An Introduction to Quantitative Research Methods in History

Course Aims and Objectives

The overall aims of the workshop were to introduce first year history students to the basic skills of presenting quantitative data in summary statistical, graphical and tabular form. The workshop aimed to enable history students to combine quantitative evidence, that has been gathered and produced using a science based approach to research methods favoured by political scientists, social scientists and, indeed increasingly in some branches of history, with the text-based interpretative evidence of the historian. In short, the workshop is an exercise in multi-disciplinary research, taught by a combination of lectures and hands-on computer laboratory work. However, there are important issues that the history student needed to confront concerning what is acceptable as knowledge. An explanation in one discipline is not necessarily accepted as warrantable knowledge in another discipline. Thus, the theory that underpins the political science approach to the study of electoral behaviour was outlined and the students were introduced to the concepts of ontology, epistemology and the consequent methodological differences between how historians and political scientists go about their work and produce knowledge about political behaviour. This divergence was explained in terms of the interpretative, subjective, impressionistic, value-laden and non-generalisable explanations and theories that some positivists accuse historians of producing. This was then contrasted with the ostensibly objective, precisely measured, accurately and unambiguously defined, value free and generalisable explanations, theories, predictions and universal laws of cause and effect that political scientist aspire to, albeit using what some historians would consider often fragmentary, distorted or biased data.

The upshot of the argument presented to the students was that each approach has its particular strengths and weaknesses. Nevertheless, each can tell us something about a political phenomenon. Indeed, in the study of electoral behaviour political scientists increasingly recognise that quantitative approaches can and should be complemented by qualitative techniques as used by the historian in order to explain contextual effects that are intrinsically difficult to measure. Likewise, in the discipline of history, especially political history, there is a recognition of the need not only to be able to analyse and present succinctly trends in electoral behaviour, but also the need to be able to engage more meaningfully in scholarly argument and debate with political scientists (see Dunleavy 1990, Kavanagh 1991, Ramsden 1992, Devine 1994,Rallings and Thrasher 1997, Bale 1999).

No prior knowledge of computers, statistics or social research methods on the part of the students was assumed or required for participation in the workshop. The workshop aimed to enable students to evaluate quantitative evidence, analyse and display raw quantitative data, integrate quantitative and qualitative approaches mindful of their respective strengths and weaknesses, and to equip students with the skills necessary to mine the multiplicity of political, social and economic data that remains inaccessible without such skills. The intended learning outcomes of the workshop included increased student awareness of inter-disciplinary research and an enhanced ability to engage with political science sources. The module aimed to improve a history student's ability to evaluate and present quantitative evidence, and to combine written work that is clearly structured and based upon wide reading of political history sources with quantitative evidence presented in appropriate graphical, tabular and statistical form. Thereby the confidence of history students in the handling quantitative historical data will be improved. The assessed skills element of the workshop was based on the ability to meld quantitative and qualitative evidence appropriately and effectively in an essay in response to a specific question on post-war voting behaviour in Britain. The delivery of the workshop was over six, two-hour sessions, in the form of lectures that preceded supervised hands-on computer laboratory sessions accompanied by step-by-step guides to data entry.

Course Schedule

The first lecture entailed an overview of positivism and the quantitative research methodology that informs the methods used by political scientists in their studies of electoral behaviour. The concepts of ontology and epistemology and what they mean in terms of acceptable methods of producing warrantable knowledge in the social sciences was contrasted with the interpretative approach of the political historian. Students were made aware of the seeming incommensurability of the quantitative and qualitative perspectives. An argument was then presented that each approach has its strengths and weaknesses and that each can tell us something about political behaviour. The main points of the lecture and a bibliography of texts that dealt with the quantitative/qualitative debate, were outlined in a student handout. The emphasis of the lecture then turned to the presentation of history and how numbers are used and can be used by historians as historical evidence. Students were made aware of the power of numbers and how they can be selected, reconstituted, redefined, reordered and displayed to suit the purposes of those that gather and use them.

