Magoosh GRE

Internet usage on a mobile phone or other forms of interaction with a hand held cell phone while driving is not only a serious offence in most developed countries but has been banned in most countries

| December 19, 2012

Introduction

Literature review

Talking on a cell phone, internet usage on a mobile phone or other forms of interaction with a hand held cell phone while driving is not only a serious offence in most developed countries but has been banned in most countries because it contributes serious distraction and danger to the drivers and other road users (WHO.2011). In spite of the serious dangers this behaviour poses to the public; evidence shows that many drivers are yet using cell phones while driving. For example, a survey conducted by Nationwide shows that the usage of a cell phone while driving increases a drivers distraction comparative to alcohol consumption beyond the legal limit of .08. The survey result also shows that cell phone usage is the primary source of distraction while driving; that users of cell phone while driving are four times more prone to get into serious fatal accidents; that driving while distracted contribute about 25 percent of reported accident by police among other factors (see www.nationwide.com).

This report was based on analysis of a study conducted for gaining  better understanding of the role as well as relative impact of the number of factors on the formation the drivers’ opinions of using a mobile phone whilst driving.  The Questionnaires  distributed to the large sample of part-time students who are studying at a UK university during the class time. This report is an attempt to conduct a quantitative analysis and interpretation of the study conducted to determine  reliability, generalizability and internal consistency of  variables utilised in  survey and the most important factor in explaining drivers’ behaviour towards using cell phone while driving. The study presumed an assumption that the theory of planned behaviour (TPB) is valid to gauge the role and impact of attitude, subjective norm and perceived behavioural control and past behavior on drivers’ opinion to use a hand held mobile phone while driving. The next section is a critical review of literatures on the theory of planned behaviour theoretical model used in the study. The second and third section presents a report of data descriptions, interpretation of the research and summary of findings in order to determine the internal consistency of the variables employed and their relative relationship and contribution to the model.

Literatures advocate that the theory of planned behavior was an expansion of the theory of reasoned action, which propose that behavioural intentions predict behaviour better than attitudes. The theory of planned behaviour was proposed by Ajzen (1975) to eliminate the limitation observed in  theory of reasoned action by adding a third variable perceived behavioural control to  TRA, later researchers in various fields have added other constructs to the model. Theory of planned behaviour is applied virtually into the every social and management sector from health to education, consumer behaviour and technology; to examine the relationship between intention, attitude, subjective norm and perceived behavioural control. Theory is specifically applicable for predict, identify, understand and explain human influences and motivation; hence its applicability to practically every social and management field. (George 2004; Mathielson, 1991; Taylor and Todd, 1995).  The theory of planned behaviour is constituted by three primary variables Figure 1 shows  components of  TPB(See Mathielson, 1991; Taylor and Todd, 1995; Kim and Han 2010, Paul and Mendel, 2006).  However, several other variables not necessarily mentioned in  report are said to contribute for predicting intention as  results and conclusion has shown.

Figure 1 TPB variables (Ajzen 1991)

TPB variables

 1.  The TPB primary components

Attitude is the number one primary variable among  TPB constituents (ATT), its often regarded as the level of the usefulness of  psychological object to the individual which determines  negative or positive behaviour of  individual towards object –of the products (Fishbein and Ajzen, 1975, 2000).  Some literature suggest that attitude is the most A number of studies which found attitude to be  key element in predicting intention in the various context. (See Gopi and Ramayah (2007), Kim and Han (2010), Sang and Chen (2010). Thus, it is expected that driver’s attitude towards the harmful effect of using a cell phone while driving will be positively related

Subjective norm (SN) in contrast is suggested to demonstrate individual perceived shared or societal demands to perform certain behaviour. Thus, it may be inferred that the general opinion of other drivers’ towards driving and using mobile phones will influence individual opinion or intention to perform this behaviour although it is negative or illegal. Literatures suggest that the subjective norm has  significant effect on the behavioural intention and the intention (e.g. Hillhouse et al., 2000; Kalafatis et al., 1999; Karahanna et al. 1999). Therefore, a positive correlation between SN and intentions or opinion of drivers to use a cell phone while driving is expected. On the other hand, perceived behavioural control could be regarded as the degree to which it is possible to perform behaviour or a given action (for example drivers’ the attitude and the intention to driving and mobile usage while driving).

