 is supported. Moreover, PEOU and NumPersonTxSMS2Resp were also positively associated with AttSMS and thus hypothesis 2 and 3 are supported. The R2, being .538, is increased by .413 which is a big jump from the explained variance purely from the control variables. This indicated that the PEffSMS4Comm, PEOU and NumPersonTxSMS2Resp are dominant factors to explain AttSMS. Comparing their influence, PEOU (β = .419***) is of greater impact on AttSMS than that of PEffSMS4Comm (β = .268***) and NumPersonTxSMS2Resp (β = .170***). Perhaps because people approach the cell phone with more focus on its interface when entering a text for the SMS, thus whether SMS would be a good communication tool is their second concern. It is reasonable as Chinese character is rather difficult to be handled especially in a cell phone setting with just having a numeric pad. For the effect of social influence, a person receiving a lot of SMS with high level of NumPersonTxSMS2Resp (β = .170***) will have a significant impact on the attitude of SMS. That is, when a person is influenced by many others sending him/her SMS, he/she would have a better attitude towards SMS.

Hypothesis 4 predicts the positive relationship of PEffSMS4Comm and AttSMS may be moderated by ConflictAvd such that the relationship would be reinforced for high level of ConflictAvd. In order to eliminate the possible covariance of the two interaction effects, we test their impact on AttSMS separately using steps 3a and 3b. According to step 3a of Table 5, our finding with the interaction effect be significant (β = .102***), shows this hypothesis is supported. Thus, when a person is concerned with conflict avoidance and he/she perceives SMS to be an effective communication means, he/she will have good attitude towards SMS.

For hypothesis 5, it hypothesizes that the positive relationship of PEffSMS4Comm and AttSMS may be moderated by PrivacyProt such that the relationship would be reinforced for high level of PrivacyProt. Our analysis, as shown in step 3b of Table 5, indicates there is significant two-way interaction between PEffSMS4Comm and PrivacyProt (β = .066*). Hence, hypothesis 5 is supported. The privacy protection of information would be a sensitive issue especially for some personal data. In order to prevent others overhear about the information, a person will use SMS for interpersonal communication once he/she perceives SMS to be an effective communication means.

To explore the form of the interactions, as shown in Figure 2, we plot the relationship between PEffSMS4Comm and AttSMS for both high and low levels of ConflictAvd. Following the procedures described by Aiken and West (1991), we add one standard deviation to the ConflictAvd. Then we perform the regression analysis. The positive relationship of PEffSMS4Comm and AttSMS is significant when ConflictAvd is at low levels. Similarly we test the slope of the relationship of PEffSMS4Comm and AttSMS for high levels of ConflictAvd by subtracting 1 standard deviation to the ConflictAvd. We found that the slope is also significant when the level of ConflictAvd is high. Accordingly, the impact of PEffSMS4Comm on AttSMS is higher when ConflictAvd is at high level as compared with ConflictAvd at low level. Hence, the moderation effect of ConflictAvd can shape the relationship of PEffSMS4Comm and AttSMS.

Figure 2.

Attitude Towards SMS Predicted by Using Two-way Interactions Between Perceived Effectiveness of SMS for Communications and the Concern for Conflict Avoidance
Note: Both slopes are significant.

On the other hand, the positive influence of PEffSMS4Comm on AttSMS is significant with respect to different levels of PrivacyProt. As shown in Figure 3, when PrivacyProt is at a low level, the impact of PEffSMS4Comm on AttSMS is mild but still significant according to the slope test from Aiken and West (1991). The relationship becomes more salient when PrivacyProt is at a high level. Comparing the moderation effects of PrivacyProt and ConflictAvd, the later one is a stronger moderator. This can be observed from the β values (0.066 versus 0.102) in Table 5. In sum, Figures 2, and 3 depict the results in different situations.

Figure 3.

Attitude Towards SMS Predicted by Using Two-way Interactions Between Perceived Effectiveness of SMS for Communications and the Concern for Privacy Protection
Note: Both slopes are significant.