In the following lecture the students were introduced to descriptive statistics and some elementary statistical techniques that arrange and display quantitative data so that basic questions can immediately be asked of the data. It became increasingly evident to the students that a table or a figure that represented the character of a mass of electoral data was extremely useful, and that by some elementary processing of figures using a statistical programme on the computer, simple measures of average or typical experience gave some notion of the range of variation in voting behaviour over time and space. It was shown that at its simplest level, quantification brought to history the ability to summarise large bodies of data, to display such data effectively and to express typical measures and values. Students were made aware of the variety of types of data and types of numbers and what this meant in terms of the kind of meaningful analysis they can be subjected to. This was followed by an overview of the growth of quantitative history, its advantages and disadvantages, and the uses of quantitative methods in the academic discipline of history were exemplified.

The third lecture had as its focus the different types of 'average' that are used to summarise information from a larger set of numbers, in this case electoral data. Given that the main purpose of statistics is to describe sets of numbers briefly and accurately it was brought to the students' attention that the so-called average can be misleading and that there can be a large departure from it. Indeed, how, for example, the aggregation of electoral data can disguise significant variations in actual voting behaviour, and that the mere indication of the central point of a distribution of numbers only allows a partial view, only an indication of typical patterns of electoral behaviour. From the pros and cons of these measures of central tendency the lecture then turned to measures of dispersion and how these descriptive statistical techniques allowed the researcher to gain a broader picture of the data and facilitated the description of any variation, i.e. the atypical so often of primary interest. This lecture ended with a recap of descriptive statistics and a very brief overview of what inferential statistics are and what they can and cannot do.

The following two lectures looked at electoral change in Britain since 1945. First, how historians have interpreted voting patterns as an expression of underlying social forces and thereby attributed developments in modern British political history to fundamental shifts in the social structure and social attitudes, and how political history is characterised by a sociological approach with electoral behaviour regarded as a barometer of social change. Studies of voting behaviour at British general elections in the 1945-1970 period were then reviewed and an era of two-party dominance, electoral stability, strong party identification, and class and party alignment was presented. In the following lecture changes in the voting behaviour of the British electorate and how these changes have been measured, evidenced and explained by political historians and political scientists was discussed. The lecture explored the decline in support for the two major British political parties and the concomitant rise of the minor parties, increased regional variations in the distribution of each party's share of the vote, increased electoral volatility and the debates that accompany these political phenomenon.

In the fifth and final lecture the requirements of the workshop's essay assignment were outlined. The general format and presentation of the essay, what was required in terms of citation and referencing of quantitative and qualitative sources of information and data. Indeed, how to cite and reference data from various sources including electronic, how to cite sources of data used in tables and charts, and how to compile a list of tables and charts and there contents. The essay assignment required the students to meld written work based on research of textual sources with quantitative evidence. The students were required to answer a question on electoral behaviour in Britain at post-war general elections and to integrate appropriate charts, graphs and tables into the text in order to support their central argument. Handouts accompanied the lectures and summarised each particular lecture and highlighted recommended reading. A workshop descriptor was given to each student in which the aims, contents, and requirements of the course were outlined, and a bibliography and glossary of terms included.

At the last of the six weekly sessions the students had an opportunity to present preliminary drafts of their work and to resolve any difficulties they may have regarding the assignment, and of course to complete any unfinished graphs, tables, charts and editing of output they had been working on during the supervised data processing sessions on the computer.

In the hands-on computer sessions the students were provided with a step-by—step guide and close supervision where necessary, that enabled students without any prior knowledge of computing to enter and analyse the electoral data provided and to create a number of summary statistical charts, figures and tables. At the end of the six one-hour computer sessions most students had completed many of the charts etc. required and had only to write up the essay and integrate the quantitative evidence appropriately.

Course Themes and Output

The principal themes of the computer sessions included the assembly and handling of data sets, the analysis of data and the presentation of findings. Students used the SPSS statistical package to manipulate and analyse British general electoral data 1945-2001. At the end of the workshop the students had gained the basic skills needed to assemble data sets by having entered their own data, had prepared data by assigning names and value labels to variables, and analysed data using a variety of elementary but nonetheless very useful statistical methods. The students had learnt how to log-on and open an SPSS data file, to enter data, create, define and label variables, to run Frequency Analysis and obtain descriptive statistical information about variables, to create new index variables, to create charts and graphs, to customise charts and place them in a word document, to print selections from a data set, to select cases and split files, to exit from SPSS and to save and retrieve the data set.

