(2 rânduri libere, 11p) Studies and Scientific Researches. Economics Edition, No 19, 2014 http://sceco.ub.ro INSIGHTS INTO QUALITATIVE VARIABLES ASSESSMENT IN MARKETING RESEARCH Laura Țimiraș “Vasile Alecsandri” University of Bacău lauratimiras@yahoo.com Abstract In this paper we present a number of issues to be taken into account in assessing the marketing qualitative variables. Thus, the opinions, the preferences, the attitudes, etc. of the consumers are qualitative variables whose measurement requires the use of different scales presented in the literature and, where appropriate, the researcher must develop scales adapted to the particularities of the study undertaken. Using a certain scale it is not random action. The literature presents both comparative methods and non-comparative scaling methods. Each of these categories generates certain types of information, and also they are complementary in the evaluation of various products, brands, organizations etc. Thus, if in a non-comparative scaling method can get information about how a product is evaluated (favorable or unfavorable, for example) by the respondents, comparative scaling method allows us to determine where that product is in a series of investigated competing products. Another aspect to be taken into account in the construction of the scales is the number of levels used. It is intended to obtain information with high degree of detail, but without the risk of increasing the rate of non-response due to inability of respondents to make assessments through scales with too many levels. Finally, the expression used to collect information from respondents is essential in obtaining accurate and comparable information, with the emphasis on avoiding ambiguity in drawing scales. Keywords marketing research; qualitative characteristics; scales JEL Classification M31 Introduction The study of consumer behavior is one of the areas of study of prime importance of marketing research. Thus, in marketing research the evaluation of qualitative variables - attitudes, expectations, intentions to buy, opinions, preferences, requirements, reasons and criteria considered in choosing the products, brands, retail units etc. and their importance – occupy a significant share of total variables investigated. Normally, the presented variables are evaluated using nominal scale which allows only the classification of investigated subjects / stimuli, often insufficient for the researcher in the study of marketing phenomena. Accordingly, preference for certain competing products can be assessed on a nominal scale dividing those products into two categories - those that are preferred and not preferred by respondents. But researcher wants to ensure greater product differentiation, depending on the intensity of respondents' preferences. For that one can build its own measuring instruments (scales). Thus, through scaling can be achieved hierarchy or can be measured "psychological distance" between the investigated stimuli depending on the variables that apparently only allowed making classifications. The marketing literature presents a number of methods of scaling that can be used to assess qualitative variables, but the researchers are able to build their own scale based 236 Țimiraș on objectives, level of detail of information they wish to obtain, the characteristics of the sample, respondents' ability to evaluate certain stimuli on a scale with more or less levels determined by the awareness of subjects regarding investigated product etc. Scaling methods in marketing literature Over time various scale methods have been developed to assess qualitative variables in order to better adapt to the specific investigated problems and increase the accuracy in measuring the investigated variables. Some of the most common methods are summarized below. Semantic differential is a scale with typically 5 or 7 levels inserted between two bipolar attributes (e.g., from very favorable to very unfavorable). Investigated subject is asked to indicate on the scale the level corresponding to his opinion about the investigated stimulus. Likert's scale requires the respondent expressing agreement or disagreement with a series of statements made with respect to the investigated issues. Stapel's scale is a scale of ten levels, five with plus sign and five with a minus sign, where investigated subjects are asked to indicate the level on the scale that best fits their opinion on the investigated stimulus. The Fishbein-Rosenberg model is a complex scaling method that enables the evaluation of investigated stimuli taking into account the attitude of the investigated subjects in terms of investigated stimuli from the perspective of each evaluation criterion and the importance attributed to these criteria in forming attitudes toward studied stimuli. Constant sum scale is a scale that involves dividing by the researched subjects of a certain amount (usually 10 or 100) between investigated stimuli, according to their own judgments about them. Rank ordering method aims to achieve hierarchies of investigated stimuli. In this sense, all stimuli are presented to the investigated subjects and they are asked to rank according to their own assessment. Paired comparisons method is also used to rank the investigated stimuli. The studied stimuli are presented in pairs to investigated subjects, asking them to indicate which of them they appreciate more, being considered "probably the easiest and most reliable method for ranking" (Cohen, 2009). If the first four presented methods involve evaluating each studied stimulus (product, brand, organization, etc..) independently of the other (non-comparative scaling methods) constant sum scale, rank ordering method, paired comparisons method are part of comparative methods of scaling, results (hierarchy of stimuli investigated) having significance only in comparative context. Using different scales depending on the particular research As we mentioned above depending on the particular research one can use one or more of the listed scaling methods or also can build other scales to meet the