1、4英文翻译模板参考样本毕 业 设 计(英文翻译)译文内容Using MCA to Segment New Car Markets(用多元类别分析进行新车市场细分)译文出处Journal of Marketing ResearchAug70, Vol. 7 Issue 3, p360-363, 4p, 2 charts系 别:经济管理学院专 业:汽车营销班 级:T553-1学生姓名:晏子淇学 号:20050530124指导教师:魏仁干 Using MCA to Segment New Car MarketsBy: Peters, William H. Journal of Marketing R
2、esearch (JMR), Aug70, Vol. 7 Issue 3, p360-363, 4p, 2 charts; (AN 5002921)PIn this study the mid-196Os new car market is analyzed by using a new, dummy variable, multivariate technique known as Multiple Classification Analysis (MCA).This approach helps provide a better understanding of this complex
3、market and gives some interesting hints as to its possible future direction.INTRODUCTIONIn the years since Smiths article 3 aroused interest in market segmentation, there has been growing concern about the power of demographic variables to segment markets for many products. In fact, whether demograp
4、hic analysis is of much help at all in defining markets or understanding buyer behavior has been questioned. In 1964 Yankelovich 6, p. 84 wrote: In neither automobiles, soap, nor cigarettes do demographic analyses reveal to the manufacturers what products to make or what products to sell to what seg
5、ments of the market.More recently the pessimism about the usefulness of demographic data in market segmentation was countered by Bass, Tigert, and Lonsdale 2. They say that demographic variables have had such a poor performance record in past studies because these studies attempted to predict indivi
6、dual family buying behavior with demographics. On this basis, because of wide variations of buyer behavior in any market group, the multiple regression R values are almost always very low. But Basset al. write that: . . . implementation of the strategy of market segmentation involves postulates abou
7、t the characteristics and behavior of groups, not persons 2, p.265. They show how important buyer groups can be identified with demographics and thereby refute the idea that it is not feasible to define market segments by socioeconomic measurements. The essence of their argument is that For market s
8、egmentation, the essential question is whether it is possible to identify groups of consumers with different mean purchase rates dependent on certain variables, such as income, age, and occupation2, p. 265.This author feels that a new, easy to use, dummy variable multiple regression program called M
9、ultiple Classification Analysis (MCA) is ideally suited to perform this task efficiently. To present and explain the MCA pro-gram is the first objective of this article. The second is to put MCA to work analyzing the mid-1960s new car market and to show the results. These results allow for a much be
10、tter understanding of this major, complex consumer market. Also, they provide some useful insights into the possible future direction of the new car marketin particular a hint at a possible reason for the current popularity of smaller cars and what this may mean for the future of small and medium-si
11、zed cars.THE STUDYData Used The data in this study were collected by personal interview for the 1966 and 1967 Survey of ConsumerFinances (SCF) by the Survey Research Center at theUniversity of Michigan. They are combined here. The subsample in this study used the automobile owned as the unit of anal
12、ysis, with appropriate economic and demographic data about the family unit associated with each car. This subsample was all cars purchased new whose model year was 1963-67 (N = 2,051).VariablesThe dependent variable used throughout the study was the automobile class to which the car belongs, with fi
13、ve new car classes established generally on the basis of size: compacts, intermediate-sized, medium-sized, large, and foreign economy cars. The explanatory variable included: total family unit in come (all earned and unearned income except capital gains), age of the family head, heads occupation, fa
14、milys housing status (rent/own), number of major income receivers (over $600 a year), number of children under 18, number of cars owned, number of adults (over 18),education and race of the head, number of drivers, who Journal of Marketing Research, Vol. VII (August 1970), 360-3USING MCA TO SEGMENT
15、NEW CAR MARKETS normally drives, distance from center of central city, and house type. Total family unit income was expressed in terms of decilesthe lowest numbered one and the highest ten. MethodologyThe current version of MCA was developed in theearly 196Os at the Institute for Social Research at
16、theUniversity of Michigan . It is a complex computer program that makes dummy variable multiple regression easy. While MCA has been used as an analytical tool in other branches of social science research (especially at the Institute for Social Research), it has yet to become well known in marketing.
