Validating the Theory of Planned Behavior Model Extended to Social Marketing Behavioral Enhancers Using Structural Equation Modeling

Abstract

Background:

This study addresses the HIV/AIDS epidemic that constitutes a major health issue in South Africa, the country most burdened by the virus in the world.

Focus of the Article:

It is an empirical study that investigates predictive behavioral patterns between traditional components of the theory of planned behavior and the previously identified social marketing behavioral enhancers and intentions to perform preventative sexual behaviors promoted under the Abstinence, Being faithful, and Condomize campaign.

Research Question:

The main question this study attempts to answer is: Is it relevant to increase the theory of planned behavior components by incorporating the social marketing behavioral enhancers’ variables to design programs that successfully influence individuals to adhere to the preventative sexual behaviors?

Importance to the Social Marketing Field:

Results will tell social marketers, through design programs fighting the spread of the HIV set within a theory of planned behavior theoretical framework, which of the social marketing behavioral enhancers are worth integrating into their model to induce behavioral change.

Methods:

Theory of planned behavior models extended to social marketing behavioral enhancers for abstinence, faithfulness, and condom use were used as theoretical frameworks to test how well they are good fits of the empirically manifested structural models. Gauteng was chosen, because three of the five metropolitan municipalities with a HIV prevalence greater than 10% are located in this province. Data were collected by means of questionnaires administered to a sample chosen randomly, using a multi-stage stratification method. A quota was determined for each suburb or city considered according to the size of its population compared to the overall Gauteng population to ensure representativeness of the study’s sample.

Results:

The study’s theoretical frameworks fitted the data well, but results also revealed insignificant causal relationships between HIV/AIDS knowledge and all Abstinence–Being faithful–Condomize intentions. Similarly, no predictive relationships were found between accessibility to HIV/AIDS information and intention to use condoms, while attitudes toward abstinence and condom use were insignificant with their respective intentions. However, their positive correlations with predictive variables suggest that they influence intentions indirectly.

Recommendation for Research:

Researchers are invited to conduct further studies to test the model in a different context. Indeed, this study does not investigate whether relationships between HIV/AIDS knowledge, accessibility to HIV/AIDS information, and attitudes toward abstinence and condom use would remain insignificant or that it could not change over time in a research ground other than Gauteng. Opportunities should be explored to augment the traditional theory of planned behavior components by variables other than the social marketing behavioral enhancers, in order to build a more robust model that will incorporate more significant factors to design successful programs.

Limitations:

Collecting data from only one province constitutes a limitation in terms of drawing conclusions for the whole South African population.

Keywords

HIV/AIDS theory of planned behavior social marketing behavioral enhancers structural equation modeling health behavior

To propose a remedy to problems that arise in society, particularly in the health sector, several theoretical models were proposed to help understand the determinants of human behaviors and the intentions to perform them (Mermelstein & Revenson, 2013). Numerous studies reveal that the trans-theoretical model and theory of planned behavior are the most extensively used models, with the latter offering the advantage that it may be applied in the analysis of virtually all significant health behaviors and, to a lesser extent, in predictive investigations and the design of health interventions (Taylor et al., 2006). Similarly, empirical evidence from several health domains indicates that the theory of planned behavior is more effective in predicting behaviors when augmented by additional variables to its traditional attitudes, subjective norms, and perceived behavioral control as suggested by Conner and Armitage (1998). Among some examples confirming the extension of the theory of planned behavior model, a study conducted at Midwestern University in Chicago that investigated the prediction of risky sexual behaviors incorporated past sexual risk behavior, anticipated affect, moral norms, sexual excitation, sexual inhibition, and sensation seeking (Turchik et al., 2010). In research conducted in Australia, Robinson et al. (2008, pp. 2559–2567) extended the theory of planned behavior to descriptive norm, moral norm, anticipated regret, and donation anxiety to explain the intention to donate blood among non-donors. Hence, in the effort made to increase the predictive validity of the theory of planned behavior through the addition of external variables, Ayikwa and De Jager (2017) proposes the integration of social marketing behavioral enhancers including accessibility to HIV/AIDS information, accessibility to condoms, and level of HIV/AIDS knowledge as factors influencing sexual behaviors and intentions in the context of the combating the epidemic.

This study thus proposes three structural models that integrate the components of the theory of planned behavior, social marketing behavioral enhancers, and the intentions to perform preventative sexual behaviors in an attempt to ascertain whether there are good fits between the elements of the empirically manifested structural models and theoretically hypothesized models. Indeed, besides its singularity in that it incorporates the social marketing behavioral enhancers alongside the traditional theory of planned behavior constructs, behavioral studies that propose to investigate concomitantly the three preventative sexual behaviors known as ABC (Abstinence–Being faithful–Condomize) are rare or non-existent.

Theoretical Background

The theory of planned behavior is an improved version of the theory of reasoned action that has added perceived behavioral control to the two previous factors of attitude and subjective norms proposed by Ajzen and Fishbein (Kiriakidis, 2015). They posit that intention is the most immediate antecedent of any voluntarily controlled behavior that captures its motivational influences (Madden et al., 1992). These beliefs were divided into two distinct concepts that took into account their behavioral and normative perspectives from which the theory of reasoned action’s components attitude and subjective norms were derived (Ajzen & Fishbein, 1974). However, in their quest to better understand behavioral patterns, scholars and academics have proposed and tested external variables that do not influence intentions through affecting either attitude and subjective norms or both. Thus, additional variables such as personal norms (Fishbein, 1967), moral obligations (Gorsuch & Ortberg, 1983; Zuckerman & Reis, 1978), and competing attitudes (Davidson & Morrison, 1983) were considered for inclusion in the theory of reasoned action model or its expansion. This did not leave indifferent one of the fathers of the theory of reasoned action, Ajzen (1985), who returned to the charge on the same subject by this time suggesting the incorporation of a third determinant emanating from the notion of perceived control over behavioral achievement. Thus was born the theory of planned behavior, which extends the boundary of the theory of reasoned action’s condition of pure volitional control by incorporating individuals’ perceptions of their ability to perform a given behavior (Pakpour & Sniehotta, 2012). The newly developed model displays a direct path between perceived behavioral control and both intention and behavior implying that it does not necessarily influence behavior through intention like attitude and subjective norms.

