Secondary Quantitative Data - Essay Example

Types of Secondary Quantitative Data Available to the Social Science Researcher And the Strengths and Weaknesses of Secondary Quantitative Data Introduction and Types of Secondary Quantitative Data Research is an important tool of the social sciences. Social scientists make use of both primary and secondary research data, and both quantitative and qualitative, but often it is difficult to carry out independent data collection. Thus, secondary quantitative data has its uses and limitations. This paper briefly details the types of secondary quantitative data available to the social science researcher and describes the strengths and weaknesses of using such data in their research. Secondary data refers to "information which has already been collected by someone else and is available for you, the researcher, to inspect" (Clark, 1997, p.57). So we are concerned with research carried out by making effective use of existing quantitative data. That is, whereas in primary research both data collection and analysis is used, in secondary research, "creative analytic techniques [are applied] to data that have been amassed by others" (Kiecolt, 1985). Another important distinction to note given that this study examines the strengths and weaknesses of using secondary quantitative data is that secondary data is that which has been collected for another purpose but later reanalyzed for use in another piece of research. This of course brings to question the validity of doing this, the rationale for conducting secondary research, its advantages and limitations. Examples of the types of secondary quantitative data are the following: Official records relating to births, marriages and deaths; records relating to crime, divorce, voting patterns etc.; the census; records held by academic, business and other organizations. The census is a special type of secondary dataset due to it being obligatory. Other regular or ad hoc surveys also provide useful statistical information. Some sources for using secondary data include surveys conducted by organizations, economic data, university academia research, national and international statistics, and opinion polls. Secondary analysis can be used on a variety of quantitative data including cohort, time-series, trend, and so on. The widespread use of secondary data in social science research probably dates back to the 'secondary data movement' of the 1960s when there was "a growing emphasis upon the use of secondary data in research, with important developments in social indicators analysis, the rise of survey archives, and the overall development of quantitative social research all playing a part." (Sobal, 1982) Secondary quantitative data is used to a great extent in economics and geography amongst the social science disciplines. Uses of Secondary Data (Strengths) Often, the greatest advantages to using secondary quantitative data are the cost and time saving benefits, and the simpler process for obtaining it. It is simply quicker and cheaper to obtain quantitative data from secondary sources than it would be from primary sources through gathering data oneself. In contrast to secondary research, primary research, specifically data collection, is a more complex procedure, typically takes a lot of time, and usually costs more to carry out. It also requires appropriate skills, access to people or sites, special equipment and other resources etc. These requirements are not an issue for obtaining secondary data. There are also issues of "declining resources for research in the social sciences" (Kiecolt, 1985) and climatic constraints, which makes it expedient to rely on existing survey data. Moreover, in this Information Age, an abundance of quantitative data is available nowadays, particularly in libraries and on the Internet. As Kiecolt points out in 'Secondary Analysis of Survey Data': "With data already collected, the costs are only those of obtaining the data, preparing them for analysis (such as ensuring that all data are computer-ready and compatible with the system), and conducting the analysis (buying computer time, and so on). Compared with the time normally required to collect data in social research, the time necessary for acquiring an appropriate data set is miniscule." Furthermore, by circumventing the entire process of data collection, "a researcher can complete a research project independently, thereby eliminating the need for ancillary research staff' [It] obviates the need for researchers to affiliate with a large organization in order to command the backing necessary for acquiring adequate survey data" (ibid). In the case of national statistics for example, these are usually obtained from government research bodies or privately commissioned ones. To obtain the data they would have huge resources at their disposal and means of access to the people or places being researched. It could be next to impossible for a lone primary researcher to carry out a similar large-scale research. "Secondary analysis of existing surveys allows researchers access to data from large, national samples - data that would be difficult for a lone researcher to gather." (Kiecolt, 1985) The national