Posthoc Analysis

Besides age and education which would be well known factors that affect attitude towards communication technologies such as SMS, our results also indicated gender is a significant factor (see Table 5). The effect of gender on attitude towards SMS would also be another interesting area to explore. According to Knowledge Networks (2007), nearly half of teenage female respondents surveyed had used Internet-based instant messaging to say something they would not say in person, compared with just over a third of teenage male respondents. This raised an interesting question—would gender in the Asian region be an important factor in understanding the attitude towards SMS? Our result shows that gender is also a significant factor on AttSMS and females have a more favorable attitude towards SMS than that of males. Hence, a follow-up analysis was conducted to investigate further on the issue: the effect of gender on AttSMS. Here we divide the data into two different groups: male (group size = 456), and female (group size = 475).

The regression models predicting AttSMS were estimated using AMOS 7.0, which provides maximum likelihood estimates (AMOS Development Corporation, 2007). The main purpose for using AMOS was that it more easily permits statistical tests of differences in a single model for two (or more) samples. When constraining a path to be equal in the two samples and simultaneously estimating parameters, one degree of freedom is freed, thereby providing a χ2 test of their equivalence.

Comparing to the regression results as shown in Table 6a, the AMOS test shows that the moderation effect of ConflictAvd is stronger for the group of females than that of males with the β values 0.153 and 0.097 respectively. In fact, the moderation effects in both cases are all significant but with different magnitudes. This would be because females are concerned more about social relationship than males. Gender schema theory suggests that such differences stem from gender roles and socialization processes reinforced from birth rather than biological gender (Bem, 1981; Bem and Allen, 1974; Kirchmeyer, 1997; Lynott and McCandless, 2000). Thus, females have stronger desire to manage conflict than males.

Table 6a.  Regression Analyses on the Effect of Gender Toward the Attitude of SMS with the Moderation of Conflict Avoidance
 MaleFemaleDifferences
Control Variables
 Age−.074***−.092*.018
 Education.121**.047.074*
Main Effects
 PEffSMS4Comm.276***.227***.049
 PEOU.372***.524***−.152**
 NumPersonTxSMS2Resp.177***.156***.021
Interaction Effects
 PEffSMS4Comm*ConflictAvd.097*.153***−.056*
Model Information
 R2.495.594−.099*

From the moderation of PrivacyProt as shown in Table 6b, females, with the two way interaction effect being significant with β value 0.116, are more eager to protect personal privacy than males, with its two way interaction effect being insignificant. This would be due to females, when compared with males, are more sensible about privacy issues. This finding is consistent with previous gender studies on privacy protection that female counterparts are more self-regulated to protect personal information privacy (Kuo et al., 2007). Moreover, females would be more risk averse than males. This will make females more anxious about the way in handling privacy information (Jianakoplos and Bernasek, 1998).

Table 6b.  Regression Analyses on the Effect of Gender Toward the Attitude of SMS with the Moderation of Privacy Protection
 MaleFemaleDifferences
Control Variables
 Age−.071−.102**.031
 Education.122**.053.069*
Main Effects
 PEffSMS4Comm.286***.242***.044
 PEOU.366***.505***−.139**
 NumPersonTxSMS2Resp.192***.161***.031
Interaction Effects
 PEffSMS4Comm*PrivacyProt.026.116**−.090*
Model Information
 R2.499.591−.092*

In sum, from our findings on both the moderation effects of ConflictAvd and PrivacyProt are significant and of larger values for females when compared with males (see Tables 6a and 6b). The AMOS tests, as indicated in Tables 6a and 6b, also show both the moderation effects of ConflictAvd and PrivacyProt have significant differences between the two gender groups. Hence, this would be consistent with the arguments that females are concerned more about their conflicts with others as well as their privacy. Contributing to the literature, this study demonstrates that the attitude towards SMS is based on the effectiveness of SMS for communications and in particular reinforced by individual concerns on SMS's effectiveness on conflict avoidance and privacy protection. Moreover, its effect will be different for different genders.