Over time, researchers realized on the basis of empirical evidence that the use of theory of planned behavior as a theoretical framework made it possible to better explain the factors motivating the performance of a behavior and/or intention, referring to it if one added other variables (Robinson et al., 2008; Turchik et al., 2010). This is why some disciplines, such as social marketing, dedicated to behavior change to promote human well-being and especially health, have begun to identify additional variables that may influence human behavior and intention. According to Andreasen (1994, p. 110): “Social marketing is the application of commercial marketing technologies to the analysis, planning, execution, and evaluation of programs designed to influence the voluntary behaviour of target audiences in order to improve their personal welfare and that of their society.” It is not a question of considering any non-profit activity but rather one that is based on market research and which uses segmentation and targeting techniques as well as marketing mix strategies to design programs intended to change a behavior deemed at risk, without resorting to coercive means (Cheng et al., 2010; French et al., 2010).

Social marketers resort to theoretical frameworks when seeking to identify influential factors and understand the context in which change in behavior shall occur in order to design effective interventions. Although they provide a huge amount of information, frequently used cognitive behavioral models from classic theories such as the theory of planned behavior propose a limited number of variables, and this slightly handicaps the understanding of behaviors and intentions (Fishbein & Yzer, 2003). A study conducted in Addis Ababa using the theory of planned behavior has showed that its variables were less sufficient in predicting voluntary HIV counseling and testing (Mirkuzie et al., 2011). Another compared the theory of reasoned action, theory of planned behavior, and an extended theory of planned behavior model which demonstrated that the augmented model had slightly better fit indexes and predictive value (Gomes & Nunes, 2017). Therefore, researchers have found it necessary to either combine a number of variables of these aforementioned theories or integrate other important aspects they have not as yet considered, as was the case with the integrated model of behavior prediction (Diteweg et al., 2013). This model differs from the theory of planned behavior by adding two additional variables to predict behavior, namely, skills (e.g., knowledge) and environmental constraints (e.g., stigma). Similarly, a study conducted in Botswana among college students by Faimau et al. (2016), by means of a systematic review of articles emphasizing on the theoretical basis of interventions conducted in sub-Saharan Africa, highlights the work of a couple of researchers who developed conceptual frameworks assuming that knowledge directly predicts sexual behavior or indirectly through intention (Michielsen et al., 2012). Likewise, a study comparing the basic theory of planned behavior model and an adapted version that integrated three information behaviors among which the exposure to HIV/AIDS information demonstrated that the proposed model improved predictions, explaining more than twice as much variance as the original information model (Meadowbrooke et al., 2014). Furthermore, researchers in the Netherlands showed that condom coverage through an HIV intervention reduced the burden of the disease by 99% (Bom et al., 2019). This evidence from various health behavior studies in relation with HIV/AIDS prompted the inclusion of knowledge, accessibility to HIV/AIDS information and condoms in this study.

In their research conducted as part of the fight against the epidemic, one of the major public health problems on the African continent, Ayikwa and De Jager (2017) have identified the three abovementioned factors that influence individuals’ adherence to preventative sexual behaviors. Although the researchers suggest that these variables be incorporated to expand the preventative sexual behaviors’ models, no attempt to test how they are structured into validated models has been made. This study intends to do so for intentions to abstain from sex, to being faithful to one partner and condomizing using the structural equation modeling techniques. Elements of the theoretical models developed to be tested are presented in Figure 1.

Figure 1.

Path model for the theory of planned behavior extended to social marketing behavioral enhancers.

Method

Participants and Sampling Method

This study consisted of a sample of 607 South African’s Gauteng Province inhabitants, of whom 310 (51.1%) were female and 297 (48.9%) male aged ≥ 18 years old. All were interviewed by well-trained surveyors, and participation was voluntary. In order to ensure that the study’s sample was a representative distribution of the Gauteng Province’s population, researchers determined a quota for each suburb or city considered according to the size of its population compared to the overall Gauteng population. Participants were chosen randomly using a multi-stage stratification method that consisted of dividing Gauteng in terms of its administrative metropolitan and district municipalities. The majority of the participants (277 or 45.6%) live in Johannesburg, followed by Tshwane (150 or 24.7%), Ekurhuleni/East Rand-Germiston (124 or 20.4%), West Rand (29 or 4.8%), and Sedibeng (27 or 4.5%). Later, the list of towns and cities served as a sampling frame from which cities and towns that made up the study’s research fields were randomly selected. The following cities and towns were selected: Alexandra, Boksburg, Centurion, KwaThema, Mamelodi, Randburg, Sandton, Sebokeng, Soshanguve, Soweto, Tembisa, and Wedela. Furthermore, the decision to conduct the study in Gauteng Province was motivated by the fact that three of the nine metropolitan municipalities that have a higher than 10% HIV prevalence rate are located in this province according to South African National HIV Survey (Human Sciences Research Council, 2012). The municipalities are City of Johannesburg Metro (11.1%), City of Tshwane Metro (11.7%), and Ekurhuleni Metro (14.3%). The Kaiser–Meyer–Olkin measure of sampling adequacy was employed to determine the suitability of the sample. Results indicated values well above .5 for all 15 constructs considered in this study, suggesting the data were sufficient for data processing.

Instruments

In this study, besides the demographic characteristics that were captured to profile participants, 15 instruments were used to measure each parameter involved in the theoretical framework. All instruments comprised items previously used to measure the parameters considered in other studies conducted in sub-Saharan Africa and more specifically the Southern Africa region as debated by scholars such as Diteweg et al. (2013), Sacolo et al. (2013), Ugwu (2012), and Mathews et al. (2009). Given that these items were gathered to form a newly non-administered questionnaire, some items were amended following consultation with academics, scholars, and professionals working in the study’s domain of interest. Their expertise helped essentially in adjusting the wording and formatting of questions found to be ambiguous or too verbose. After these amendments were made, a pilot study was conducted to optimize its reliability and validity. After performing an exploratory factor analysis (EFA) and reliability tests using the software program SPSS Version 23, 31 of the 173 items retained were dropped as highlighted in Table 1. Items removed from this study’s questionnaire were left aside on the basis that they either had a factor loading < .3 or showed correlation > .9 with another item within the same construct, causing a collinearity issue or because it decreased the internal consistency estimates (Cronbach’s alpha) or had low item discrimination ability (i.e., a low item–total correlation).