requirements of the study. It should be noted that comparative scaling methods can be complementary to other methods; the use of both comparative and a non-comparative scaling methods may lead to a deeper analysis of assessments/preferences/opinions to different stimuli. Thus, non-comparative scaling methods involve evaluating each stimulus independently of others, while comparative methods involve evaluating stimuli through comparison (as their very name indicates). Considering the valuation of consumer assessments of certain competing products, non-comparative scaling methods allow, for each product, to know whether these 237 Studies and Scientific Researches. Economics Edition, No 19, 2014 http://sceco.ub.ro respondents have favorable or unfavorable evaluations, as appropriate. Further comparison can be made between goods which have been subject to investigation at different times. Thus, if a research result for a given product A discovers unfavorably evaluation of the consumers, and in another study for a competing product B consumer evaluations are favorable, the correct interpretation is that the product B is positively appreciated by the consumers and the product A is not appreciated, product B being situated in a better position than product A in the reference market in terms of consumer evaluations. On the other hand, comparative methods allow the knowledge of the hierarchy of products in terms of respondents' assessments, whether they are favorable appreciation or unfavorable. More specifically, the first place owned by a product among competing products doesn’t mean that the assessments of respondents are necessarily favorable to it (the product can hold first place between products that respondent did not appreciate) and n-th place between n compared products does not mean that product is not appreciated by the respondents (it can hold the last place of n highly rated products by respondents). The results obtained by applying comparative scaling methods can be used only in the context of generated compared context. Accordingly, whether through a research were studied three competing products A, B, C, and the generated hierarchy is: first place - B, second place - C, third place - A and in another research were studied other two competing products, D and E with the following hierarchy: first place - D, second place – E, based on the results of the two research one cannot compared together the results regarding the five products (A, B, C, D, E) and cannot draw conclusions like: product A (third place obtained in the first research) is less appreciated than the product D (first place obtained in the second research), each of these products having other items of comparison. Also, if the products were evaluated with the help of a non-comparative scaling method, the results for each of the five products might point unfavorable or favorable assessment, comparative scale methods not being able to provide such information. The combined use of comparative and non-comparative scaling method is particularly useful in cases of evaluated products where the assessments of respondents are relatively close. Thus, starting from the evaluation of three brands in terms of consumer appreciation using, for example, a 5-level scale - from very unfavorable (score 1) to the very favorable (score 5) (semantic differential), there is a possibility that all 3 brands to receive from a respondent "score" 5 (actually proving very good reviews from the respondent for all three stimuli). However, the respondent consumes one particular brands. This demonstrates that although all three brands were rated with "very favorable", one of them stands out from the other two in terms of respondents' preferences. In this context, using comparative scale methods can determine which brand is preferred, which is the one that precedes it etc. Thus, requiring, for example, from respondents to assess simultaneously the three brands and establish their hierarchy in terms of preferences (rank ordering method) one will get results that will attest the brand that comes first, which is the second and which is on the third place (which is not contrary to the result obtained by non-comparative scaling method that revealed very positive feedback for all brands, especially since in this case non-comparative scale differentiation capacity is relatively low ). Regarding the number of levels of the assessment scale, the researcher will always want to obtain information with a greater degree of detail (using scales with large number of levels). Of course in this case we refer to non-comparative scales category. On the other hand, too many levels of scale sometimes make the respondent tired and even make him unable to indicate the corresponding level on the scale for his likes/opinions/preferences etc. More specifically it is easy to specify whether or not we consider a product (two-level scale), being relatively more difficult to "point" on a scale with five levels (from level 1 - "very unfavorable" to level 5 - "very favorable") 238 Țimiraș and is even difficult to indicate a certain level on a scale with 10 levels. In this case, however, there are differences from one category of respondents to another. There are respondents who know in greater measure the product thus making it easier to achieve its evaluation on a scale with a larger number of levels, however, respondents who know to a lesser extent a rated product will be placed in a difficult situation where ask to use scales with number high of levels. The major differences in this regard are recorded in business markets and respective in consumer markets. If in the case of business markets, potential customers know in detail the product and may evaluate each characteristics of it, even on a scale with the high number of levels (for example, Stapel Scale), in the consumer markets, goods are little known in particular in terms of their characteristics. It is thus advisable to use scales with a small number of levels (e.g. up to 5 levels) in the evaluation of consumer products. On the other hand, the degree of knowledge