Measuring Consumer Attitude towards Imported Poultry Meat Products in a Developing Market : An Assessment of Reliability , Validity and Dimensionality of the Tri-Component Attitude Model

As competition increases the world over, it has become imperative for marketers to understand their target customers. In this regard, consumer attitude towards imports has received increased attention in research. Despite its importance in marketing, consumer attitude has not been measured consistently. This study, therefore, sought to assess the measurement of consumer attitude by considering the reliability, validity and dimensionality of consumer attitude towards imported poultry products in Zimbabwe—a developing market—using the Tri-component attitude model as the basis for the research model. A cross-section of 400 poultry consumers was taken in Harare and Bulawayo using an interviewer-administered questionnaire. The structural equation modelling technique was applied to analyse data. The study found that consumer attitude toward imported poultry products comprises three dimensions, namely beliefs (cognition), feelings (affect) and intentions to buy (conation). Marketers and researchers are advised to consider these factors when measuring consumer attitude toward imported poultry products in developing countries.


Introduction
There has been increased globalisation for over the years.This phenomenon has not only presented opportunities to marketers, but also challenges (Ranjbarian, Rojuee & Mirzaei, 2010:372).The world trade barriers have continued to tumble, giving rise to more and more exposure of consumers to various products from other countries (Teo, Mohamad & Ramayah, 2011:2805;Ranjbarian et al., 2010:372;Saffu & Walker, 2006:168).The global poultry sector has also witnessed massive international trade of poultry products.Consequently, competition has heightened in the global poultry sector (Vukasovi , 2009:65).In this competitive environment, it has become a matter of survival for marketers to understand their target markets.As such, marketers now seek to have a clear understanding of every bit of information about their target customers (Mangnale, Potluri & Degufu, 2011:241).In this regard, the concept of consumer attitude in consumer behaviour studies has received much attention as marketers endeavour to understand their target markets (Du Plessis, Rousseau, Boshoff, Ehlers, Engelbrecht, Joubert & Sanders, 2007:260;Blackwell, Miniard & Engel, 2006:375;Assael, 2004:216;Solomon, 2002:527).
Consumer attitudes play a crucial role in the behaviour of consumers.Therefore, in order to adequately understand the behaviour of consumers, one must also understand consumer attitudes (Schiffman & Kanuk, 2004:253;Wilcock, Pun, Khanona & Aung, 2004:56).In the context of consumer behaviour, an attitude refers to the consistent tendency of consumers to behave, favourably or unfavourably, with regards to a specific product or brand (Schiffman & Kanuk, 2004:253).Attitudes play an important part in developing marketing strategy.Likewise, Assael (2004:222) suggests that attitudes help marketers to identify benefit segments, develop new products, and formulate and evaluate promotional strategies.Similarly, Wilcock et al., (2004:56) observe that the knowledge of consumer attitudes is critical in that it enables marketers to predict consumer behaviour.On the other hand, consumer behaviour is believed also to influence consumer attitudes (Schiffman & Kanuk, 2004:265).Moreover, Juric and Worsely (1998:43) posit that in order to succeed in the international market place, marketers need to consider the attitudes of consumers toward foreign products.In this regard, Schiffman and Kanuk (2004:269) acknowledge that changing the attitudes of consumers is a fundamental strategy consideration for most marketers.As such, an accurate measurement of consumer attitude is particularly important in the success of marketing strategy formulation.
Despite the importance of consumer attitude in influencing consumer behaviour, this construct has not been measured consistently in literature.Evidence from Bruner II, Hensel and James (2005:4-139), Assael (2004:218) and Schiffman and Kanuk (2004:259) suggests that there are as many scales to measure consumer attitude as there are scholars.The implication of this is that the construct of consumer attitude has not been measured consistently.The Tricomponent attitude model is the basic model of consumer attitude.It specifies that consumer attitude comprises three components, namely beliefs, feelings and intentions to buy.This model, however, has been developed in developed countries (Assael, 2004:216;Schiffman & Kanuk, 2004:256;Solomon, 2002:200).Calls have been made to conduct more consumer behaviour studies in developing and emerging economies because most of these studies have been conducted in developed nations.This makes the generalizability of findings to developing countries questionable.More so, developing countries now present more business opportunities than ever before, with developed countries consisting of an ever-shrinking part of the global economy (Opoku & Arkoli, 2009:350;Klein, Ettenson & Krishnan, 2006:305).This study, therefore, assesses the reliability, validity and dimensionality of consumer attitude towards imported poultry products in Zimbabwe-a developing market-using the Tri-component attitude model as the basis for the research model.The decision to consider imported poultry products in Zimbabwe was motivated by the fact that the country has received a great deal of imported poultry products mainly from Brazil, South Africa and USA since 2009 after the adoption of multi-currencies (Irvine's, 2012:22).