Table 1.

Validity and Reliability Analysis Using Exploratory Factor Analysis (EFA) and Cronbach Alpha.

Construct	Item retained
Attitude toward abstinence from sex	It is against your values for you to have sexual intercourse before marriage (AttAbs1). Having sexual intercourse is something only married people should do (AttAbs2). Having sex is a good thing to do (AttAbs3). Having sex is okay if adolescent and partner think that it’s okay (AttAbs4). Having sex is not a way of proving love to someone (AttAbs5). I feel comfortable talking about abstinence (AttAbs6). Abstinence requires discipline and willpower (AttAbs7). Having sex okay if dating same person for one year or more (AttAbs8). Sexual intercourse is not the only way two people can get to know each other (AttAbs9). Before engaging into sexual relation, one should think about it carefully (AttAbs10). Having sex is okay if birth control used (AttAbs11). Having sex is okay if plan to marry (AttAbs12). It is possible to postpone sexual relations until a person is able to take responsibility for his or her action (AttAbs13).
Attitude toward faithfulness	Faithfulness is very important in a relationship (AttFaith1). It is possible to contract HIV/AIDS from a regular partner (AttFaith2). It is important that one can talk with his or her partner about HIV/AIDS (AttFaith3). Only the two people who trust each other completely should have sexual relations (AttFaith4). It is boring to have the same sexual partner for a long time (AttFaith5). I don’t see anything wrong in having multiple sexual partners (AttFaith6). It will be easy for me to stay faithful to my partner if I choose to (AttFaith7).
Attitude toward condom use	Having sex without a condom is an expression of love to your partner (AttCond1). Buying condom is embarrassing (AttCond2). Using condom to prevent the spread of HIV/AIDS is more trouble than its worth (AttCond3). Condoms are not easy to use (AttCond4). Condom reduces the quality of sex so much that it is better not to use it at all (AttCond5). Using condom is very unpleasant (AttCond6). Using condom is very healthy (AttCond7). Using condom is very bad thing to do (AttCond8). Using condom is very beneficial (AttCond9). Using condom is not pleasurable at all (AttCond10). Condoms should be used in casual encounters and not in steady relationships (AttCond11). Use of condom with a steady partner indicates lack of trust (AttCond12). The idea of using condoms does not appeal to me (AttCond13). Condoms are not comfortable (AttCond14).
Subjective norms about abstinence from sex	It is against the commonly shared values of your customs to have sexual intercourse before marriage (SubnormAbs1). It is against the commonly shared values of your religion to have sexual intercourse before marriage (SubnormAbs2). Your parents and/or siblings think that sexual intercourse is something only married people should do (SubnormAbs3). Your friends or peers think that sexual intercourse is something only married people should do (SubnormAbs4). Your colleagues think that sexual intercourse is something only married people should do (SubnormAbs5). Your religious leader(s) think that sexual intercourse is something only married people should do (SubnormAbs6). You are under pressure from your siblings to engage in a sexual relationship (SubnormAbs7). You are under pressure from your friends or peers to engage in a sexual relationship (SubnormAbs8).
Subjective norms about faithfulness	You are under pressure from your parents (mother and/or father) to stay faithful to your only partner (SubnormFaith1). You are under pressure from your friends to stay faithful to your only partner (SubnormFaith2). Your partner would approve of you being faithful to him or her (SubnormFaith3). Your parents and or siblings would approve of you being faithful to one sexual partner (SubnormFaith4). Your friends would approve of you being faithful to one sexual partner (SubnormFaith5). Your colleagues would approve of you being faithful to one sexual partner (SubnormFaith6). Your religion or religious leader(s) would approve of you being faithful to one sexual partner (SubnormFaith7). Your customs would approve of you being faithful to one sexual partner (SubnormFaith8). Your partner expects of you being faithful to him or her (SubnormFaith9). Your parents and/or siblings expect of you being faithful to one sexual partner (SubnormFaith10). Your friends expect of you being faithful to one sexual partner (SubnormFaith11). Your colleagues expect of you being faithful to one sexual partner (SubnormFaith12). Your religion or religious leader(s) expect of you being faithful to one sexual partner (SubnormFaith13). Your religion or religious leader(s) promote polygamy (SubnormFaith14). Your customs promote polygamy (SubnormFaith15). Your religion or religious leader(s) allow extra marital relationship (SubnormFaith16).
Subjective norms about condom use	Your parents and/or siblings think that you should use a condom every time you have sex (SubnormCond1). Your friends and/or peers think that you should use a condom every time you have sex (SubnormCond2). Your colleagues think that you should use a condom every time you have sex (SubnormCond3). Your religion and/or religious leader(s) think that you should use a condom every time you have sex (SubnormCond4). Your parents and/or siblings expect of you that you use a condom every time you have sex (SubnormCond5). Your friends and/or peers expect of you that you use a condom every time you have sex (SubnormCond6). Your colleagues expect of you that you use a condom every time you have sex (SubnormCond7). Your parents and/or siblings would disapprove your using of condom every time you have sex (SubnormCond8). Your friends and/or peers would disapprove your using of condom every time you have sex (SubnormCond9). Your colleagues would disapprove your using of condom every time you have sex (SubnormCond10). Your religion and/or religious leader(s) would disapprove your using of condom every time you have sex (SubnormCond11).
Perceived behavioral control over abstinence	If you wanted, you could abstain from sex before marriage (PerBehavControlAbs1). It is mostly up to you whether or not to abstain from sex before marriage (PerBehavControlAbs2).
Perceived behavioral control over faithfulness	If you wanted, you could be faithful to one sexual partner (PerBehavControlFaith1). It is mostly up to you whether or not to have more than one sexual partner (PerBehavControlFaith2).
Perceived behavioral control over condom use	It is easy for you to use a condom every time you have sex (PerBehavControlCond1). If you wanted, you could use a condom every time you have sex (PerBehavControlCond2). It is mostly up to you whether or not to use a condom every time you have sex (PerBehavControlCond3).
Accessibility to HIV/AIDS information	Articles in magazines (Channel1). Articles in newspapers (Channel2). Animated banners (Channel3). Flash banners (Channel4). Static banners (Channel5). Billboards (Channel6). Internet (Channel7). Leaflets and information booklets (Channel8). Murals (Channel9). Plays or drama (Channel10). Posters (Channel11). Radio advertisements (Channel12). Radio dramas, talk shows, and other programs (Channel13). Radio news (Channel14). Television advertisements (Channel15). Television dramas, talk shows, and other programs (Channel16). Television news (Channel17).
Accessibility to condoms	Pharmacy (AccessCond1). Kiosk (AccessCond2). Health center/clinic (AccessCond3). Supermarket (AccessCond4). Friends (AccessCond5). Other (Specify) (AccessCond6).
HIV/AIDS level of knowledge	HIV/AIDS is a sexually transmitted disease (Knowledge1). HIV/AIDS is a preventable disease (Knowledge2). HIV/AIDS is a manageable disease (Knowledge3). HIV/AIDS is a blood-borne disease (Knowledge4). Swollen lymph glands (Knowledge5). Purplish spots on the skin (Knowledge6). Night sweats (Knowledge7). Persistent fever (Knowledge8). Fatigue (Knowledge9). Long-lasting diarrhea (Few weeks) (Knowledge10). Persistent cough (Knowledge11). Shortness of breath (Knowledge12). Abdominal discomfort (Knowledge13). Yeast infections in the mouth or vagina (Knowledge14). Itching in the genital area (Knowledge15). Pain during sex (Knowledge16). Sores on the genitals (Knowledge17). Burning sensation when urinating (Knowledge18). Unusual discharge (Knowledge19). A person can get HIV/AIDS if engaged in unprotected sex (Knowledge20). A person can get HIV/AIDS through transfusion of unscreened blood (Knowledge21). A person can get HIV/AIDS through sharing anyone of the following: needles, razors, or blades (Knowledge22). A person can get HIV/AIDS through sharing anyone of the following: toilet or room with an infected person (Knowledge23). A person can get HIV/AIDS through sharing meals with an infected person (Knowledge24). A person can get HIV/AIDS through insects/ animal’s bite (Knowledge25). A person can get HIV/AIDS through handshaking with an infected person (Knowledge26). A person can get HIV/AIDS through hugging with an infected person (Knowledge27). A person can get HIV/AIDS through kissing with an infected person (Knowledge28). A mother can transmit HIV/AIDS to her unborn baby during pregnancy (Knowledge29). A mother can transmit HIV/AIDS to her new born baby during breastfeeding (Knowledge30). Avoid transfusion (Knowledge31). Avoid injections (Knowledge32). Avoid kissing (Knowledge33). Seek protection from traditional healer (Knowledge34). Take a shower immediately after sex (Knowledge35).
Intention to abstain from sex	You intend to abstain from sex before marriage (IntentAbs1). You never expect to have sex before marriage (IntentAbs2). You are certain about your intentions to avoid sex before marriage (IntentAbs3).
Intention to being faithful to one partner	You intend to be faithful to one partner (IntentFaith1). You are certain about your intentions to have only one partner (IntentFaith2).
Intention to using condoms during sexual intercourse	You intend to use a condom every time you have sex (IntentCond1). You are certain about your intentions to use a condom every time you have sex (IntentCond2). It is likely that you will use a condom every time you have sex (IntentCond3).