17、 It has potential advantages in exploring multivariate marketing problems, and especially for group-type marketing segmentation, as suggested by Basset al. The main advantages of MCA are:1. It makes deck preparation when doing dummy variable regressions very simple.。takes all the data coded into mut
18、ually exclusive categories (both nominal and interval scale predictors) and automatically converts each subcategory of each predictor variable into a 1-0 dummy variable.(It only uses one computer card column for each predictor variable, regardless of the number of subcategories. With conventional du
19、mmy variable regressions, the researcher has to write a special computer program to help do this time-consuming job, and even then one computer card column is used for each subcategory of each predictor variable. 2. MCA handles nonlinear data.3. The output is very easy to understand (a briefexample
20、follows). MCA gives the effect of eachexplanatory variable upon the dependent variablein two wayswith and without taking into ac-count the other explanatory variables. The chief disadvantage of MCA is that it cannotefficiently handle interaction effects. It assumes the absence of interaction.。While
21、the MCA program was designed to operate on an interval scale dependent variable, it can handle a one-zero nominal scale dependent variable, which, as in the following example, is useful for studying product ownership among market segments. Table 1 shows part of a typical table constructed from MCA o
22、utput during this study. In this example, the dependent variable was a one-zero dichotomy: one if the car owned was a compact and zero if it was not. (In the study a separate MCA regression was run for each car class.) The meaning of the important statistics in Table 1 is as follows:1. Grand meanThe
23、 output for the regression will show a grand mean for all the input data. In this example the grand mean is .1424, the 14.24 percent of the entire sample who own a compact car.2. Difference from grand mean, gross总体平均数离差值。This shows the unadjusted effect of being in a specific category of an explanat
24、ory variable. This gross difference is computed without accounting for the effects of the other explanatory variables.3. Difference from grand mean, netThis difference gives the effect of being in a given subgroup of the population after the effects of the other explanatory variables have been taken
25、 into account.。The net difference is the deviation of the subgroup mean from the grand mean that would occur when only the effect of membership in the subgroup affects the result. This gets at the true effect of being in that subgroup of the population. In the report of the results that follows, onl
26、y adjusted frequencies based on net differences that were at least significant at the .05 level are reported (significant adjusted frequencies).4. StatisticThis measures the explanatory variables (the entire variable with all its subgroups) ability to explain variation in the dependent variable afte
27、r adjusting for the effects of all other explanatory variables (not in terms of percent of variance explained). Those variables with higher /8 levels are better predictors.RESULTS AND DISCUSSIONBefore the MCA results are presented it should benoted that market segmentation via the traditionalJOURNAL
28、 OF MARKETING RESEARCH, AUGUST 1970 method of cross-tabulating the explanatory variables one at a time against the dependent variable (car class) was attempted with data. Without showing the details of this procedure, this analysis indicated that separatemarket segments for the five classes of new c
29、ars cannot be clearly differentiated using such explanatory variables as income, age of head, occupation, number of children, etc. The only new car market segment that did lend itself to fairly clear differentiation from the whole was for large cars. The mass market, however, could not be separated
30、so that clear market segments, in terms of these demographic variables, were distinguishable for compacts vs. intermediates vs. medium-sized cars vs. the foreign economy class.。From the point of view of this type of analysis, the new car market seems to be a very uniform one.。The next step in the an
31、alysis was to do the MCAregressionsa separate regression for each car class with the value of the dependent variable being one if the car owned was in the class being studied and zero if it were not. The first thing one does with such results, of course, is to look at the R- values to see the joint
32、effect of the variables. As in the previously published work where demographics were used as predictors of buyer behavior, the R values were all very low. None of them went above .05 for the new car dataBut now to the more interesting results from theMCA analyses the true effect of each explanatory variable upon new car ownership as expressed by the net differences of the subgroup means from the totalownership mean of the entire population. The effect on the grand mean probability as expresse