Ch.14 Exam Prep - Secondary Analysis And Official Statistics

Chapter 14 - Secondary analysis and official statistics

Test Bank

Type: true-false

Title: Chapter 14 - Question 01

01) Secondary analysis is the analysis of data that we have collected ourselves

a. True

b. False

• the secondary analysis of data that have been collected by other researchers;
• the secondary analysis of data that have been collected by other organizations in the course of their business.

Type: true-false

Title: Chapter 14 - Question 02

02) Secondary research data can only have been collected by governments.

a. True

b. False

• the secondary analysis of data that have been collected by other researchers;
• the secondary analysis of data that have been collected by other organizations in the course of their business.

Type: true-false

Title: Chapter 14 - Question 03

03) Most secondary analysis data sets are low quality.

a. True

b. False

First, the sampling procedures have been rigorous, in most cases resulting in samples that are as close to being representative as one is likely to achieve. While the organizations responsible for these studies suffer the same problems of survey non-response as anybody else, well-established procedures are usually in place for following up non-respondents and thereby keeping this problem to a minimum.

Secondly, the samples are often national samples or at least cover a wide variety of regions. In addition, some datasets enable cross-national comparison (see Research in focus 14.4). The degree of geographical spread and the sample size of such datasets are invariably attained only in research that attracts quite substantial resources. It is certainly inconceivable that student projects could even get close to the coverage that such datasets attain.

Thirdly, many datasets have been generated by highly experienced researchers and, in the case of some of the large datasets, such as the WERS (see Research in focus 3.14) and the LFS (see Research in focus 14.7), the data have been gathered by research organizations that have developed structures and control procedures to check on the quality of the emerging data. Some large datasets that are suitable for secondary analysis are described in Table 14.1.

Type: multiple choice question

Title: Chapter 14 - Question 04

04) Which of the following is not an advantage of secondary analysis?

Cost and time. As noted at the outset, secondary analysis offers the prospect of having access to good-quality data, such as that available from the UK Data Archive (see section on ‘Accessing the UK Data Archive’ below), for a tiny fraction of the resources involved in carrying out a data collection exercise yourself.

High-quality data. Many of the datasets that are employed most frequently for secondary analysis are of extremely high quality. By this we mean several things. First, the sampling procedures have been rigorous, in most cases resulting in samples that are as close to being representative as one is likely to achieve. While the organizations responsible for these studies suffer the same problems of survey non-response as anybody else, well-established procedures are usually in place for following up non-respondents and thereby keeping this problem to a minimum. Secondly, the samples are often national samples or at least cover a wide variety of regions. In addition, some datasets enable cross-national comparison (see Research in focus 14.4). The degree of geographical spread and the sample size of such datasets are invariably attained only in research that attracts quite substantial resources. It is certainly inconceivable that student projects could even get close to the coverage that such datasets attain. Thirdly, many datasets have been generated by highly experienced researchers and, in the case of some of the large datasets, such as the WERS (see Research in focus 3.14) and the LFS (see Research in focus 14.7), the data have been gathered by research organizations that have developed structures and control procedures to check on the quality of the emerging data. Some large datasets that are suitable for secondary analysis are described in Table 14.1.

Opportunity for longitudinal analysis. Partly linked to the last point is the fact that secondary analysis can offer the opportunity for longitudinal research, which, as noted in Chapter 3, is rather rare in business and management research because of the time and cost involved. Sometimes, as with the WERS, a panel design has been employed and it is possible to chart trends and connections over time. Such data are sometimes analysed cross-sectionally, but there are obviously opportunities for longitudinal analysis as well. Also, with datasets such as the LFS, where similar data are collected over time, usually because certain interview questions are recycled each year, trends (such as changes in working time or shifting patterns of employment) can be identified over time. With such datasets, respondents differ from year to year so that causal inferences over time cannot be readily established, but nonetheless it is still possible to gauge trends. For example, although the study by Knight and Latreille (2000) was confined to use of the 1998 WERS data (see Research in focus 14.5), the authors made frequent comparison with analyses from the 1990 WERS data to show that there had been relatively little change in patterns and rates of disciplinary sanctions and dismissals and complaints to employment tribunals during this time period. Similarly, a study by Addison and Belfield (2000) used data from the 1998 WERS to replicate research done by other researchers who had used data from the 1990 WERS in order to test whether or not e orts to boost employee participation have had any effect. Undoubtedly, the publication of findings from WERS 2011 will have enabled researchers to compare data from the most recent study with data from previous iterations of the survey.

