Case 2: Bakhill Foods Group 1 R99741002 providers to research a Yu-wen R99741006 supplier Kenichi Lin Yun-chiao The the

R99741010 supplier Kenichi Liu plantago containing the R99741052 supplier Kenichi Chen Ying bamboo The B96701120 BBA four Pan Y H. Fang Groups following is an analysis of the structure of the case, we will analyze the case of experimental processes and methods and make recommendations, final conclusions, integrated back A case Q1 and Q2. Experimental purposes, The pretest hypothesis and types Sample selection Experimental Design Analysis of Results

Advertising under test design First, the purpose of the experiment, hypothesis and types (1) The purpose of the experiment Finding subjects a) test the way b) Scale Attractiveness and consumers find advertising spokesperson for the causal relationship between advertising attitude, in addition, taking into consideration both product attributes, advertisers and advertising agency for the endorsers gender selection concerns, and therefore also deemed to affect the gender of the endorsers experiment of one of the variables (shown at right). (2) the experimental hypothesis

Advertising spokesperson attractive vs unattractive Spokesperson gender

Consumer attitude Positive vs negative According to Michelle Gill, the hypothesis of the experiment for the physical attractiveness of the endorsers will affect the attitude of the consumer advertising; spokesperson appearance to attract consumers, consumers have a positive attitude towards the ad. (3) experimental species This experiment is designed as a true experimental design, emphasized extracted samples and "random" "random" distribution treatment expectations through random (randomization) experiment can finished than able to achieve external validity, this experiment, and there is no experimental group (experimental group ) and control group (control group). Second, pre-test: The purpose of the experiment is to find four model (Jude Law The female, unattractive female, Jude Law The male, unattractive male) as measured advertising spokesperson. The experimental method see the model image to the subjects (10 males, 10 females), issuing the degree of the questionnaires model attractive. Sampling method is convenient sample, internal staff or relatives likely to advertising companies. The seven-point scale questionnaire, observation tools required for the model selected based on the average scores for the highest score and the lowest score. The Comments and suggestions 1. Instrumentation (1) for a single Model tested only seen a single photo, but the photo quality, the camera angles are affecting the participants' perception and do not necessarily reflect the attractiveness of the Model I photo shoot quality, angle and other variables, should strive for the same. (2) 20 Model itself with the popularity of the Model, may also affect the judgment because of the stereotype. Do not look for celebrity endorsement can avoid subjects to affect the outcome of the likes and dislikes of celebrities. 2. Selection Bias convenient sample is not representative samples for internal staff, aesthetics may be different from the general public. four Beautiful / handsome Top Model sources can vote directly from the Internet access more convenient sample with credibility and kind. The larger space. Third, the selection of the sample

Sampling methods for random sampling of 96 people (48 male, 48 female) from the New York City phone book and give remuneration and traveling expenses. The Comments and suggestions 1. Frame error (1) New Yorkers are more fashionable than people in other states, aesthetics may be different from the other people of the state. (2) published phone mostly for low-income earners, and not representative of the overall sample. consider excessive cost subjects, selected by the National states had such cases the sampling should gather demographic variables (such as New York City published characteristics are citizens of the phone), to analyze the impact of the results, and then decide to exclude or incorporate. 2. Selection Bias (1) to give reward as well as traveling expenses, although more serious subjects can Tianda. However, in the high income earners may not have time to participate, so that the participants are mostly poor results due measured by a different standard of living, and cause different aesthetics and affect. (2) Due to the the measured small number of uncontrolled rater age, living habits and other variables, unfortunately recruited by measured by age high or low, or even find not drink coffee habits of consumers, also affect the results. standard of living, drink coffee habits and brand attitudes may affect the results of the tests should investigate these variables. But it is in the quiz The former investigations may result in the Interactive Testing Effect, it is suggested that the kind of problem to the exam. Fourth, the experimental design This experiment 2 (attractiveness: attractive vs. Unattractive) x 2 (gender: female vs. Male) design group, a total of four ads. 48 male subjects and 48 female subjects were each divided into 12 groups, a group of four people, a total of 24 groups, each group of four people each to see which one in four tests ads plus three dummy ads. The Comments and suggestions Group design compared with the advantages of group design will not let the subjects were aware of Surveying content, but the drawback is that the subjects will be affected by the other three commercial interruptions. To the subjects were randomly assigned advertising, although you can compensate for the selection bias, but the sample size is not large enough, may produce biased. (For example, in this experiment only segmentation sex, not separate different ages, plus a people only see one tested advertising may all see that the majority of older persons in the case with an ad). first by matching the way the cycle can clustering of age, income level, and then randomly selected in proportion Watch a different ad. (A) test Before the start of the test, the test will do the Introduction, inform the answer after reading an advertisement for each fill out a scale, and must be "itself" to judge each ad can not be used in conjunction with other, also can not go back to answer, and the real subject sort of measuring advertising third. The Comments and suggestions 1. Ordering Effect

