3.

3.1.

Research method

Sample and data collection

A structured questionnaire was developed to collect data for this study. Using the snowball sampling technique, the

Downloaded by [101.87.164.135] at 05:14 07 April 2015

98 Y.-K. Seock et al.

questionnaire was distributed to Chinese female consumers, ranging in age from their 20s to their 50s, all residing in Beijing, China. In order to collect data, the researchers first contacted various local companies, factories, and retail stores in which Chinese females were employed, and a time was arranged to visit these places for the purpose of admin- istrating the questionnaires. When they visited the arranged places, the researchers asked the respondents to pass along the survey to their friends and acquaintances who may be interested or willing to participate in the survey. Of the total of 356 returned, 312 questionnaires were used for data anal- yses after eliminating those surveys that were completed incorrectly or missing too many responses. Of the total 312 respondents, 29.8% were aged in their 20s, 27.6% were in their 30s, 23.1% were in their 40s, and 19.6% were in their 50s. A vast majority of the respondents (82.7%) had jobs and more than half (58.3%) of the respondents were married.

3.2. Instrument

A structured questionnaire was developed to collect data on the variables in the study. Research variables included infor- mation sources, demographic characteristics, and socio- economic variables. In order to measure the respondents’ perceptions of the relative importance of information sources used for purchasing apparel products, 18 items were adapted from previous studies on information sources (Hwang, 2009; Kim & Lee, 2008; Ok, 2000; Park & Zhang, 2005). Information sources for purchasing clothing prod- ucts were classified into internal search (e.g. already known brand image, price of clothing, store image, and fash- ion information) and external search. External search was pided into two categories: either using a human infor- mation source (e.g. advice of friend or colleague, family or close acquaintance, salesperson or service provider, street or surrounding person’s clothes, and entertainers’ clothes) or using a non-human information source (e.g. clothing adver- tisement through Internet, TV advertisement, radio adver- tisement, fashion journal, newspaper, fashion show, outdoor advertisement, bus advertisement, sign, window/store dis- play, clothing laid out in store, and fashion company’s cata- logue). Respondents were asked to indicate how much they place importance on each information source when search- ing for information about apparel products. Responses were measured on a 5-point Likert-type scale (1 = not important at all, 5 = very important). The next section of the question- naire inquired about the respondents’ basic demographic and socio-economic characteristics such as age, marital sta- tus, education, occupation, household income, social class, and clothing expenditure.

Two different versions of a questionnaire were prepared utilising cross-cultural validation and back-translation pro- cedures. The questionnaire was first completed in Chinese and then translated into Korean by a bilingual Chinese

translator who speaks both Chinese and Korean. The Korean version of the questionnaire was then translated back into Chinese by a different bilingual inpidual and compared to the original version for consistency. Several items were modified to improve accuracy. Cronbach’s alpha was used to assess internal consistency reliability of the scales. Cronbach’s alpha coefficients of internal search, non-human information sources, and human information sources were 0.72, 0.61, and 0.70, respectively.

4. Results

Pillai’s Trace multivariate analyses of variance (MANOVA) were used to identify the differences of Chinese female con- sumers’ information search behaviour in shopping apparel products varied by socio-economic status. Pillai’s crite- rion was used to test for significant differences between groups because Pillai’s criterion is more robust with small or unequal sample sizes and against violations of the homogeneity of variance/covariance assumption than other multivariate test criteria (Olson, 1979). Tukey’s honestly significant difference (HSD) post hoc tests were used for pair-wise comparisons of the groups in order to determine which group(s) differ from each other.