Subgroup or subset analysis. When large samples are the source of data (as in the WERS), there is the opportunity to study what can often be quite sizeable sub- groups of individuals or subsets of questions. Very often, in order to study specialized categories of individuals, small, localized studies are the only feasible way forward because of costs. However, large datasets can frequently yield quite large nationally representative samples of specialized categories of individuals, such as workers in a particular industry or occupation or with a particular set of personal characteristics. These can form the basis for representative sampling of individuals. Similarly, when a large-scale survey covers several topic areas, analysis may involve focusing on a smaller subset of questions that are covered by the survey. For example, Addison and Belfield (2000) were interested in the effects of European works councils on organizational performance and employee attitudes. They therefore analysed the responses from just one question in the 1998 WERS, which related to the status of these new institutional arrangements.

Opportunity for cross-cultural analysis. Cross-cultural research has considerable appeal at a time when social scientists are more attuned to the processes associated with globalization and to cultural differences, though it is easy to forget that many findings should not be taken to apply to countries other than the one in which the research was conducted.

However, cross-cultural research presents barriers to the social scientist. There are obvious barriers to do with the cost and practical difficulties of doing research in a different country, especially when language and cultural differences are likely to be significant. The secondary analysis of comparable data from two or more countries provides one possible model for conducting cross-cultural research. The ISSP is explicitly concerned with bringing together findings from existing social science surveys from different countries and contexts. An example of the kind of cross-cultural analysis the programme has produced is given in Research in focus 14.4. Another example to illustrate how data from more than one country can be compared is a study by Coutrot (1998), in which he compared the industrial relations systems of France and the UK through statistical analysis of two broadly similar datasets—WERS 1990 and Relations Professionnelles et Négociations d’Entreprise (REPONSE) 1992 (a large-scale survey that covers similar issues to WERS and is based on interviews with managers and employee representatives in France). However, in order for a cross-cultural analysis to be conducted, some coordination is necessary so that the questions asked are comparable. Differences between countries in the definitions used and the criteria for inclusion can make this di cult, as the example relating to the use of official statistics, provided by Jackie Davies (2001; see Research in focus 14.6), illustrates.

a. Cost and time

b. More reliable results

c. High quality data

d. Opportunity for cross-cultural analysis

Type: multiple choice question question

Title: Chapter 14 - Question 05

05) Outline two disadvantages of secondary analysis. Please select all that apply

Lack of familiarity with data. When you collect your own data, when the dataset is generated, it is hardly surprising that you are very familiar with the structure and contours of your data. However, with data collected by others, a period of familiarization is necessary. You have to get to grips with the range of variables, the ways in which the variables have been coded, and various aspects of the organization of the data. The period of familiarization can be quite substantial with large complex datasets and should not be underestimated.

Complexity of the data. Some of the best-known data- sets that are employed for secondary analysis, such as the WERS, are very large in the sense of having large numbers of both respondents and variables. Some- times, the sheer volume of data can present problems with the management of the information at hand, and, again, a period of acclimatization may be required.