PBC is one of the primary elements of the theory of planned behaviour and a co-determinant of the intention along with the attitude and the subjective norm. (Ajzen, 1991; Mathielson, 1991; Taylor and Todd, 1995).Ajzen (1991) suggest that the PBC has only one construct however, critics says that it has two constructs: self efficacy and controllability; casting doubts on the nature and measurement of PBC, doubting Ajzen’s (1991) hypothesis that PBC has only one construct rather than  critic suggestion (Paul and Mendel, 2006). However, it was  suggested that  the possibility of this two construct does not invalidate  original unitary assertion (Taylor and Todd, 1995; Kim and Han 2010). Thus, we propose  null hypothesis that PBC will  positively correlated with drivers’ intention for using  mobile phone while driving; or PBC will not be positively correlated with drivers’ intention to use a mobile phone while driving.

Table 1: TPB main constructs (Ajzen, 2000)

Behaviour:

 

It is  transmission of intention or perceived behavioural control into the action.
Behavioural Intention: It is an indication of how hard people who are willing for trying and of how much  effort they are planning for exerting, in order to performing behaviour. Influenced by three components: person’s attitudetoward performing the behaviour;  perceived social pressure; called subjective norm and perceived behavioural control.

 

Attitude: It is the first determinant of behavioural intention. It is  degree to which  person has a favourable or unfavourable evaluation of the behaviour in question.

 

Subjective Norm: It is considered the second predictor of behavioural intention. This is an influence of social pressure that is perceived by the individual (normative beliefs)for performing or not performing the certain behaviour. This weighted by the individual’s motivation to comply with those perceived expectations (motivation to comply).

 

Perceived Behavioural Control: Antecedent of behavioural intention. This construct is defined as the individual’s belief concerning how easy or difficult to perform  behaviour will be. It often reflects actual behavioural control

2   Analysis and result 

This section is an analysis of the information on factors that influence and affect drivers’ intention to use a mobile phone while driving based on a study conducted  UK. The data collected will be analyzed and interpreted to explain the properties of the measurement scales and the items that constitute the survey conducting a reliability analysis to show the relationship of individual items in the scale (Morris, 2002).

In general, we want to find out whether the survey measure drivers opinion in a valid and sufficient way? Using reliability analysis, we want to deduce the extent to which the items in the questionnaire are related to each other, in order to deduce an overall index of the repeatability or internal consistency of the scale as a whole, and identify problem items that should be excluded from the scale. The proceeding section identifies significant correlation and regression to determine the predictive ability values of intention on attitude, subjective norm, perceived behavioural control and past behavior. Behavioural intention is suggested to be the most important predictor of the behaviour (Teo and Lee (2010). Apparently, this may not be applicable in all the cases.

2.1   Sample profile statistics and results

The respondents comprises of participants from a UK university with no missing data for all variables. The descriptive statistics does not contain demographic information or frequency of usage.

2.2 Data analysis reliability measures

This study utilize Likert-type scales to report overall scale and subscale internal consistency reliability estimates in the analysis of the individual respondent to a survey scale items. The study which this report presents its findings used 18 multiple likert items based on the constructs of the theory of planned behaviour to determine drivers’ opinion to using mobile phone while driving. Each question seeks a response for example how “very likely” to very unlikely” in order to understand whether the respondents to this questionnaire agree to presumed harmful effect of using a mobile phone while driving.  In order to measure the reliability analysis for each of the TPB constructs Cronbach’s alpha is most commonly used to measure internal consistency of multiple likert scale.  Literature suggest that a scale to be reliable must contain multiple items to measure a quantitative response and respondents are asked to rate each items on a scale of response which reflects their best response to the item (Nunnally and Bernstein (1994), and Spector, 1992). It is noted that past behaviour was combined in the SPSS output wit intention since cronbach alpha cannot measure a single item. The proceeding shows the item-analysis output from SPSS for the multi-item scale of student opinion towards their using a mobile phone while driving. The Cronbach’s alpha is 0.495, which indicates a low level of internal consistency for the scales items used for this study as shown in table 2.1.