Factor analysis used principal component analysis and orthogonal varimax rotation extraction method for scaled variables to group items in a data set called factor on the basis of strong correlations, exploring the underlying theoretical structure of the phenomena (Tabachnick & Fidell, 2013). Therefore, in view of the structural equation modeling procedure, items loading within the same factor explaining a latent variable have been combined as measurement of the underlying dimension they represent (Streiner & Norman, 2003). In case of one factor explaining a latent variable, items loading within the specific factor are considered singularly as observed variable at confirmatory factor analysis stage as illustrated in Table 2. However, prior performing the EFA based on the unweighted least squares extraction method for variables measured through dichotomous (binary) options, a tetrachoric correlation matrix was estimated using the TETRA-COM Macro for SPSS as suggested by Lorenzo-Seva and Ferrando (2012, pp. 1191–1196). Attitude, subjective norms, perceived behavioral control, and intentions toward abstinence, being faithful to one partner, and condom use required a 5-point Likert-type scale measure (from totally disagree to totally agree), while accessibility to HIV/AIDS information and level of knowledge regarding HIV/AIDS utilized dichotomous answers (yes or no). Accessibility to condoms was assessed through a 3-point Likert-type scale (not easy, easy, and very easy).

Table 2.

CFA Level: Validity and Reliability Analysis on Constructed Scales Using Average Variance Extracted (AVE) and Composite Reliability (CR).