Also, some of the most prominent datasets that have been employed for secondary analysis are known as hierarchical datasets, such as the WERS. The difficulty here is that the data are collected and presented at the level of both the organization and the individual, as well as other levels. The secondary analyst must decide which level of analysis is going to be employed. If the decision is to analyse individual-level data, the individual-level data must then be extracted from the dataset. Different data will apply to each level. Thus, at the organizational level, the WERS provides data on such variables as number of employees and level of owner- ship, while, at the individual level, data on age, qualifications, and salary level can be found. For example, Hoque (2003) was interested in the impact of Investors in People (IiP) accreditation on workplace training practice. He used data from the 1998 WERS managers’ survey to extract organization-level data to build up a profile of workplaces that have IiP accreditation. How- ever, in order to evaluate the impact of IiP accreditation on training practice, Hoque relied on individual-level data, in the form of data about training activity taken from the survey of employees. These included questions about the number of days spent on training that were paid for or organized by the employer and whether or not the employee had, in the previous twelve months, discussed his or her training needs with his or her supervisor. He used these data to draw conclusions at the level of the organization, and to make comparisons of the effectiveness of training practice in accredited versus non-accredited workplaces.

No control over data quality. The point has been made on several occasions that secondary analysis offers the opportunity for students and others to examine data of far higher quality than they could collect themselves. However, this point applies mainly to datasets from a regulated source such as the UK Data Service (see Table 14.1). These tend to be commissioned by a government department and conducted by researchers who are regarded as independent or at least somewhat distanced from the issues that are being investigated, such as academics working for a university research unit. While the quality of data should never be taken for granted, in the case of such datasets it is reasonably assured, though that is not to say that the data will necessarily meet all of a prospective secondary analyst’s needs, since they may not have been collected on an aspect of a topic that would have been of considerable interest. With other datasets, somewhat more caution may be necessary in connection with assessment of data quality. This may be of particular concern when using data that are the result of commercially commissioned research, as is the case in market research or when using surveys that have been conducted in-house by a company that wants, for example, to measure the effectiveness of its HRM strategy.

Absence of key variables. Because secondary analysis entails the analysis of data collected by others for their own purposes, it may be that one or more key variables may not be present. You may, for example, want to examine whether or not a relationship between two variables holds when one or more other variables are taken into account. Such an analysis is known as multivariate analysis, an area that will be touched on in Chapter 15. The inability to examine the significance or otherwise of a theoretically important variable can be frustrating and can arise when, for example, a theoretical approach that has emerged since the collection of the data suggests its importance. This is also a drawback in meta-analysis (Key concept 14.9 and Research in focus 14.10), some- times making it di cult for researchers to generate unambiguous conclusions as a result of the analysis). Obviously, when researchers collect primary data themselves, the prospect of this happening should be less pronounced.

a. Lack of familiarity with data

b. Lack of access to data

c. Not as robust as primary analysis

d. Complexity of the data

Type: true-false

Title: Chapter 14 - Question 06

06) Secondary research generally does not provide the opportunity for longitudinal analysis.

a. True

b. False

Type: true-false

Title: Chapter 14 - Question 07

07) Cross-cultural analysis can be conducted using the official statistics of a particular country.

a. True

b. False

Type: true-false

Title: Chapter 14 - Question 08

08) Issues of validity and reliability are much less considerations for secondary as opposed to primary analysis.

a. True

b. False

Type: true-false

Title: Chapter 14 - Question 09

09) Official statistics are not generally used for business research.

a. True

b. False

Type: true-false

Title: Chapter 14 - Question 10

10) The BSA refers to the British Social Attitudes survey which covers a wide range of areas of social attitudes and behaviour.

a. True

b. False

Type: true-false

Title: Chapter 14 - Question 11

11) Most governments restrict the data they collect to researchers due to privacy and confidentiality concerns.

a. True

b. False

Type: true-false

Title: Chapter 14 - Question 12

12) Meta-analysis is the study of big data online to reveal macro patterns into digital behaviour.

a. True

b. False

Type: true-false

Title: Chapter 14 - Question 13

13) The ecological fallacy is the error of assuming that inferences about individuals or organisations cannot be made from findings relating to aggregate data.

a. True

b. False

Type: true-false

Title: Chapter 14 - Question 14

14) An unobtrusive measure is any method of observation that directly removes the observer from the set of interactions or events being studied.

a. True

b. False

Type: true-false

Title: Chapter 14 - Question 15

15) Official statistics are considered an intrusive form of measurement.

a. True

b. False