Table 2.1 Reliability Statistics
Cronbach’s Alpha Cronbach’s Alpha Based on Standardized Items N of Items
.495 .465 18

The Item-Total Statistics table 2.2 presents the Cronbach’s Alpha if item deleted in the final column, as shown below:

Table 2.2 Item-Total Statistics
  Scale Mean if Item Deleted Scale Variance if Item Deleted Corrected Item-Total Correlation Squared Multiple Correlation Cronbach’s Alpha if Item Deleted
Int1 57.65 73.085 .427 .717 .413
Int2 58.37 68.352 .477 .797 .387
att1 60.80 97.597 -.321 .690 .556
att2 61.17 90.661 -.060 .686 .515
att3 61.24 90.979 -.068 .625 .514
att4 60.14 104.081 -.461 .601 .599
att5 60.23 95.106 -.217 .450 .552
att6 61.59 89.215 .032 .566 .499
sn1 57.36 85.957 .100 .279 .491
sn2 58.88 72.326 .480 .574 .402
sn3 58.46 75.902 .365 .562 .432
sn4 59.37 77.453 .230 .426 .463
sn5 57.14 84.501 .178 .335 .478
pbc1 59.55 70.365 .451 .687 .400
pbc2 60.65 75.230 .398 .586 .425
pbc3 57.17 80.869 .250 .236 .462
pbc4 61.50 82.774 .212 .242 .471
Past Behaviour 61.92 87.334 .438 .386 .476

This column presents the value that Cronbach’s alpha would be if that particular item was deleted from the scale. It suggests that removal of any question except att4, would result in a lower Cronbach’s alpha. Therefore, it will not give us a better result even if att4 is delete from these questions. Removal of question (att4) would lead to a small improvement in Cronbach’s alpha and we can also see that the Corrected Item-Total Correlation value was low (-0.461) for this item. This might lead us to consider whether we should remove this item. Appendix B suggest that if all attitude question are deleted the cronbach alpha will increase to 0.657 which is a much better result. Table 2.2 column 4  total correlation result for all items shows there is  positive correlation between intention and Past behavior (.438; SN-item 2 (.480); and pbc – item 1 (.451) but a weak and the negative correlation between intention and attitude (-.461).

Table 2.3 Reliability Statistics
Cronbach’s Alpha Cronbach’s Alpha Based on Standardized Items N of Items
.599 .548 17

Table 2.3 shows that a removal of question (att4) increases the cronbach alpha 0.495 to 0.599 and also increases the individual items internal consistency cronbach alpha.  Table 2.4 shows the measure of spread around the mean (63.13) with a dispersion SD: 9.531.

Table 2.4 Scale Statistics
Mean Variance Std. Deviation N of Items
63.13 90.838 9.531 18

Cronbach’s alpha reliability the coefficient normally ranges between 0 and 1. However, there is actually no lower limit to coefficient. Closer Cronbach’s alpha coefficient is to 1.0 ,greater the internal consistency of the items in the scale. Based upon the  formula _ = rk / [1 + (k -1) r] where k is  number of items which is considered and r is  mean of the inter-item correlations the size of alpha is determined by both  number of items in  scale and  mean inter-item of correlations. George and Mallery (2003) provide  following rules of thethumb: “_ > .9 – Excellent, _ > .8 – Good, _ > .7 – Acceptable, _ > .6 – Questionable, _ > .5 – Poor, and _ < .5 – Unacceptable” (p. 231). While increasing  value of the alpha is partially dependent upon  the number of items in the scale, it should be noted that this has a diminishing returns. It should also be noted that an alpha of .8 is probably a reasonable goal some studies report alpha of .7 as a suitable. George & Mallery (2003) also noted that while a high value for Cronbach’s alpha indicates good internal consistency of the items in the scale, it does not mean that  scale is undimensional explaining that this factor analysis which is not dealt with in this study can be useful.

3. Relationship between the variables

3. 1  Regression Analysis

Regression was used for exploring  relationship between intention (dependent variable) and the independent variable of the attitude, SN, PBC, and past behavior the items are supposed to be averaged. According to Cohen (2007) linear regression analyzes is most appropriate to evaluate relationship between two or more predictor variables and the dependent variable (intention). Table 3.1 provides the R and R2 value. The R value is attitude 0.671; subjective norm 0.702; and pbc 0.788 respectively for which represents the relative contribution of each of the variable items in the scale with pbc contributing more in the regression model. Thus, it means that the percentage of variation in intention is explained by the independent variables of attitude, subjective norm and perceived behavioural control. The R2 value indicates how much of the dependent variable, intention, can be explained by the independent variables. In this case, 57.5% can be explained, which is relatively above average. Representing 42.5% of intention not explained by the TPB variables which are not included in the model and therefore other factors may contribute significantly to variation in intention which is not explained. It also means that about that percent of variance in intention variable can be explained by the regression model. The F-test result was 10.391 with significance (“Sig”) of .000. This meant that the probability of this results occurring by chance was less than 0.05. Thus, a significant relationship was present between intention of students and actual behaviour of using a mobile phone while driving.