Construct	Item	Factor Loading	AVE	CR
Attitude toward abstinence	AttAbs3	0.48	.542	.890
	AttAbs5	0.67
	AttAbs6	0.74
	AttAbs7	0.85
	AttAbs9	0.76
	AttAbs10	0.80
	AttAbs13	0.79
Attitude toward faithfulness	AttFaith1 +AttFaith2 +AttFaith3 +AttFaith4 +AttFaith7	0.99	.470	.763
Attitude toward faithfulness	AttFaith5 +AttFaith6	0.54
Attitude toward condom use	AttCond1 +AttCond2 +AttCond3 +AttCond4 +AttCond5 +AttCond6 +AttCond10 +AttCond11 +AttCond12 +AttCond13 +AttCond14	0.63	.659	.788
Attitude toward condom use	AttCond7 +AttCond8 +AttCond9	0.96
Subjective norms about abstinence from sex	SubnormAbs1	0.81	.583	.847
	SubnormAbs2	0.84
	SubnormAbs3	0.66
	SubnormAbs6	0.73
Subjective norms about faithfulness	SubnormFaith3 +SubnormFaith4 +SubnormFaith7 +SubnormFaith8 +SubnormFaith9 +SubnormFaith12 +SubnormFaith13	0.99	.779	.874
Subjective norms about faithfulness	SubnormFaith5 +SubnormFaith6 +SubnormFaith10 +SubnormFaith11	0.76
Subjective norms about condom use	SubnormCond1	0.89	.710	.945
	SubnormCond2	0.89
	SubnormCond3	0.74
	SubnormCond4	0.83
	SubnormCond5	0.90
	SubnormCond6	0.89
	SubnormCond7	0.74
Perceived behavioral control over abstinence	PerBehavControlAbs1	0.73	.636	.777
Perceived behavioral control over abstinence	PerBehavControlAbs2	0.86
Perceived behavioral control over faithfulness	PerBehavControlFaith1	0.71	.666	.797
Perceived behavioral control over faithfulness	PerBehavControlFaith2	0.91
Perceived behavioral control over condom use	PerBehavControlCond1	0.79	.664	.854
	PerBehavControlCond2	0.93
	PerBehavControlCond3	0.71
Accessibility to HIV/AIDS information	Channel2 +Channel10 +Channel12 +Channel13 +Channel14 +Channel15 +Channel16 +Channel17	0.75	.442	.683
	Channel1 +Channel3 +Channel4 +Channel6 +Channel7 +Channel8 +Channel11	0.80
	Channel5 +Channel9	0.35
Accessibility to condoms	AccessCond1 +AccessCond2 +AccessCond3 +AccessCond4	0.43	.429	.578
Accessibility to condoms	AccessCond5 +AccessCond6	0.82
HIV/AIDS level of Knowledge	Knowledge14 +Knowledge15 +Knowledge16 +Knowledge17 +Knowledge18 +Knowledge19	0.53	.449	.691
	Knowledge5 +Knowledge6 +Knowledge7 +Knowledge9 +Knowledge10 +Knowledge12 +Knowledge13	0.92
	Knowledge8 +Knowledge11	0.47
Intention to abstain from sex	IntentAbs1	0.85	.617	.828
	IntentAbs2	0.71
	IntentAbs3	0.79
intention to being faithful to one partner	IntentFaith1	1.00	.861	.925
intention to being faithful to one partner	IntentFaith2	0.85
Intention to using condoms during sexual intercourse	IntentCond1	0.92	.704	.877
	IntentCond2	0.83
	IntentCond3	0.76

Research Design and Data Analysis

The three preventative sexual behaviors of intentions toward abstinence, being faithful to one partner, and condom use promoted by social marketing campaigners addressing the epidemic under the well-known Abstinence–Being faithful–Condomize label were the dependent variables. The independent variables were constituted by the three theory of planned behavior components of attitude, subjective norms, and perceived behavioral control in relation to abstinence from sex, faithfulness to one partner, and condom use as well as social marketing behavioral enhancers including accessibility to HIV/AIDS information and condoms and level of knowledge regarding HIV/AIDS. The dependent and independent variables covered all constructs in the theoretical frameworks developed in accordance to the suggestion by Ayikwa and De Jager (2017).

The data that captured information of items retained and reorganized by component variables that best describe each construct’s dimension, identified after EFA, were processed to assess the study’s hypothesized model through structural equation modeling techniques using the software program AMOS Version 23.

Results

Validity and Reliability Analysis on the Constructed Scales

Prior to drawing all the study’s constructs in the three final models with regard to the Abstinence–Being faithful–Condomize intentions, a confirmatory factor analysis was performed on each dimension providing its respective reliabilities and discriminant validities through average variance extracted and composite reliability tests. The components that showed factor loading < .4 were removed for being the major causes of poor fitting of the construct’s measurement model, while the threshold used for average variance extracted and composite reliability was set at ≥ .5 and ≥ .6, respectively, as suggested in Hair et al. (2010). However, a lesser value of the standard factor loading for convergent validity (≥ .35) and/or average variance extracted (≥ .4) was accepted in order to proceed with the overall model fit on the condition that offending estimate does not occur (Huang, 2007). Offending estimate is manifested by negative error covariance and/or standardized regression coefficient close to 1 (Shyu et al., 2013). Similarly, composite reliability value ≥ .5 was accepted to ascertain reliability following Nunaly’s suggestion (Shyu et al., 2013). Thus, accessibility to HIV/AIDS information and condoms has been able to keep all their component variables. The performing of validity and reliability tests at confirmatory factor analysis level, by means of average variance extracted and composite reliability, enabled the study to drop 40 more items contained in underperforming factors, as highlighted in Table 2. Consequently, items of factors that qualified as a unique underlying dimension of their respective constructs were used as observed variables.

Structural Equation Modeling Analysis

All three tested extended theory of planned behavior models, with regard to Abstinence–Being faithful–Condomize intentions, fitted the data well. All fit statistics were at an acceptable level for intentions to abstain from sex, being faithful to one partner, and using condoms during sexual intercourse as the following demonstrates: Intention to abstain: CMIN = 553.11; DF = 234; CMIN/DF = 2.36; p = .00; GFI = .93; NFI = .91; CFI = .94; TLI = .93; RMSEA = .05; SRMR = .04. Intention to being faithful to one partner: CMIN = 300.85; DF = 89; CMIN/DF = 3.38; p = .00; GFI = .94; NFI = .93; CFI = .95; TLI = .94; RMSEA = .06; SRMR = .05. Intention to use condoms during sexual intercourse: CMIN = 661.95; DF = 211; CMIN/DF = 3.14; p = .00; GFI = .91; NFI = .92; CFI = .94; TLI = .93; RMSEA = .06; SRMR = .05. The overall evaluation of these indices thus led to the validation of the extended TPB models for Abstinence–Being faithful–Condomize intentions. However, results in the path diagrams displayed in Figure 2, 3, and 4 highlighted the non-significance of HIV/AIDS knowledge as predictor of all three Abstinence–Being faithful–Condomize intentions. Similarly, no causal relationship was found between attitudes toward abstinence from sex and condom use during sexual intercourse and their related intentions. Results likewise indicated that accessibility to HIV/AIDS information was statistically insignificant in predicting the intention to use condoms during sexual intercourse. All the other paths shown in the extended theory of planned behavior models for Abstinence–Being faithful–Condomize intentions are significant, as critical ratios were above 1.96.