Implications on commercial practice

From this study, the concern for conflict avoidance would enhance the perceived effectiveness of SMS for communications. This is especially valid for female users. In practice, SMS is getting more common as an effective medium to say “no” without much embarrassment or conflict. In order to encourage people to use SMS for this purpose, the service provider or mobile phone manufacturer may consider setting up their phones with some preformatted wordings or phrases that help users to express themselves. Moreover, some cell phone would memorize the words or phrases of a user that he/she is used to type when sending an SMS. Thus, in typing a new SMS, the users may save their effort with the hints of possible wordings in the cell phone's memory.

According to our findings, the influence of the concern for privacy protection is less than that of SMS for conflict avoidance. There would be some reasons behind this. First, in the current practice, there is no sophisticated encryption associated with sending an SMS. Thus, the SMS may still be trapped or overheard by others in the mobile phone network. In this regard, it is nice if the mobile service provider would allow users to encrypt their SMSs with some personal passwords before sending them. This practice would raise the security level and encourage more people to use SMS for privacy information transfer. This also fits into the growing concern about privacy information of consumers (Caudill and Murphy, 2000; Franzak et al., 2001).

In this study, we found that females are more sensible when managing conflicts and are more concerned about their privacy. Sending SMS would be a more useful tool for them to resolve the conflict and privacy issues. Hence the mobile service provider would consider having some kind of package associated with a certain amount of air time as well as extensive amount of SMS so as to attract more potential female customers.

Limitations and Conclusion

As is the case with all empirical research, this investigation too has several limitations. A notable weakness lies in the cross-sectional research design, where all measurement items were collected at the same point of time from the students' perspective. Given that the investigated constructs are not supposed to remain unchanged over time, this research method may not fully capture the dynamics of the attitude towards SMS. This constraint may also lead to the problem of same source bias, which is fortunately not a serious concern as confirmed in our analysis. To address the above issues, future research should consider employing multimethods and longitudinal research designs. A longitudinal study combining qualitative and quantitative data would enable a process-oriented perspective that cannot be achieved by using a variance-based approach, such as the one employed here.

People communicate in numerous ways: face to face interactions, talking over a phone, letters, e-mails, blog, SMS, etc. Each communication means may have its unique characteristics and may be more favorable than the others in some situations. We argue that SMS, being more and more popular due to its unique characteristics such as mobility, asynchrony, and intimacy, facilitates people to communicate ubiquitously. From time to time, it would be used for conflict avoidance and privacy protection. Our findings have validated the general attitude towards SMS will be mainly affected by people's perception of its effectiveness for communication, its perceived ease of use, and the social influence. The effect of SMS effectiveness for communication towards the attitude of SMS would be enhanced by the concerns of using SMS for conflict avoidance as well as for privacy protection.

Acknowledgements

We are grateful to two anonymous referees for their constructive comments on an earlier version of this paper. This research was supported in part by The Hong Kong Polytechnic University under grant number P6GE.

About the Authors

Vincent Cho is an associate Professor in the Department of Management and Marketing, The Hong Kong Polytechnic University. He obtained his Ph.D. from the Hong Kong University of Science and Technology. His teaching interests are MIS, e-commerce and IT strategies. His research interests lie with technology adoption, data mining, and expert systems. Recently, he has involved with various consultancy projects relating to software vendor partnership, logistics in supply chain management, and demand forecasting.

Address: Department of Management and Marketing, The Hong Kong Polytechnic University, Hung Hom, Hong Kong.

Humphry Hung is currently a senior teaching fellow of the Department of Management and Marketing in the Hong Kong Polytechnic University. He obtained his Ph.D. from the City University of Hong Kong. His research interests include mobile communication, strategic management, creativity and innovation management, and entrepreneurship. Dr. Hung has published in several international refereed journals and conference proceedings.

Address: Department of Management and Marketing, The Hong Kong Polytechnic University, Hung Hom, Hong Kong.

Ancillary