Table 3.1 Model Summaryd
Model R R Square Adjusted R Square Std. Error of the Estimate Change Statistics
R Square Change F Change df1 df2 Sig. F Change
1 .671a .450 .425 1.685 .450 18.018 6 132 .000
2 .702b .493 .449 1.649 .043 2.159 5 127 .063
3 .788c .621 .575 1.449 .128 10.391 4 123 .000
a. Predictors: (Constant), att.avg
b. Predictors: (Constant),  att.avg, sn.avg
c. Predictors: (Constant), att.avg, sn.avg, pbc.avg
d. Dependent Variable: int.avg

Table 3.2 below tells us whether or not the regression result is meaningful using the last column (p-value at .05). The ANOVA result is less than .005 meaning that the regression model is appropriate and that a significant amount of variation in the dependent variable is explained by the independent variables, overall, the model applied is significantly good enough in predicting the outcome variable.

Table 3.2 ANOVAd
Model Sum of Squares Df Mean Square F Sig.
1 Regression 306.931 6 51.155 18.018 .000a
Residual 374.753 132 2.839    
Total 681.683 138      
2 Regression 336.290 11 30.572 11.241 .000b
Residual 345.394 127 2.720    
Total 681.683 138      
3 Regression 423.524 15 28.235 13.453 .000c
Residual 258.159 123 2.099    
Total 681.683 138      
a. Predictors: (Constant), att.avg
b. Predictors: (Constant),  att.avg, sn.avg
c. Predictors: (Constant), att.avg, sn.avg, pbc.avg
d. Dependent Variable: int.avg

The coefficients table (3.3) indicates which of the three independent variables predict significantly the dependent variable of intention better. The result shows that all three TPB variables have a role in predicting intention, however, we can see that subjective norm (.452) have a stronger role, then attitude (.273)  and pbc (.057) of the standardized coefficients in fourth red row has a much weaker role to play. Hence, the SN and ATT p-value is less than .000 which means that it is more significant than PBC (.401).

The correlation result (table 3.3) between the dependent variable (intention) and attitude shows that there is a fairly weak but positive correlation between intention and attitude.401, and SN .521 but intention and PBC have a negative correlation -0.12. Also, there is a very weak and no significant relationship between attitude and PBC.    However, there is weak but positive relation between intention, SN and Attitude (table 3.3).

3.3 Co efficientsa
Model Unstandardized Coefficients Standardized Coefficients t Sig. Correlations
B Std. Error Beta Zero-order Partial Part
1 (Constant) 1.470 .420   3.496 .001      
Attitude – .538 .101 .401 5.319 .000 .401 .401 .401
2 (Constant) .793 .389   2.040 .043      
Attitude – .371 .094 .276 3.955 .000 .401 .310 .265
sn_avg .435 .069 .443 6.348 .000 .521 .464 .425
3 (Constant) -.074 1.099   -.067 .947      
Attitude – .367 .094 .273 3.908 .000 .401 .308 .262
sn_avg .445 .070 .452 6.395 .000 .521 .468 .429
pbc_avg .154 .182 .057 .843 .401 -.012 .070 .056
a. Dependent Variable: int_avg

The regression coefficient analysis revealed that subjective norm was a highly significant predictor of intention ((β = 44.5%, p value=.000, and t value = .005), attitude (β = 36.7%, p value=.000, and t value = 3.908) and PBC (β = 15.4%, p value=.401, and t value = 0.843). In summary SN predict approximately 45% of intention, attitude 37% and PBC with the lowest contribution (15).

The part correlation, last column, shows the unique contribution of individual variable to the dependent variable (Intention) indicating that subjective norm (45.2%) contribute more and has a greater significance (.000) to predicting intention than attitude (27.3%) contribution and the lowest contribution with PBC (05.7%) contribution (p-value=.401) inferring that PBC is less significant.  Further analysis was conducted to examine the additional contribution of past behaviour to intention.