Results in Figure 2, 3, and 4 suggest the existence of causal relationships between the intention to abstain from sex and accessibility to HIV/AIDS information (β = .35; p ≤ .01), subjective norms (β = −.25; p ≤ .05), and perceived behavioral control (β = .28; p ≤ .01). In line with the aforementioned SEM results, it has been concluded that the respondents’ reported uncertainty about abstaining from sex (intention to abstain from sex: M = 3.20; σ = 1.29) is driven by their poor exposure to HIV/AIDS information (accessibility to information: M = .19; σ = .22) and personal choice not to refrain from engaging in sex, admitting that abstinence is under their control (perceived behavioral control over abstaining from sex: M = 3.95; σ = 1.33), while challenged by family and religious related values (subjective norms about abstaining from sex: M = 3.57; σ = 1.57) that sex should be performed within wedlock. Social marketers should therefore design programs that consider better exposure to HIV/AIDS information, promote abstinence, and increase family and social networks’ influence with regard to abstaining from engaging in sex before marriage and consequently enhance individuals’ will to refrain from sexual intercourse.

Figure 2.

Structural model of the theory of planned behavior extended to social marketing behavioral enhancers for abstinence.

The structural equation modeling results of the extended theory of planned behavior model for intention to being faithful to one partner identified accessibility to HIV/AIDS information (β = −.21; p ≤ .01), attitude (β = .77; p ≤ .01), subjective norms (β = .78; p ≤ .01), and perceived behavioral control (β = .97; p ≤ .01) as predictors of the intention to being faithful. These results lead to the conclusion that respondents’ reported intention to being faithful to one partner is driven by their favorable attitude toward faithfulness (attitude toward faithfulness: M = 4.07; σ = .22) and family and social network values that deem it important to be faithful to the partner (subjective norms about faithfulness: M = 3.93; σ = 1.32) as well as their perceived behavioral control over being faithful (perceived behavioral control over faithfulness: M = 4.16; σ = 1.35), while challenged by their poor exposure to HIV/AIDS information (accessibility to information: M = .19; σ = .22). Thus, social marketers are recommended to implement programs that consider better exposure to HIV/AIDS information, promoting faithfulness, while reinforcing individuals’ favorable attitude and the influence of family and social networks with regard to faithfulness as well as their perceived control over being faithful to one partner.

Figure 3.

Structural model of the theory of planned behavior extended to social marketing behavioral enhancers for faithfulness.

Results of the extended theory of planned behavior model for the intention to use condoms in Figure 2, 3, and 4 identified accessibility to condoms (β = .37; p ≤ .01), subjective norms (β = .23; p ≤ .01), and perceived behavioral control (β = .66; p ≤ .01) as predictors of intention to use condoms during sexual intercourse. These results reflect that respondents’ reported intention to use condoms during sexual intercourse is driven by the ease of obtaining condoms (accessibility to condoms: M = 2.01; σ = .66), their family and social networks’ opinion about condom use (subjective norms about condom use: M = 3.64; σ = 1.31), and their perceived behavioral control over condom use (perceived behavioral control over condom use: M = 4.01; σ = 1.31). Hence, social marketers are invited to design programs that increase distribution of condoms, while reinforcing family and social networks’ opinion with regard to the use of condoms as well as individuals’ perceived control over using condoms during sexual intercourse.

Figure 4.

Structural model of the theory of planned behavior extended to social marketing behavioral enhancers for condom use.

Although no significant causal relationship was found between HIV/AIDS knowledge and all three Abstinence–Being faithful–Condomize intentions and between accessibility to HIV/AIDS information and intention to use condoms during sexual intercourse as well as between attitudes toward abstinence and condom use and their related intentions, they remain variables of importance for influencing the Abstinence–Being faithful–Condomize intentions because of their significant correlations with predictive variables as shown in Table 3.

Table 3.

Correlation Between Latent Independent Variables.

Extended TPB Model for Intention to Abstain From Sex			Extended TPB Model for Intention to Being Faithful to One Partner			Extended TPB Model for Intention to Use Condoms During Sexual Intercourse
Variables	r	p	Variables	r	p	Variables	r	p
Accessibility to HIV/AIDS information: HIV/AIDS level of knowledge	.082	.160	Accessibility to HIV/AIDS information: HIV/AIDS level of knowledge	.062	.207	Accessibility to HIV/AIDS information: HIV/AIDS level of knowledge	.066	.192
HIV/AIDS level of knowledge: Accessibility to condoms	—	—	HIV/AIDS level of knowledge: Accessibility to condoms	—	—	HIV/AIDS level of knowledge: Accessibility to condoms	.111	.146
Accessibility to HIV/AIDS information: Accessibility to condoms	—	—	Accessibility to HIV/AIDS information: Accessibility to condoms	—	—	Accessibility to HIV/AIDS information: Accessibility to condoms	.123^a	.000
Subjective norms: Attitude	.635^b	.000	Subjective norms: Attitude	.793^b	.000	Subjective norms: Attitude	.541^b	.000
Attitude: Perceived behavioral control	.735^b	.000	Attitude: Perceived behavioral control	.799^b	.000	Attitude: Perceived behavioral control	.690^b	.000
Subjective norms: Perceived behavioral control	.643^b	.000	Subjective norms: Perceived behavioral control	.807^b	.000	Subjective norms: Perceived behavioral control	.637^b	.011
Subjective norms: Accessibility to HIV/AIDS information	.109	.075	Subjective norms: Accessibility to HIV/AIDS information	.106^a	.031	Subjective norms: Accessibility to HIV/AIDS information	.124^a	.007
Subjective norms: HIV/AIDS level of knowledge	.086	.070	Subjective norms: HIV/AIDS level of knowledge	.122^a	.008	Subjective norms: HIV/AIDS level of knowledge	.128^a	.048
Subjective norms: Accessibility to condoms	—	—	Subjective norms: Accessibility to condoms	—	—	Subjective norms: Accessibility to condoms	.143	.188
Attitude: Accessibility to HIV/AIDS information	.156^a	.014	Attitude: Accessibility to HIV/AIDS information	.108^a	.028	Attitude: Accessibility to HIV/AIDS information	.064^a	.045
Attitude: HIV/AIDS level of knowledge	.114^a	.017	Attitude: HIV/AIDS level of knowledge	.101^a	.024	Attitude: HIV/AIDS level of knowledge	.094	.106
Attitude: Accessibility to condoms	—	—	Attitude: Accessibility to condoms	—	—	Attitude: Accessibility to condoms	.118^a	.010
Perceived behavioral control: Accessibility to HIV/AIDS information	.128^a	.050	Perceived behavioral control: Accessibility to HIV/AIDS information	.064	.169	Perceived behavioral control: Accessibility to HIV/AIDS information	.132	.058
Perceived behavioral control: HIV/AIDS level of knowledge	.061	.220	Perceived behavioral control: HIV/AIDS level of knowledge	.101	.052	Perceived behavioral control: HIV/AIDS level of knowledge	.092	.842
Perceived behavioral control: Accessibility to condoms	—	—	Perceived behavioral control: Accessibility to condoms	—	—	Perceived behavioral control: Accessibility to condoms	.015^a	.000

Note. ^aSmall effect size. ^bMedium effect size. ^cLarge effect size.