census is a case in point. Then there are longitudinal studies that are carried out over long periods of time. This too may be inconvenient for the researcher to replicate. Another example of a time related complication is that of past data such as old climate records or data relating to individuals when they were younger or those who have passed away. Without a time machine it would be impossible to obtain this data again. Similarly, far away places may not be easily accessible, so data pertaining to those places may need to be obtained from secondary sources. Thus, secondary research is ideal for large data sets, historical and longitudinal data, and remote data too distant for the researcher to acquire. Additionally, secondary data could turn out to be of high quality, very reliable, and able to provide "unrivalled contextual material" (Clark, 1997). This of course depends on the source. Not all sources can guarantee quality, reliability and good contextualization, but some could. These would include large, independent and impartial organizations such as universities and research centres. Secondary as well as artificial data would also be more convenient for teachers to demonstrate with and students to use in the case of the initial stages of learning about social research before plunging into collecting original data for real. Secondary data can also then serve as a 'sample' or model. Obtaining secondary data is of course unobtrusive whereas obtaining primary data is an obtrusive process into people's lives. Some unforeseen discoveries may also be made along the way (Weijun, 2008). Secondary quantitative data, especially in the area of social research is becoming increasingly available for researchers to use through ease of access. Moreover, many of these are of high quality such as government administrative datasets geared towards policy-oriented research. These are usually based on large-scale data sets that would not be possible for a single primary research to conduct. In the field of education for example, quality large-scale data sets are becoming more widely available on an unprecedented scale. This provides the researcher with access to copious and good quality secondary quantitative data that was not available before. "Analyses of these data have the potential to radically improve the robustness and generalisability of educational research." (Hansen, 2007) In the UK, survey data from the Department fir Children, Schools and Families is a good example (DCSF) of a publicly accessible high quality administrative dataset, as are statistics from the Department for Work and Pensions (DWP). The former hold a National Pupil Database (NPD) containing a good range of information on all children in the UK education system. Administrative datasets such as the aforementioned NPD can offer invaluable information for secondary researchers though it is unlikely in most cases to be rich enough. It is possible however to combine the benefits of administrative data with rich survey data as in longitudinal studies. This makes the administrative data linked to rich survey data. An example is the Longitudinal Study of Young People in England carried out by the Economic and Social Data Service (ESDS). This is a longitudinal study that aims to identify, and enable analysis and understanding of key factors that affect young people's progress. These and other similarly available longitudinal studies can provide the secondary researcher with rich survey data to combine with other available administrative data. Some examples of major annual or continuous surveys are the International Passenger Survey, National Readership Survey, Labour Force Survey, Family Reources Survey, and the General Household Survey. The latter for example began in 1971, thus providing decades worth of data. Another great source of secondary quantitative data based in the UK is the UK Data Archive (UKDA). This houses thousands of datasets, one of "the largest collection of digital data in the social sciences and humanities in the UK" (UKDA). The advantage this archive offers over many others is that the research is prepared with academic research in mind for the use of data analysts, and has a support structure as well. There are many examples of social science researchers using secondary quantitative data. A few recent ones will be mentioned from different disciplines of social science as well as research in which a combination of methods were used for the purpose of demonstrating its strengths. Historians frequently make use of secondary sources for research. Mostly it is of a qualitative nature based on document analysis, especially if the time period concerned is further back in time. However, quantitative data is also used, for example in studying old census records. In either case, primary research is just not possible except in studying present day events or using living subjects to research the past. In Sociology, secondary quantitative data would be used comparatively more. In the area of Health, Gruffudd (2008) studied the current state and use of 'user evaluation consultation' of an air ambulance service in North Wales. Although primary research was conducted on 10 participants