The output generated by the students included; a multiple-line graph that depicted trends in each party's share of the vote at British general elections, and a line graph that depicted changes over time in the two-party share of the vote i.e. the sum of the two major British parties, Labour and Conservative. The students used the electoral data to create an index variable that measured the level of net electoral volatility at each successive general election and presented the results in the form of a bar graph, similarly they created an index variable to measure trends in class voting at British general elections. Changes in the support for the Liberal Party were charted in a line graph that depicted the party's percentage share of the vote at successive elections and contrasted in a further chart with the percentage seats with which the first past-the-post system rewards minor parties. The students also produced appropriate tables to accompany the charts and graphs. These were the minimum requirements of the workshop in order for students to be able to visually present quantitative data in summary statistical, tabular and graphical form and adequately evidence their response to an essay question on British voting behaviour in the 1945-2001 period.

Course Evaluation

On completion of the module students received module evaluation forms that gave them the opportunity to anonymously express their views on the quality of the module. The completed forms, returned to the faculty office by the students, requested that they circle the appropriate number for each of the following questions using a scale : 1 = unsatisfactory, 2 = below average, 3 = satisfactory, 4 = good, 5 = very good/excellent.

• Q 1. Were the aims and objectives/learning outcomes of the module presented clearly?
• Q 2. Were the assessment requirements made clear and fully discussed?
• Q 3. Was the library provision adequate for the module?
• Q 4. On a week-by-week basis, was the module well-organised and effectively run?
• Q 5. Was the module taught in a stimulating way?
• Q 6. If known, was your written work returned punctually, with adequate feedback?
• Q 7. Did the module deliver what it promised, in terms of content, aims, skills etc?
• Q 8. All things considered, what is your verdict on this module?

The evaluation form also invited students to comment upon what they particularly liked or disliked about the module, and to make suggestions for future improvements. Feedback from the students, (sixteen out of twenty-seven returned their module evaluation forms) was encouraging. As can be seen in the statistics outlined in the Table 1, the indicators reflect a positive experience by the students. Library provision apart, the means for six of the eight indicators were equal to or more than 4, categorised as good on the scale. More specifically, in terms of how the module delivered on its promised content, aims and skill development, 31% of the students reported satisfactory, 50% good, and the remainder excellent (Table 8). Furthermore, many students commented upon the improvement to their IT skills that the module had made. Less pleasing was that 56% of students thought that library provision of core module texts was only below average to satisfactory (Table 4), a point reiterated in the comments made by students and one which will be addressed. Although the majority of the students reported that the module was taught in a stimulating way and was well organised and effectively run (Tables 4 and 5), there were nonetheless comments made by some students about the limited number of computers with the SPSS program available for their use on campus and the difficulty for those students who lived off campus in that few had computers at home let alone ones with the SPSS program. In the main students were able to complete the data analysis requirements of the module during supervised laboratory sessions and had only to integrate their charts, tables and figures into the text of their essays in their own time. Clearly, for slower students and especially those living off campus without access to a computer completion of the module meant longer hours at the university and in some cases extra costs in travel and time. Nevertheless, these are problems that can be overcome by an increased proportion of teaching time allocated to hands on supervised computer laboratory and a corresponding decrease in that allocated to lectures. These problems apart, the statistics outlined in Tables 1-9, and the general tone of the comments reflected a very positive experience by the students.

Assessment of the module was determined by the student's ability to meld quantitative and qualitative evidence appropriately and effectively in answer to a specific question on post-war voting behaviour in Britain at parliamentary elections. The assignment required students to answer one of a choice of questions in no more than one thousand words plus charts, graphs and tables that they considered appropriate to substantiate their argument. The essays had to be supported with footnotes/endnotes where appropriate and a bibliography in all cases. Citation of all sources of data used in tables and charts was also required. The number of students who completed the module was twenty-two, the remaining five students failed to present work for assessment. The minimum mark awarded was 41% and the maximum mark 70%. The mean mark achieved was 55.6% and the standard deviation of the marks awarded 8.2%. Every one of the twenty-two students who completed the module surpassed the minimum pass rate of 40%. One student scored a first class grade of 70%+, seven students achieved upper-second grades of 60-69%, eight students lower second grades 50-59%, and six students third class grades 40-49%.