Conclusion and Recommendations for Further Research

This study validated the theoretical framework proposed by Ayikwa and De Jager (2017) of predicting intentions to perform preventative sexual behaviors using structural equation modeling procedures. It contributed in demonstrating the relevance of extending the theory of planned behavior to social marketing behavioral enhancers to better predict human intentions to perform Abstinence–Being faithful–Condomize behaviors based on empirical evidence. Insignificant variables in predicting intentions to adhere to preventative sexual behaviors such as knowledge about HIV/AIDS, accessibility to HIV/AIDS information, as well as attitudes toward abstinence and condom use can be considered as strategic factors that indirectly induce the needed behavior change’s intention through influencing other predictive variables. Thus, it is worth integrating the identified social marketing behavioral enhancers to the theory of planned behavior models to predict preventative sexual behaviors and when designing and implementing HIV/AIDS programs to combat the spread of the virus and eliminate the disease.

Although this study is a logical continuation of Ayikwa et al. (2019), it is nevertheless different in many ways. First, the type of models built previously were measurement models, while this study comprises a structural part highlighting the nature and strength of causal relationships between the components of the theory of planned behavior extended to social marketing behavioral enhancers and intentions to perform preventative sexual behaviors. Second, the inclusion of intentions enabled this study to have both independent and dependent variables. In the previous study, all variables of the theory of planned behavior extended to social marketing behavioral enhancers were independent variables. Third, the final extended theory of planned behavior measurement models in this study does not consider escalating accessibility to HIV/AIDS information, accessibility to condoms, and HIV/AIDS level of knowledge as reflective latent variables of social marketing behavioral enhancers’ second-order latent variable. Indeed, the previous study was a comparative one between first and second-order extended theory of planned behavior models extended to social marketing behavioral enhancers. Finally, the previous study investigated simultaneously abstinence, faithfulness, and condom use in the same model, while the present one analyzes each preventative sexual behavior independently from others.

A suggestion for future research is to examine the prospect of other variables other than the social marketing behavioral enhancers considered in this study that have the potential to influence intentions to perform preventative sexual behaviors to further strengthen the theory of planned behavior in order to affect behavior change. Inclusion of other variables for predicting preventative sexual behaviors identified by previous researches in the context of the struggle against the HIV/AIDS alongside the theory of planned behavior and social marketing behavioral enhancers’ components in empirical testing of the model could be pertinent. The synergy they might create can lead to the design and implementation of powerful and successful social marketing programs and campaigns. Another aspect to be examined would be the predictive and mediating role of demographic characteristics in the prediction of preventative sexual behaviors. Broadening the field of data to be collected to provinces other than Gauteng is necessary to get a larger and representative sample of South Africa to help authorities to implement a nationwide policy. The South African experience can also be replicated in the whole Southern Africa region and elsewhere across the continent to alleviate the devastation caused by the epidemic. Furthermore, longitudinal research using this study’s questionnaire would be interesting to test the stability of the models. Thus, the impact of cyclical effects could be disassociated from the structural to keep a universal version of the models.

It clearly appears from the above that this study encountered two major limitations that are related to time horizon and geographical coverage. Indeed, it employed a cross-sectional approach for data collection that consisted of the observance of the unit of analysis at one point in time. Furthermore, the study was conducted only in Gauteng Province, while the strengthening of its findings requires that research be conducted in more, if not all, South African provinces.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Lutete Christian Ayikwa

References

Ajzen

(1985). From intentions to actions: A theory of planned behavior. In Kuhl

Beckmann

(Eds.), Action control: From cognition to behavior. Springer.

Ajzen

Fishbein

(1974). Factors influencing intentions and the intention-behavior relation. Human Relations, 27(1), 1–15. https://doi.org/10.1177/001872677402700101

Andreasen

A. R.

(1994). Social marketing: Its definition and domain. Journal of Public Policy & Marketing, 13(1), 108–114.

Ayikwa

L. C.

De Jager

J. W.

(2017). Identifying variables to incorporate in an extended TPB model with regards to HIV testing in Kinshasa, DR Congo. International Journal of Environmental & Science Education, 12(7), 1595–1603.

Ayikwa

L. C.

De Jager

J. W.

Van Zyl

(2019). Enhancing the theory of planned behaviour by incorporating social marketing behavioural enhancers: A first VS second order confirmatory factor analysis approach. Journal of Economics and Behavioral Studies, 11(1), 139–151.

Bom

R. J. M.

Van der Linden

Matser

Poulin

Van der Loeff

M. F. S

Bakker

B. H. W.

Van Boven

T. F.

(2019). The effects of free condom distribution on HIV and other sexually transmitted infections in men who have sex with men. BMC Infectious Diseases, 19, 222. https://doi.org/10.1186/s12879-019-3839-0

Cheng

Kotler

Lee

(2010). Social marketing for public health. Jones & Bartlett Learning.

Conner

Armitage

C. J.

(1998). Extending the theory of planned behavior: A review and avenues for further research. Journal of Applied Social Psychology, 28(15), 1429–1464.

Davidson

Morrison

(1983). Predicting contraceptive behavior from attitudes: A comparison of within- versus across-subject procedures. Journal of Personality and Social Psychology, 45(5), 997–1009.

10.

Diteweg

Van Oostwaard

Tempelman

Vermeer

Appels

Van der Schaaf

Maree

(2013). AIDS awareness and VCT behaviour: An application of the integrated model of behaviour prediction. Health SA Gesondheid, 18(1). https://dx-doi-org.web.bisu.edu.cn/10.4102/hsag.v18i1.530

11.

