Abstract

(ProQuest: ... denotes formulae omitted.)INTRODUCTION:Market players have several ways of obtaining information to be used in making strategic decisions regarding their products. One of which is the use of surveys. Marketers may choose between a census which enumerates completely every unit in a population and a sample survey which consists of just a part of them. Designing a full-blown statistical survey may prove to be difficult at times especially when no definite frame of intended respondents are available. In addition, this requires economical and organizational efforts that are seen, in many cases, as burdens to small firms. As a common practice, marketers simply send out questionnaires to all intended respondents, hoping they grant answer. However, many of them choose to disregard this request. Consequently, unreliable estimates become basis of the firm's marketing strategy if it fails to adjust for self -selection bias incorporated in the results of the census. Self -selection bias is observed when the unit under study is allowed to independently choose whether or not to participate in a census, determining some amount of non - responses in the process. The units which chose to participate in the census constitute a non - probabilistic sample.This paper intends to show the results of a study that explores the use of Propensity Score Matching (PSM) in eliminating self -selection bias in market surveys. The propensity score approach was first used by Rosenbaum & Rubin (1983) in observational resea rch to balance treatment and control subjects. Propensity score was then defined as a probability of exposure to a treatment given observed In the case of market surveys, the propensity score developed in this paper represents the conditional probability of being a self -selected participant given specific observed covariates. Although corrective strategies demonstrated in this paper is not encouraged particularly if there is a way for market researchers to use probabilistic samples , the study is made to introduce cautious measures if using convenience samples as base for making major o r even minor marketing decisions.THE PROBLEM OF SELF-SELECTION:Self-selection bias is observed when respondents are allowed to decide entirely for themselves whether to take part in a census or not. The units which chose to participate in the census constitute a non - probabilistic sample. To illustrate how this bias occurs, consider a finite population of N units. After the census operation, the population is basically divided into two groups: (1) participating group, and (2) non-participating group. This situation is shown in Figure 1.LetR = Number of participating unitsM = Number of non-participating unitsN = R+M = Population sizeR = R/N = Proportion of participating units in the populationM = M/N = Proportion of non-participating units in the populationThen, the population total and mean may be written as... (1) and... (2)where Y and YR are the total and mean of a variable of interest for participating units, respectively, whereas Y and YM are the total and mean of a variable of interest for non -participating units, respectively. If no compensation is made for non -response, the population mean will be declared as YR when in fact it should be YN . The self-selection bias is computed as,... (3)The preceding equation (3) suggests that the estimated' mean YR is approximately unbiased for the population mean YN if either M is small or the mean for participating units is close to that of the non-particip ating units. Since there is no way to control the difference between the means for participating and non - participating units, the only way to ensure that the selection bias is small is by increasing the rate of participation (or decreasing the rate of non -participation). However, this cannot also be done, since participation is assumed to be independently decided, particularly in marketing surveys. …

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call

Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.