Definition and scope of consumer attitudes
Solomon (2002:527) describes an attitude as a lasting, general evaluation of an object.An object may refer to people (including oneself) or issues.Arnould, Price and Zinkhan (2002:459) submit that an attitude is simply "a summary of consumer thoughts, feelings and actions".Blackwell et al. (2006:375) view an attitude as a global evaluative judgement of products or brands-a favourable attitude towards a product denotes that the person likes the product or brand while an unfavourable attitude denotes that the person does not like it.In the context of consumer behaviour, Assael (2004:216) and Schiffman and Kanuk (2004:253) view an attitude as a favourable or unfavourable tendency that directs the behaviour of consumers towards certain objects.In the consumer-oriented definition, Schiffman and Kanuk (2004:253) assert that the word object refers to something specific in consumer behaviour such as a product, product category, brand, and an advertisement.Thus marketers tend to be object-specific when conducting marketing research.In that case, it is difficult to observe attitudes directly.Therefore, attitudes must be inferred from the actions of the consumers or from what they say (Schiffman & Kanuk, 2004:253).Schiffman and Kanuk (2004:253) identify three characteristics of attitudes that emerge from the definitions of attitude.First, attitudes are a learned disposition.The implication of this is that attitudes are learned.Consumer attitude towards a product is formed as a result of the consumer's direct experience with the product, exposure to the marketer's communications and word-of-mouth information.An element of motivation is present in attitudes i.e. consumer attitudes may propel or repel the consumer from a particular product (Schiffman & Kanuk, 2004:253).Second, attitudes are relatively consistent with the behaviour that they reflect.However, Schiffman and Kanuk (2004:253) argue that this does not mean that attitudes are permanent.Attitudes may change with time.Third, attitudes occur within a situation.It is possible for an individual to have various attitudes toward a product or brand based on different situations.Therefore, when measuring attitude, the marketer should pay particular attention to the situation in which behaviour is taking place, otherwise a misinterpretation of the relationship between attitude and behaviour may result (Schiffman & Kanuk, 2004:255).
The concept of consumer attitude has been given wide attention in marketing literature.An understanding of consumer attitudes is important in that it enables marketers to influence how consumers behave toward products or brands (Argyriou & Melewar, 2011:431).More so, attitudes play an important part in developing marketing strategy.Attitudes help marketers to identify benefit segments, develop new products, and formulate and evaluate promotional strategies (Assael, 2004:222).Similarly, Wilcock et al. (2004:56) submit that the knowledge of consumer attitudes is critical in that it enables marketers to predict consumer behaviour.