of end consumer product in the market depends on the nature of the product. Depending on the degree of involvement in the purchase, frequency of use, the nature of need they meet etc. characteristics of end consumer products can be known in a greater or lesser extent by them. Finally, there are types of consumers that are characterized by the desire to be informed and involved in studying in detail the offers of the market, this category being therefore able to evaluate various characteristics of products using scales with several levels, and uninformed consumers who most likely will not make judgments on such a scale, in the best case being able to classify products using scale with two levels: yes/no, I prefer/I do not prefer. In the present context in assessing qualitative characteristics, the researcher will consider the construction of scales with the number of levels that can be used to collect data in a more accurate and complete manner - thus leading to a balance between the desire of the researcher to obtain detailed information and the ability of the respondents to give it. Special attention in the construction of the scales that will be the basis for assessing various marketing phenomena should be given not only to the number of levels of scale but also to how to request information from the respondents. The modality used in asking for data in order to avoid ambiguity is crucial in the approach of collecting the information that reflects reality. For example, considering the importance of the "price" feature in the decision to purchase a particular product X, one may choose to use a 4-level scale from "very important" (4) to "unimportant" (1). In this case, there is a risk of confusion, i.e. the same value on the scale has different meanings for respondents. Considering for example value 4 on the scale, it could be indicated by both the respondents who are very sensitive to price, without sacrificing quality (the category of those seeking the best quality at the lowest price), and by those devoted exclusively to the price even at the expense of quality (accept a lower quality product for a low price). An alternative to this scale could be much clearer formulation on investigated matters, asking the respondents to express their agreement or disagreement (Likert's scale), for example: "I always buy superior products in terms of quality but looking at the sales price ","I buy cheaper products even if their quality is lower ", etc. In the same manner, if you want to determine whether respondents are loyal to a particular brand is preferable to avoid questions like "are you loyal to brand X?, a consumer “yes” answer being given because it consumes only brand X, or where appropriate, consumes most common brand X. Therefore, to avoid confusion one can use statements such as "I always buy brand X", "If I don`t find brand X I postpone buying", "If I cannot find another brand I buy brand X" etc. respondents being asked to indicate agreement/disagreement to these statements. In this way we can clearly identify the category of those who consume only studied brand, those who prefer to consume the brand, but if, for example, is not available they opt in for another brand. 239 Studies and Scientific Researches. Economics Edition, No 19, 2014 http://sceco.ub.ro The language used also has to be adapted to the understanding of the respondents, technical language, strictly specialized, being used only for research on organizational consumer. In conclusion, measuring qualitative variables requires a process of comparison and a choice of between several types of scales corresponding to the necessity of ongoing research. You can combine both comparative and non-comparative scaling methods in order to know the consumer options to different stimuli and to generate hierarchies in light of the findings, especially when evaluation of stimuli is difficult by using the non-comparative scaling methods. The number of levels of the used scales will be determined according to the knowledge of the respondents' ability to differentiate higher or lower levels of stimulus intensity that have certain properties. Thus, given characteristics of the business market, one may be used for his evaluation a scale with higher number of levels (i.e. 1 to 10), but the researches in end consumers market will use scales with smaller number of levels, corresponding to the lower awareness of the full features of the product. Finally, in order to obtain accurate, comparable data, in the evaluation of qualitative variables should be avoided ambiguity in the expression used to collect information and should be used of an accessible language to all categories of respondents. References Aaker, D. A., Kumar, V., Day, G.S. (1998), Marketing research, 6th Edition, John Wiley & Sons, Inc. Ajzen, I., Fishbein M. (2008), Scaling and Testing Multiplicative Combinations in the Expectancy–Value Model of Attitudes, Journal of Applied Social Psychology, 38, 9, 2222–2247. American Marketing Association Dictionary (2014) on the webpage https://www.ama.org/resources/Pages/Dictionary.aspx Boier, R, Ţimiraş, L.C. (2006), Cercetarea de marketing, Iaşi Editura Performantica Cătoiu, Iacob (coordonator) (2002), Cercetări de marketing, Editura Uranus, Bucureşti. Cohen, E. (2009), Applying best-worst scaling to wine marketing, International Journal of Wine Business Research, Volume 21, Number 1, 8-23. Gilmore, A., McMullan, R. (2009), Scales in services marketing research: a critique and way forward, European Journal of Marketing, Volume 23, Number 5/6, 640-651. Kumar, V., Aaker, D. A., Day, G.S. (1999), Essentials of marketing research, John Wiley & Sons, Inc. Malhotra, N.K. (2004), Marketing Research an Applied Orientation. Fourth edition, Upper Saddle River, New Jersey, Pearson Education. 240 In this paper we present a number of issues to be taken into account in assessing the marketing qualitative variables. Thus, the opinions, the preferences, the attitudes, etc. of the consumers are qualitative variables whose measurement requires the u... Scaling methods in marketing literature