3.2 Additional contributions of past behaviour to predicting Intention

Further regression was explored on the additional contribution of past behavior to intention. The r value (47.8%) adjusted R square for this is 22.9%. It is expected that the ANOVA (table 3.3) result will be significant at .000. The coefficients table (3.4) indicates a higher contribution from country to the model (47.8%) comparative to the key predictor variables. Thus, we may infer that past behavior has a higher contribution to predict intention better than the main constituent variables as shown in above (see table 3.3).

Table 3.2 Model Summaryb
Model R R Square Adjusted R Square Std. Error of the Estimate Change Statistics
R Square Change F Change df1 df2 Sig. F Change
1 .478a .229 .223 .259 .229 40.951 1 138 .000
a. Predictors: (Constant), Past Behaviour
b. Dependent Variable: Int.avg

 

Table 3.3 ANOVAb
Model Sum of Squares Df Mean Square F Sig.
1 Regression 157.140 1 157.140 40.951 .000a
Residual 529.545 138 3.837    
Total 686.686 139      
a. Predictors: (Constant), Past Behaviour
b. Dependent Variable: Int.avg

 

Table 3.4 Coefficientsa
Model Unstandardized Coefficients Standardized Coefficients t Sig. Correlations
B Std. Error Beta Zero-order Partial Part
1 (Constant) 1.776 .496   3.578 .000      
Past Behaviour 2.452 .383 .478 6.399 .000 .478 .478 .478
a. Dependent Variable: Int2

4. Discussion and Conclusion

This study examines that the theoryof planned behaviour (TPB) in the context of intention to use a mobile phone while the driving attempt for apply TPB to cross examines  major determinants of the intention for using a mobile phone while driving.   Additionally, past behavior was explored. A sample of students who provided these information about the driving behaviour, as well as views, attitudes and intention regarding for keeping to the regulation banning the use of the mobile phone while driving. The result is however not consistent  TPB studies (Ajzen, 1991, Devellis et al,. 1990). The result shows that the subjective norm and the attitude predicted intention better than the PBC. According to (Ajzen, 1991) perceived behavioural control plays  important role when the behaviour in question becomes more non-volitional. The result of this study also shows that it is more in agreement with intention predictive ability from other studies using the model TPB. The model fit well in the ANOVA t-test, indicating that as can be seen from table 3.2. This study is consistent with other such results to confirm that the TPB model is valid and is significant in predicting intentions to drive using a mobile phone. Results shows that a percentage of intention (table 3.1) may be explained by other variables confirming other studies conclusion that suggested the extension of TPB  by other variables such as past behaviour, self identity among others been added in various studies to explain the variations in intention. The results also show a significant relationship between the variables except pbc which has the lowest correlation (Table 3.3).

Based on  foregoing,  the primary hypothesis that  the theory of planned behaviour is significant in predicting intention is accepted that is consistent with other studies (Kim and Han, 2010; Sang and Chen 2010;Gopi and Ramayah, 2007; Vackier, 2005 etc). Secondly, we noticed that the contribution of the other variables to predicting intention may be responsible to explain the higher percentage variation in intention. Thus, most studies include other variables in order to reduce  limitation of the present study. Results of the study show past behavior rather than the TPB variables to be pertinent, accounting for 47.8% of the variance in sample. Some of the limitations of this study include a low level of internal consistency result for perceived behavioural control, these might be responsible for the null-effect result for PBC which is however not consistent with previous studies (e.g. Pavlou et al, 2006). Empirical evidence may not be available to suggest that this result is applicable to other European countries in terms of generalization.  Indeed, perceived behavioural control has  lowest variance compared to  other variables in our study (.067), and it might therefore be less relevant in prediction of the intentions in this case. A second limit might be lack of consideration which is given to additional variables and lack of the measurement which is given to demographic factors and the actual behaviour, which might be contribute to  outcome. Finally, this study concludes that perceived behavioural control has the lowest significant contribution for predicting intention as we can see that the subjective norm (.452) has stronger role, then the attitude (.273) and the PBC (.057).

Reference

1. Ajzen, I. (1991) The theory of planned behaviour. Organisational Behaviour and Human Decision Processes, 50, 179-211.

2.Armitage, C. J. & Conner, M. (2001) Efficacy and the theory of planned behavior: A meta-analytic review. British Journal of Social Psychology, 40, 471-499.

Tags: ,

Category: Free Essays, Information Technology