The Tri-component attitude model
The desire to understand consumer attitudes and their relationship with consumer buying behaviour has motivated psychologists to come up with models or theories that capture the underlying dimensions of attitudes (Schiffman & Kanuk, 2004:256).As noted by Arnould et al. (2002:459), attitude models explain how a consumer processes information that influence the choice processes.Included in the consumer information process are cognitions and emotions.Prominent consumer attitude models or theories discussed in literature include the Tri-component attitude model, hierarchy of effects model, attitude toward object model, attitude toward behaviour model, theory of reasoned action and the theory of planned behaviour.These models provide different perspectives on the attitude components and how these components are related.However, the Tri-component attitude model is the basic model of consumer attitude (Blythe, 2008:146-147;Assael, 2004:216-218;Schiffman & Kanuk, 2004:260;Arnould et al., 2002:462-464;Solomon, 2002:200;Verbeke, 2000:526;Conner & Armitage, 1998:1429).
A person's cognitions (beliefs) represent the first part of the tri-component attitude model.Cognitions refer to the knowledge and perceptions that are acquired by consumers by way of a combination of direct experience with the attitude object as well as related information from various sources (Schiffman & Kanuk, 2004:256).Assael (2004:216) and Schiffman and Kanuk (2004:256) agree that consumers hold beliefs about products or brands.For example, they may believe that the product or brand has certain attributes.They may also believe that specific behaviour will result in a particular outcome.Because of this, the cognitive component is understood to consist of consumer beliefs.Consumer beliefs represent "subjective judgements about the relationship between two or more things".That is, consumer beliefs about a particular product are the characteristics that they ascribe to it (Blackwell et al., 2006:375).The role of the marketer is to develop a vocabulary of product attributes and benefits and then use it to develop questions used to solicit consumer data that will lead to a sound understanding of the beliefs of consumers (Assael, 2004:217).Assael (2004:217), Schiffman and Kanuk (2004:257) and Solomon (2002:200) concur that the affective component is described as the feelings or emotions of consumers about a particular product or brand.They also view this component as the overall brand evaluation i.e. of the three components, only the affective component-brand evaluation-is central to the study of attitudes because it summarises the consumer's predisposition.They argue that beliefs are relevant only to the extent that they influence brand evaluations, which are the primary determinants of intended behaviour.While the consumer's beliefs about a particular product or brand are multi-dimensional because they represent the brand attributes consumers perceive, the affective component is one-dimensional.Therefore, the overall consumer brand or product evaluation can be measured by rating the brand from 'poor' to 'excellent', from 'prefer least' to 'prefer most', from 'good' to 'bad', or from 'favourable' to 'unfavourable' (Assael, 2004:217;Schiffman & Kanuk, 2004:257).Brand evaluations are a product of brand beliefs i.e. affect-laden experiences manifest themselves in the form of emotionally charged states such as surprise or happiness, etc. (Schiffman & Kanuk, 2004:257).The conative component is concerned with the intention to buy, which is also known as behavioural intention (Assael, 2004:217-218;Schiffman & Kanuk, 2004:258).The conative component is the consumer's tendency to act towards a particular product or brand.In some interpretations, conation may refer to the behaviour itself (Schiffman & Kanuk, 2004:257).However, Assael (2004:218) opposes this view.He argues that sometimes beliefs and evaluations of a particular product do not have to change for consumers to establish an intention to buy particularly if the economic inducement is large enough, for example, a sharp reduction in price.It is important to measure the behavioural intention when developing marketing strategy.The behavioural intention can be used by marketers as the closest substitute of the actual buying behaviour (Assael, 2004:218).

Sampling and data collection
A cross-section of 400 consumers was taken in Harare and Bulawayo, the two major cities in Zimbabwe, using a structured interview guide.Consumers were intercepted in supermarkets as they carried out their shopping.Only 305 questionnaires were completed and usable i.e. 175 consumers from Harare and 130 consumers from Bulawayo.