who had used the service and rightly so, the methodology incorporated secondary quantitative data to gather key performance indicators. This demonstrates the usefulness of obtaining secondary quantitative data to supplement primary research. In short, secondary quantitative data, which is typically in the form of published aggregate statistics, is becoming increasingly available and therefore useful for secondary research. It may be obtained from sources that carry out routine administrative procedures or the results of special large-scale surveys. When data already exists and is easily accessible, it is simply more practical to use it instead of conducting a repeat study unless necessary. An example of a study in which secondary data would be a necessity however is historical research such as research on some social issue relating to the past. It could also be unavoidable in conducting comparative research. For example, it may be possible to gather original data on suicide statistics but if these need to be compared to suicide rates in several other countries, the use of secondary data would be more convenient than going on a world tour. The use of secondary quantitative data for historical or comparative research is a justification on methodological rather than practical grounds. Another use for secondary data could be for triangulation purposes. For example, the results of a small-scale survey may be compared to a larger scale official survey to check for reliability of the original research. (Rowlingson, 2004, p.139) The traditional source of secondary data is the library. But an important factor in the greater availability and use of secondary quantitative data is the computer. Apart from being responsible for the availability of published data online, which is a convenient source not confined to specific locations as the library, the availability of computer software has also contributed to the popularity of using and analyzing secondary data - in the form of statistical computing programs such as SPSS and SAS. It is now easier to locate, access and manipulate survey data than ever before thanks to computer technology. Inappropriateness of Secondary Data (Weaknesses) WHEN/HOW NOT The greatest weakness of secondary data is its inflexibility because it is not prepared specifically for the problem or question at hand. Therefore, it is unlikely to always provide the precise information that is required. In other words, the primary researcher usually has specific requirements for his or her research that may not be precisely catered for by secondary quantitative data sources. Only the researcher better understands what kind of data is required or may be customized to meet the requirements. For example, the administrative data mentioned above is prepared specifically to aid policy making, and is not designed in a way that would suit researchers directly. Greater involvement of academia could however rectify this situation so that the data suits more users. The above-mentioned UK Data Archive may be a good example of an accessible and well-described archive of quantitative data but not all data is archived in the same way. Also, secondary quantitative data is either likely to be unavailable or difficult to obtain for very localized or uncommon topics of research or easily available but too general to be applied to the specific area of research intended. Even if the right kind of data is known to be available, it may not be freely and easily available, and therefore more costly to obtain. Even if obtaining the data is a simple process, the researcher may then have to figure out the actual issue or problem the data was prepared for, which may differ from his or her objectives and therefore affect the study being conducted. That is, the purpose, scope, direction, perspectives etc. of the researcher's study may be in significant conflict with those of the researcher or source from whom the secondary data is obtained. In such a case, the secondary data cannot be directly relevant. Moreover, "a certain degree of skill is needed in manipulating and analyzing data which can often be large-scale and complex." (Rowlingson, 2004) An example of a complex official survey is the Family Expenditure Survey. Other weaknesses of using secondary quantitative data could be due to not being able to be verified especially if the source is anonymous or untraceable. And, if the research was conducted in unusual circumstances, it may not be easily replicable. Furthermore, for some kinds of research focused on current quantifiable values in a rapidly changing world, secondary data can very easily become obsolete. For example, opinion polls for current affairs must be taken afresh to give the current perspective. It may be no use referring to secondary data conducted even in the recent past or pertaining to another location or country. To sum up, the limitations of using secondary quantitative data are due to its inflexibility, not being precisely fit for the purpose it is intended to be used for i.e. not always quite valid, the existence of differences of interpretation, differences in scope, perhaps not being easily available, coming from unverifiable sources, and