Critical Review: Security Managers Report on Shoplifting - Assignment Example

Critical review: Security Manager’s Report on Shoplifting This paper gives a critical review of the Shoplifting Report, with a particular focus on the methodology used in the research. The first point discussed is whether the report gives a good enough outline of the research problem and aim. After this the quality of the data is discussed, that is the methodology, reliability and validity. The security manager’s conclusions are also discussed. The paper includes suggestions for how the research could have been improved. Research problem The research problem is given in broad terms by the statement that ‘this research will be about shoplifting’. This statement is valid. The security manager (who I will refer to as male for the sake of simplicity) then attempts to narrow the focus to his own specific shop. But this step is not well justified. He makes sweeping statements such that previous research on shoplifting ‘is probably not that relevant’ because it does not pertain to his own shop. He also writes ‘it is clear that most theft is committed by customers, and not much is committed by staff’. He gives no references to research that shows this and it comes across as being his own opinion. Aim of the research In this section, the security manager repeats the research problem (‘to study shoplifting by customers’) and this is unnecessary. He also then gives specific focus points which should have been included in his discussion of the research problem, namely ‘a number of aspects… including the amount of shoplifting, the influence of new CCTV equipment on shoplifting and address why people shoplift’. The last point about addressing why people shoplift, is important because it means that he would have to interview shoplifters to ask about their reasons for shoplifting. This implies talking to people who have actually done shoplifting themselves in the past. Because this is a key point, it should have been discussed separately. The other points, namely ‘the amount of shoplifting’ and ‘the influence of new CCTV equipment’ are more concrete because they can probably be measured and may not have to rely on interviews. Later in this paper I will examine the attempt to measure these aspects. In discussing the purpose of the research, the security manager should have also said that the reason for trying to understand more about shoplifting in his store is to help prevent it. The research is not just an academic exercise but has real-life implications. He should have made it clear that there are two aspects to the purpose of the research. The first is to try and understand more about why people shoplift. The second is to identify areas where the physical security of the store is weak so that it can be improved. Choice of methods The security manager made a good choice of methods for this study. It is true that a combination of qualitative and quantitative data can give the fullest picture in a case like this. However, he evidently has no grasp of how to do simple statistical calculations and so he tries to use the raw data only, which is inappropriate. He also does not understand the interview technique or what makes it valid with certain people but not others. So although he has correctly chosen the methodology, he has not applied the correct procedures. He intends to use the store’s statistical records of stock loss as quantitative data, plus interviews as qualitative (in-depth) data, which on the surface of things is a suitable methodology for this kind of research. Quantitative data The report gives the raw data only without any attempt at statistical calculation. This makes it very hard to compare the monthly figures for stock loss, because the baseline figures are different for each month. With a simple calculation, the figures could have been presented as percentages and this would have allowed for a comparison from one month to the next. This table would look as follows: If the data is shown in this way, we see that although there is indeed an increase in stock loss from May to August, the pattern has not been uniform. More information would be required for additional months to see whether there really is a trend. For example, In July the stock loss was actually at its lowest (4.75%), but the following month it peaked to its highest level (7.41%). It would also be a good idea to compare the data from the previous year or two years as well. It is not enough to give data just for four months and then conclude there is a trend. In the same way, the security manager has tried to use stock loss figures to see whether installing the CCTV helped to reduce theft or not. Once again he should have converted the raw data to percentages in order to make a comparison, because the baseline figures are different (268142 units in December versus 139023 units in February). If he had converted to percentages the data would have looked like this: We can see from this that there was in fact no change in the level of stock not accounted for before and after the installation of CCTV. Based on this we would conclude that CCTV made no difference at all to theft. However, once again the store manager should have looked at the bigger trend. He should have included data from the entire year (or at least 6 months) preceding the CCTV installation, versus the data from the entire year (or at least 6 months) following it (or as much monthly data as was available). In terms of validity, the numerical data supplied in the report is valid, because it comes from the shop’s own numerical records of stock. In terms of reliability there is no way of repeating or replicating this data because it is not an experiment. Therefore, issues of reliability are not relevant in this particular research context. The table that shows a breakdown of results by gender (p.5) provides data that does not meet the standards for quantitative research. The sample is far too small (3 men and 3 women) to draw any conclusions from. The same is true for his entire sample – it is too small to count as quantitative data, which needs about 50 subjects at least. With the gender breakdown the problem is particularly bad as the security manager has drawn erroneous conclusions based on this tiny sample, namely that more women are serious shoplifters than men. With samples of only 3 to 10 people, as in this study, one cannot use quantitative techniques, apart from reporting simple statistics like percentages. Qualitative data The qualitative data in the report is very poor. The security manager states that he conducted two semi-structured interviews with other staff members, but from his descriptions we see that they were not really interviews but were rather—in his own words (p. 2)—informal chats. The problem with this is that people may be distracted, may give replies that are not clearly thought out, or may be worried that others might overhear what they are saying. In the first ‘interview’ it is clear that Jason Smith is merely giving a personal opinion, without having really studied the store’s stock reports. His comments are reported as personal impressions (‘he thought..’; ‘he was pretty sure…’, p.2). Jason raised the point that retailers who sell similar merchandise probably have a similar problem with shoplifting. But as already noted the security manager does not even attempt to provide such figures, or to make comparisons between his shop and other similar shops. This point is worth following up and he should have done