Research instrument
The questionnaire was designed based on the Tri-component attitude model in order to capture the three dimensions of attitudes, namely beliefs (cognition), feelings (affect) and intentions to buy (conation) as proposed by Assael (2004:216), Schiffman and Kanuk (2004:256) and Solomon (2002:200).Several scales and insights from various authors (Northcutt, 2009;Brewer & Rojas, 2008:5;Bruner II et al., 2005:4-139;Assael, 2004:218;Schiffman & Kanuk, 2004:259;Groom, 1990:205) were considered in order to come up with a holistic scale that measured the three components of consumer attitude with regards to imported poultry meat products.A 15 item scale was developed i.e. each of the three components was measured by 5 items (see Table 1).A 7 point ordinal Likert scale was used to measure the three components i.e. items ATTI1-ATTI5 measured consumer beliefs (cognition), items ATTI6-ATTI10 measured consumer feelings (affect), and items ATTI11-ATTI15 measured consumer intentions (conation).

Data analysis procedures
The first step was to clean data.Then, data were coded and entered into SPSS Version 21.This was followed by exploratory factor analysis (EFA) so as to validate the components underlying the construct of consumer attitude.Cronbach's alpha was then used to determine the reliabilities of the sub-scales measuring the three components.Lastly, confirmatory factor analysis was performed in AMOS Version 21.

Sample profile
The study sample consisted of 54.4% females and 45.6% males.Age was distributed as follows: -1.3% for consumers aged 18-19, 18.4% for consumers aged 20-24, 26.2% for consumers aged 25-29, 30.5% for consumers aged 30-39, 20.0% for consumers aged 40-49 and 3.6% for consumers aged 50-59.The distribution of the highest level of education attained by the respondents was as follows: -0.3% for those who attained Grade 7, 5.9% for those who attained ZJC, 6.9% for those who attained O-Level, 10.8% for those who attained A-Level, 32.1% for those who attained Diploma, 33.4% for those who attained Bachelor's degree, 9.2% for those who attained Master's degree and 1.3% for the Doctoral degreed.The monthly income was distributed as follows: -less than US$500 was earned by 42.0%, US$500 to 999 was earned by 28.2%, US$1,000 to 1,499 was earned by 8.0%, US$1,500 to 1,999 was earned by 5.9% and at least US$2,000 was earned by 5.9%.

Exploratory factor analysis
Exploratory factor analysis was executed in order to validate the factors underlying the construct of consumer attitude.There was a need to test for the adequacy of the sample and also to test whether or not data permitted factor analysis.To test for sampling adequacy, a Kaiser-Meyer-Olkin (KMO) Measure of Sampling Adequacy was used.The KMO statistic was computed as 0.917.This indicates that the sample was very adequate (Field, 2005).The Bartlett's Test of Sphericity was also applied as an assessment of whether or not factor analysis could be executed on the data.As recommended by Field (2005), data permitted factor analysis to be performed (Chi-square = 2712.073;Degrees of freedom = 105; significant at p 0.001).The results of EFA are summarised in Table 2.
As illustrated in Table 2, three factors were extracted from the data.Items ATTI11-15 were linked to the consumer purchase intentions toward imported poultry products; hence this factor was named 'intentions'.Items ATTI1-4 explained the beliefs that consumers held about imported poultry products; hence the factor was referred to as 'beliefs'.Items ATTI5-10 except ATTI8 explained the feelings consumers had with regards to imported poultry products; hence the factor was called 'feelings'.Item ATTI8 was deleted because it could not load onto any of the three factors.Item ATTI5 was interpreted as a feeling instead of a belief.The reliabilities of the subscales measuring these items were assessed using the Cronbach's alpha.All the subscales displayed acceptable levels of reliability (Cronbach's 0.70) as indicated in Table 2 (Zikmund, Babin, Carr & Griffin, 2010;Bryman, 2008).

