"Consumer Behaviour is an essential study because the more you know your consumer the more you can earn within that particular market. (Woodruff and Gardial, 1996) This essay discusses Goldsmith et al (1997) 'The Impact of social values on food-related attitudes'.

It consists of a short review concentrating on the main points of each section followed by critical remarks on strengths and weakness of the study and a conclusion.The introduction familiarizes the reader with the author's goal: to investigate the relationship between self reported values, attitudes and consumption (behaviour) for low-involvement food varieties. The authors decide interviewing consumers using self administered questionnaires is the most efficient method. The background describes previous studies relating to this research. Findings show a relationship between social values and purchasing new clothes, (Rose et all, 1994).

Another study described shows associations between social values and the use of natural foods (Homer and Kahle, 1998)The method is split into four sections; the convenience sample is described followed by the structured interview methods used to obtain the values, attitudes and behaviours of women shoppers relating to food issues. Values were obtained through a seven point response format; this could be considered 'positivist', a reductionist approach putting human values onto a numerical scale, critics suggest positivism falsely represents human social action, (Max Horkheimer, 1937) but it is advantageous also because it reduces interview variability; there are set possible answers in comparison to open ended questions, (Bryman and bell, 2003).However, in this research the order of values listed was rotated each interview to avoid bias from order effect, meaning participants did not receive identical, standardized interviews which subsequently reduces validity. Duncan and Schuman, (1977) have suggested 'changes in question order might alter the patterns of relationship between the response given to questions as well as altering the frequency of particular answers' this suggests the question order should have been planned more thoroughly (S. G McFarland, 1981).Food attitudes and food shopping were split into categories; 'pro snacking', 'convenience shopping' and 'likes cooking'.

To measure purchase 11 food-categories was read to participants who indicated how often they purchased each using a four point response format. Once again this point-response is reductionist. The researcher could have asked open ended questions where more detail could have been collected to overcome this problem. Results confirm an association between attitude and behaviour existing to be stronger than the association between values and attitudes/behaviour.Despite this research being important I believe there could be ways it could be improved.

Firstly, the sample used women experienced food shoppers of all ages and races; this is advantageous as no population bias existed however, men were excluded, some might consider it sexist to ignore men being experienced food shoppers. The sample used a minimum age of 24 years which ignores younger age groups, I can personally guarantee university students living away from home (18-22 years) are experienced food shoppers.To overcome this problem the age range of the sample could be lowered, one might find using a younger age supports existing findings of the association between attitudes and behaviour; it is likely students would have 'convenience shopping' attitudes thus buy 'convenience food' items. Further research with younger age ranges might be useful providing insight into food values of adolescents; psychological changes of adolescent's such as their search for independence, identity and concern for appearance can have a strong impact upon food choices, (Mary Story, 1986).Adolescents are particularly concerned with body image through media and peer pressures, at any one time a large number of girls are on diets ranging from 25% (Leon et al, 1989) to 63%, (Rosen and Gross, 1987) which undoubtedly would have an impact upon food choice; broadening the age range might find stronger results between values and food behaviour, particularly 'sense of belonging' category because previous studies have shown adolescent girls 'believing being thinner would impact upon their lives,' (Paxton, 1990). This study is an American perspective, this limits our ability to generalise results worldwide.

The sample was taken in a small city; it is likely most women participants belonged to a similar social class/environment making the study bias and more specific in nature. A convenience sample creates bias as certain women might be more likely to stop and participate when shopping or be available at a certain time of day which may result in only a subset of the target population being represented (Zina O'Leary, 2004). An improved sample might be found in a larger city where there is scope for broader ranges of participants from different environments and age groups.A large sample was taken (323 people) but this might not be large in relation to the population size of the city.

The research method was interviews using self administered questionnaires. With this methodology the quality of interview is dependent upon the interviewers' skills; the author overcame this problem by training student interviewers prior. A disadvantage of interviews is that interviewer characteristics e. g. ethnicity, gender and social background have been found to influence answers, (Bryman and Bell, 2003).

Using postal questionnaires or telephone surveys instead could overcome this issue.The use of student interviewees could be questioned; students might encourage/discourage potential participants, this might explain why there were a greater number of twenty year olds used in the sample (40%). However, an interviewer is advantageous in explaining questions to participants if they don't understand preventing partially answered questionnaires and skewed variables, (Bryman and Bell, 2003). However, probing could create bias disadvantages, interventions could influence the respondents answer and are likely to differ across interviewers producing variations in respondent's replies, (Bryan and Bell, 2003).Schuman and Presser, (1981) support this further; they've shown small variations in wording questions can influence replies therefore this research could be improved by giving interviewers strict guidance upon probing.

Surveys are advantageous as they generate standardised, quantifiable, empirical data making it easier to generalise findings, (Zina O'Leary, 2004) however, relying upon standardisation forces the researcher to develop general questions which are minimally appropriate for all respondents, potentially overlooking what is more appropriate.The interview nature might prevent women revealing all information relating to their eating habits as this is a sensitive issue for some and could cause embarrassment. The quality of research could be improved by combining methods of research; self administered questionnaires and interviews which would enable richer data to be collected. These findings may be limited by the minimal selection of foods included in the study, this could be improved by broadening the selection used.Social values may more strongly influence other types of food, natural foods for example (Homer and Kahle, 1988) Taking everything into consideration, there are aspects of this research that could have been improved to produce more validity however, this article is of academic value to teach the role of attitudes in relation to consumer behaviour for low-involvement products.

It is useful to bear these factors in mind for marketing strategic, designing persuasive programmes relating to customers high valued end states.