Measured advertising arrangement order may produce endures in sorting and time variables bias, thereby affecting the internal validity of the experiment. The subjects psychological may be subject to other advertising impact, resulting in unfair judgment standard, for example: that the first two ads are very boring subjects, so compared to the next, to see a third ad will suddenly feel very attractive, and the This judgment does not come from the attitude of the model advertising. In order to reduce the impact caused by the ordering effect of the experiment, when sorting, so the other three dummy ads Changes in the sequence, a total of six kinds of permutation. Therefore, 24 groups in each of the four groups (two male, two female) is a unit, each of the four groups will work together to see its In a sort of way. In analyzing the results, and will be able to measure the the ordering effect generated. 2. Maturation Over the test of time, the subjects may be impatient, bored, resulting in the latter half of the answer tone casual or a more negative attitude, these are not caused by advertisements feeling, however, is purely receiver emotional effects brought about by differences. control the length of the scale, so Its moderate and not too long, avoid subjects irritability. 3. Instrumentation Four ads to focus on consistent quality cases mentioned three dummy ads, so pay attention to that the ad quality should not be too gap to avoid subjects based on "ad quality" rather than based on the ads themselves have a different attitude to make Experimental results bias. wide Announce presented and must be designed to require no difference in quality, to ensure that the use of the tool with consistency. From the overall experimental design, advertising under test in the design must pay attention to other factors interfere with the purpose of the experiment. For example, a wide Divisions in avoid showing Bakhill Coffee brand, in order to avoid the subjects itself to brand attitudes and preferences will affect the views on advertising. advertising brand cover letters or substitute 1. (B) Scale to fill Subjects after reading each ad, you have to fill out a scale. Scale contains cognitive, affective and conative information. The Comments and suggestions 1. Scale staggered on the scale in the opposite direction Although ensures the subjects attentive and serious answer, but it is likely to cause confusion and erroneous error scoring. Although this problem does not lead to the inference error, but will continue to affect the subsequent analysis. recommended that positive words are from the same side of the start or Who is let all agree that the scale on the same side, but tests whether antonyms to seriously answer. 2 The lack of confirmation of the problem The experimental questionnaire after three questions, did not as a general questionnaire or scale, in order to ensure that the subjects to be true to fill their reaction irresponsible replies synonym ask again the same problem, it is difficult to ensure that the subjects fill A reliability at the time. extra few questions test reliability problems in the problem column. V. Analysis of Results

According to the information of the case, the results plotted as the table at right. This set of assumptions, the advertising company as an advertising protagonist select attractive male reason for these six projects, attractive male most times get a high score. The Comments and suggestions 1. Formation of the attitude of the three elements may be present interaction Cognitive Affective Conative From the terms of the purpose of the experiment with the hypothesis that this experiment would like to know whether there is a positive relationship between the "appearance" and "attitude" (That is, to find the attractiveness spokesperson will let consumers have a favorable impression of the product), and the attitude is "cognitive", "affective "," Conative "consisting of, interaction between these three elements may exist, if the individual judgments may result in bias. Judgment of the experimental results The group select attractive male Backhill Foods to consider not comprehensive enough, and interactive relationship may exist between the three elements especially the influence attitudes, so this set of recommendations Backhill Foods must explore the relationship between the three elements first and find The criteria for measuring attitudes (must be continuous data types), once after setting the criteria, can be done by the following statistical regression analysis: Let D1 representatives gender (D1 = 0 represents for boys, D1 = 1 on behalf of girls); D2 behalf of attractiveness (D2 = 0 on behalf of Jude Law The D2 = 1 on behalf of as unattractive); Y = the effectiveness of advertising; X = ad attitude; Can be listed in the following regression:, followed by regression analysis can find the best combination of the 2X2 matrix that is able to identify the most suitable of the model. VI Conclusion May be based on cost considerations, the experimental sampling more briefly, simply through a single phonebook, randomly selected subjects, did not consider the subject's life, such as age and drinking coffee becomes possible experimental results that could produce serious bias; addition, as measured by measured by the attitude questionnaire to measure "cognitive", "affective", "conative" three levels, however, three of the experiment did not interact relations into account. Furthermore, the group judge experimental results, Backhill statistical regression analysis, was not too arbitrary way to decide the model allows for more complete interpretation of experimental results, not To produce the wrong decisions. 1, however, also pointed out that the average person on the selection of the letters and preferences. If the adoption of letters of alternative brands, should pay attention to this point may have an impact for the results.