Confirmatory factor analysis
Having extracted three factors, namely intentions, beliefs and feelings from the data, the next step was to assess the construct validity of the dimensions of consumer attitude.Confirmatory factor analysis was therefore executed in AMOS.The results are presented in Table 3 and Figure 2. As illustrated in Table 3, the model fit indices were highly acceptable (CMIN/df = 1.820;GFI = 0.951; AGFI = 0.922; NFI = 0.955; TLI = 0.972; RMSEA = 0.052).The decision to accept these fit indices was based on the recommendations by Hooper, Coughlan and Mullen (2008) and Reisinger and Mavondo (2007) that CMIN/df should be less than 3 and that GFI, AGFI, NFI and TLI should be close to 1. Hooper et al. (2008) also recommend that the value of RMSEA range between 0.05 and 0.1.
Model estimates were also assessed to determine their significance.The model estimates were all significant.As illustrated in Figure 2, all items loaded significantly (p 0.001) on the latent variables and were sufficiently larger than 0.4 (Field, 2005).The critical ratios (C.R.) ranged between 7.076 and 19.184; thus were sufficiently large for each path.All correlations between latent variables were high and significant.The correlation between beliefs and feelings was estimated at 0.75; significant at p 0.001, C.R. = 7.948.The correlation between feelings and intentions was estimated at 0.70; significant at p 0.001, C.R. = 7.808.The correlation between beliefs and intentions was estimated at 0.66; significant at p 0.001, C.R. = 8.419.These findings indicate convergence validity of the latent variables, even though Zimbabwean consumers still regard them as different.

Discussion of Results and Implications
The study sought to assess the reliability, validity and dimensionality of consumer attitude towards imported poultry meat products in Zimbabwe-a developing market.The three dimensions of the Tri-component attitude model, namely beliefs (cognition), feelings (affect) and intentions to buy (conation) as suggested by many authors (Assael, 2004:216;Schiffman & Kanuk, 2004:256;Arnould et al., 2002:462;Solomon, 2002:200) were confirmed as valid and reliable measures of consumer attitude in a developing country.This implies that consumer beliefs, feelings and intentions to purchase are important factors that influence the decisions of consumers when purchasing important poultry products in developing countries such as Zimbabwe.
Zimbabwean consumers hold beliefs that imported poultry products have attributes and that certain behaviours will lead to specific outcomes (Assael, 2004:216;Schiffman & Kanuk, 2004:256).This can be explained by the fact that beliefs are subjective judgments about the relationship between two or more things (Blackwell et al., 2006:375).Assael (2004:217), Schiffman and Kanuk (2004:257) and Solomon (2002:200) concur that the feelings or emotions of consumers result from brand beliefs.As indicated by the positive and significant correlation between beliefs and feelings, the beliefs of Zimbabwean consumers with regards to the products are likely to influence their feelings toward the products.The conative component is concerned with the intention to buy.It leads to behaviour or purchase action.It is the consumer's tendency to act toward an object, or the intention to buy (Assael, 2004:217;Schiffman & Kanuk, 2004:258).The significant and positive correlation between consumer feelings toward and consumer intentions to buy imported poultry products suggests that the favourable the feelings of consumers, the more they are prepared to buy imported poultry products; the unfavourable the feelings of consumers, the more unlikely they are to buy imported poultry products.Consumer beliefs about imported poultry products are also positively correlated with intentions to buy imported poultry products.This suggests that consumer intentions to buy imported poultry products may also be influenced by the beliefs that consumers hold regarding imported poultry products.The significant and high correlations between consumer beliefs, feelings and intentions to purchase suggest convergence validity i.e. these components converge on measuring consumer attitude.However, Zimbabwean consumers still consider these components as distinct because the correlations are not extremely high.
A contribution to marketing literature is made through an enriched understanding of the dimensionality of consumer attitudes.As such marketing practitioners and researchers are advised to consider the three dimensions of consumer attitude when measuring consumer attitudes toward imported poultry meat products in developing markets such as Zimbabwe.

Table 1 .
Items Used to Measure Consumer Attitudes toward Imported Poultry Products

Table 2 .
Results of Exploratory Factor Analysis

Table 3 .
Model Fit Indices