so. Obtaining stock loss figures from other shops would also have increased the reliability of the current data, because it is the closest one can come to replicating the information. In his second ‘interview’ with Stella, the manager showed a lack of neutrality because instead of trying to get Stella’s opinion, he used the data to influence her. However, he did not even give her the complete set of data. If he had showed Stella the percentage figures, she would have formed a different opinion. This interview was therefore worthless. The only point which Stella made independently was that the CCTV made female staff feel safer, rather than having any positive impact on lessening theft. Thus while interviewing staff members is a potentially valid source of data, in the current case it cannot be regarded as valid. This is because the interviewer tried to influence the respondents, and also because he did not conduct the interviews in a professional or confidential setting. There was a third interview, with a customer who had evidently just shoplifted. It is hard to believe the scenario that is reported, especially coming from a security manager who would know that there is a protocol to be followed. Anything said by a person in this situation must be disregarded because this is once again not a proper interview, but rather an informal conversation conducted under pressured circumstances. Regarding the questionnaires which the store manager drew up and had people fill in, unfortunately this information is also completely invalid. There are several reasons for this. The most important is that he gave the questionnaire to inappropriate people. He gave it to his own friends (and their friends) rather than to people convicted of shoplifting. Thus the sample which he used was not at all representative of the population in question. About a third of his friends answered the questionnaires but their answers must be regarded as purely fictitious. The fact that he used his own friends reduces the validity all the more, because his friends (and their friends) probably would not want him to know if they had ever really done any shoplifting, so their answers may not have been honest. Also his friendship circle is probably limited in terms of socio-economic status and other social variables. A representative sample drawn from the population of genuine shoplifters would theoretically have allowed every known shoplifter an equal chance of being selected for the research. Similarly, a representative sample drawn from the population of the general public would have had to allow every individual (whether a shoplifter or not) an equal chance of being selected. Either way, one might expect people from a variety of racial and cultural or language backgrounds, as well as people from different occupational or status groups, and both men and women, to have been included. The report does not give any indication of whether the security manager’s friends are similar to, or very different from, the general population or the population of known shoplifters. For this reason, the sample and all results from the questionnaires must be regarded as invalid because they are not a representative sample of the population being studied. The issue of reliability is also made meaningless because of the lack of a representative sample, as the research is not worth replicating. The manner of reporting the results from the questionnaire is also very poor. He is giving quantitative data (numericals) but in a text format, whereas the accepted format would be a table containing numericals. In addition, he should have included a copy of the actual questionnaire to show exactly what the questions were and what the scoring system was. Conclusions contained in the report The report does not present the conclusions in a coherent format. The security manager’s opinions or conclusions are scattered throughout the report instead of in a clearly labelled section. Also, due to all of the above weaknesses and problems, any conclusions which he drew based on his spurious data, must be regarded as purely speculative. For example, the statement that shoplifting ‘is increasing at an alarming rate’ is not valid in the context of the few figures that are reported. The same applies for the gender breakdown. It is a dangerous and invalid conclusion that more women shoplift than men, as this conclusion is not supported by the data. The statement that CCTV reduces stock theft is also clearly not justified when one looks at the stock loss figures as a percentage rather than in raw data format. The report also states that ‘there is a great deal of shoplifting’ but it does not compare the actual percentages of ‘Inventory Not Accounted For’ with figures from any other research. There is no indication of the figure recorded by most shops that carry similar types of merchandise. Without making this comparison, the security manager should not conclude that his shop’s percentages indicate ‘a great deal of shoplifting’. It could be that the percentages for stock loss in his store are actually well below average. Ethics The ethical standards of the report are very poor. The security manager evidently made verbal threats to a customer whom he saw shoplifting, which is unacceptable in a research context. He also placed the respondents of his questionnaires in the awkward position of probably having to imagine or lie about shoplifting. Or possibly some of them really had shoplifted and then the question is raised as to whether this information is handled in a confidential manner. There is no attention given to confidentiality anywhere in the report, with the full names of staff members being given. This could result in those people being harassed or penalised in some way by other staff members. There is no mention of whether any of the research subjects were told that they did not have to participate in the study, or if they gave their voluntary consent to do so. Finally, a very serious ethical issue is raised by the conclusion which the security manger reaches when he says that more women than men are serious shoplifters, and he suggests installing CCTV in departments visited more by women. This is a form of prejudice and harassment of women, because the conclusion is not statistically supported by the data. Improvement of methodology Throughout my paper I have highlighted which areas of the research and the report should be improved on. I have included these suggestions (e.g. the numerical tables) in the relevant sections of my critique, and these suggestions will not be repeated here. In summary, while the security manager did correctly choose a combined qualitative and quantitative approach to the study, he should have made fuller use of the data available in his shop’s stock-keeping records, and should have converted raw data to percentages. This would have allowed him to draw more sensible and valid conclusions. He also should have used a totally different sample of respondents for the questionnaire, so that the findings would be representative of shoplifters. He also should not have tried to draw conclusions about whether men or women are more serious shoplifters, because he did not have a sufficient sample to do statistical research based on gender. References McCall, R.B. (1990) Fundamental Statistics for Behavioural Sciences (5th ed.) Harcourt Brace Jovanovich Publishers, New York Kerlinger, F.N. (1986) Foundations of Behavioural Research (3rd ed.) Harcourt Brace Jovanovich Publishers, New York Read More