Faimau1

Maunganidze

Tapera

Mosomane

L. C. K.

Apau

(2016). Knowledge of HIV/AIDS, attitudes towards sexual risk behaviour and perceived behavioural control among college students in Botswana. Cogent Social Sciences, 2(1164932). https://dx-doi-org.web.bisu.edu.cn/10.1080/23311886.2016.1164932

12.

Fishbein

(1967). Readings in attitude theory and measurement. Wiley.

13.

Fishbein

Yzer

M. C.

(2003). Using theory to design effective health behavior interventions. Communication Theory, 2, 164–183. https://dx-doi-org.web.bisu.edu.cn/10.1111/j.1468-2885.2003.tb00287.x;https://dx-doi-org.web.bisu.edu.cn/10.1093/ct/13.2.164

14.

French

Blair-Stevens

Mcvey

Merritt

(2010). Social marketing and public health theory and practice. Oxford University Press.

15.

Gomes

A. I. C. S.

Nunes

M. C. S.

(2017). Predicting condom use: A comparison of the theory of reasoned action, the theory of planned behavior and an extended model of TPB. Psicologia: Teoria e Pesquisa, 33, 1–8. https://dx-doi-org.web.bisu.edu.cn/10.1590/0102.3772e33422

16.

Gorsuch

Ortberg

(1983). Moral obligations and attitudes: Their relation to behavioral intentions. Journal of Personality and Social Psychology, 44, 1025–1028.

17.

Hair

Black

Babin

Anderson

(2010). Multivariate data analysis: A global perspective. 7th ed. Pearson Education.

18.

Huang

F. M.

(2007). Structural equation modeling: theory and application. Wu Nan.

19.

Human Sciences Research Council. (2012). South African national HIV survey. https://www.hsrcpress.ac.za/books/south-african-national-hiv-prevalence-incidence-and-behaviour-survey-2012

20.

Kiriakidis

(2015). Theory of planned behaviour: The intention-behaviour relationship and the perceived behavioural control (PBC) relationship with intention and behaviour. International Journal of Strategic Innovative Marketing, 03, 40–51. https://doi.org/10.15556/IJSIM.02.03.004

21.

Lorenzo-Seva

Ferrando

P. J.

(2012). TETRA-COM: A comprehensive SPSS program for estimating the tetrachoric correlation. Behaviour Research, 44, 1191–1196. https://doi.org/10.3758/s13428-012-0200-6

22.

Madden

Ellen

Ajzen

(1992). A comparison of the theory of planned behavior and the theory of reasoned action. Personality and Social Psychology Bulletin, 18(1), 3–9.

23.

Mathews

Aarø

Flisher

Mukoma

Wubs

Schaalma

(2009). Predictors of early first sexual intercourse among adolescents in Cape Town, South Africa. Health Education Research, 24(1), 1–10. https://doi.org/10.1093/her/cym079

24.

Meadowbrooke

C. C.

Veinot

T. C.

Loveluck

Hickok

Bauermeister

J. A.

(2014). Information behavior and HIV testing intentions among young men at risk for HIV/AIDS. Journal of the American Society for Information Science and Technology, 65(3), 609–620. https://doi.org/10.1002/asi.23001

25.

Mermelstein

Revenson

(2013). Applying theory across settings, behaviors and populations: Translational challenges and opportunities. Health Psychology, 32(5), 592–596. https://doi.org/10.1037/a0030582

26.

Michielsen

Chersich

Temmerman

Dooms

Van Rossem

(2012). Nothing as practical as a good theory? The theoretical basis of HIV prevention interventions for young people in sub-Saharan Africa: A systematic review. AIDS Research and Treatment. https://doi.org/10.1155/2012/345327

27.

Mirkuzie

A. H.

Sisay

M. M.

Moland

K. M.

Åstrøm

A. N.

(2011). Applying the theory of planned behaviour to explain HIV testing in antenatal settings in Addis Ababa: A cohort study. BMC Health Services Research, 11, 196. http://www.biomedcentral.com/1472-6963/11/196

28.

Pakpour

Sniehotta

(2012). Perceived behavioural control and coping planning predict dental brushing behaviour among Iranian adolescents. Journal of Clinical Periondontology, 39(2), 132–137. https://doi.org/10.1111/j.1600-051X.2011.01826.x

29.

Robinson

Masser

White

Hyde

Terry

(2008). Predicting intentions to donate blood among non-donors in Australia: An extended theory of planned behavior. Transfusion, 48(12), 2559–2567.

30.

Sacolo

Chung

M-H.

Chu

Liao

Y-M.

Chen

C-H.

K-L.

Chou

K-R.

(2013). High risk sexual behaviors for HIV among the in-school youth in Swaziland: A structural equation modeling approach. PLoS One, 8(7). https://doi.org/10.1371/journal.pone.0067289

31.

Shyu

C-S.

Y-L.

Tang

(2013). Applying confirmatory factor analysis on the measure for restaurant over-service. The Journal of International Management Studies, 8(2), 10–16.

32.

Streiner

D. L.

Norman

G. R.

(2003). Health measurement scales: A practical guide to their development and use. Oxford University Press.

33.

Tabachnick

B. G.

Fidell

L. S.

(2013). Using multivariate statistics. 6th ed. Pearson Education.

34.

Taylor

Bury

Campling

Carter

Garfied

Newbould

Rennie

(2006). A review of the use of the health belief model (HBM), theory of reasoned action (TRA), theory of planned behaviour (TPB), and trans-theoretical model (TTM) to study and predict health related behaviour change. University of London.

35.

Turchik

Probst

Irvin

Chau

Gidycz

(2010). Prediction of sexual assault experiences in college women based on rape scripts: A prospective analysis. Psychology of Violence, 1(5), 76–83.

36.

Ugwu

O. M.

(2012). Knowledge and attitude of secondary school students towards sexually transmitted diseases in Nsukka education zone [Unpublished master thesis]. University of Nigeria.

37.

Zuckerman

Reis

(1978). Comparison of three models for predicting altruistic behavior. Journal of Personality and Social Psychology, 36, 498–510.