being obsolete. All these factors assume of course that suitable secondary data is available and usable in the first place. Advantages and Disadvantages of Using Official Statistics As for the use of official statistics themselves, the social science researcher has access to a great variety of social, economic, business and political data, which are all readily available. Many examples of official statistics especially from government sources are already mentioned previously. Their advantage is that they "allow the examination of trends over time, comparisons between social groups and geographical regions, and 'before and after studies''" (Slattery, 1986). However, while official statistics are numerous and widely available, they do pose a number of disadvantages. We shall consider these from a sociological point of view in order to show a starker conflict of interests: Because official statistics are collected for administrative and political purposes, they do not directly cater to the research interests of sociologists. Definitions used in official statistics often differ from those used by sociologists. Moreover, definitions can change over time. Official statistics are often obsolete by the time they are published. Although plenty of statistics are published, there is much that is not made available due to "the cost involved or for political reasons" (Slattery, 1986, p.29). Even from the perspective of political science, there could be issues of public confidence in official statistics given that by their very nature, statistics can very easily be distorted or used in a way to support a particular opinion or give a particular impression. In other words, there can be "concerns about the massaging of figures" (Harrison, 2001, p.68). Conclusions It is important to be clear on the extent of need to use secondary quantitative data as opposed to conducting primary research. As discussed above, in certain circumstances or to satisfy particular study objectives it may be more suitable to use either one or the other. On the other hand, it is also possible to combine the use of both primary and secondary data. However, this study highlighted that in certain cases only the latter option may be possible as with research of a historical nature. Furthermore, to overcome some of the limitations of secondary data, it is also possible to combine the use of more than one source of data as mentioned above with combining administrative data and quantitative data from longitudinal studies. The validity of data is also an important issue to be kept in mind with quantitative data. But overall, the advantages of using secondary quantitative data outweigh gathering primary data assuming the right kind of data can be obtained. In some cases, there could be a perceived lack of relevance or a lack of knowledge about the available sources that may be hindering its use. However, this study also pointed out the limitations of relying on secondary research alone. In fact, the intended research cannot be carried out if the data required is not available. Difficulties are thus encountered in first locating the required data. For effective research, to make any secondary data useful, the researcher must be clear of his or her precise requirements, be aware of possible sources of secondary data, be able to access the data, and finally able to analyse or manipulate it as required. If appropriate to the research at hand, secondary research, official statistics in particular provide a very convenient, cheap and timesaving source of data. References Clark, G. 1997. 'Secondary data'. In Flowerdew, Robin and Martin, David (Ed's.). 2005. Methods in Human Geography: a guide for students doing research projects. 2nd edition. Prentice Hall, London. Gruffudd, Gwilym Sion ap. 2008. A user evaluation of a local air ambulance service in North Wales. International Journal of Health Care Quality Assurance. Vol. 21, Issue 6, pp. 585-597. Hansen, K. and Vignoles, A. 2007. The use of large scale data-sets in educational research. London: TLRP. Online at http://www.tlrp.org/capacity/rm/wt/vignoles (accessed 17 August 2009) Harrison, Lisa. 2001. Political research: an introduction. Routledge. Taylor and Francis Group. Kiecolt, Jill K. and Nathan, Laura E. 1985. Secondary Analysis of Survey Data (Quantitative Applications in the Social Sciences). SAGE Publications. Rowlingson, Karen. In Becker, Saul and Bryman, Alan. 2004. Understanding research for social policy and practice: themes, methods and approaches. Part 3: Using Existing Research. Bristol, The Policy Press. Slattery, Martin. 1986. Official Statistics. Taylor and Francis. Sobal, Jeff. 1982. The Role of Secondary Data Analysis in Teaching the Social Sciences. Library Trends. Pp. 479-488. http://www.ideals.uiuc.edu/bitstream/handle/2142/7216/librarytrendsv30i3m_opt.pdf'sequence=1 (accessed 15 August 2009) Weijun, Tang. 2008. Research Methods for Business Studies. Ch.7. http://www.drtang.org/lecture/RM/RM_07.pdf (accessed 16 August 2009) Websites: DCSF. http://www.dcsf.gov.uk/rsgateway/ DWP. http://www.dwp.gov.uk/research-and-statistics/ ESDS. http://www.esds.ac.uk/longitudinal/access/lsype/L5545.asp UKDA. http://data-archive.ac.uk/ Read More