Quantitative Methods

1

• Quantitative Methods
• and Computer Applications in the
• Historical and Social Sciences

Roman Studer Nuffield College roman.studer_at_nuffiel
d.ox.ac.uk
2
Aim of the Course

• Provide an introduction to elementary
  quantitative methods and statistics, covering
  some of the techniques most widely used in
  research in the historical and social sciences
• Introduce the statistical software package STATA
  and show how the relevant statistical
  applications can be performed with this software

3
Motivation of the Course

• Understanding quantitative research
• Quantitative methods are widely used both in
  History as well as in (other) Social Sciences
• Being able to make use of quantitative research,
  but also being able to assess its limits

• Using quantitative methods in your own research
• How representative or reliable are observed
  patterns?
• If many factors determine an outcome, the
  relative importance of the factors can be
  determined
• Graphs and tables can be very persuasive tools,
  they can also give valuable hints and raise new
  questions

• General usage
• We live in a data world, hence being able to
  handle data and quantitative studies properly is
  a very useful qualification both in the job
  market as well as for everyday life

4
Lecture Plan

• Week 1 Introduction
• Week 2 Descriptive Statistics
• Week 3 Correlation
• Week 4 Simple Linear Regression
• Week 5 Statistical Significance I
• Week 6 Statistical Significance II
• Week 7 Multiple Linear Regression
• Week 8 Dummy Variables

5
Course Arrangements

• Textbook
• The principal text will be Feinstein, C. H. and
  M. Thomas (2002). Making History Count A Primer
  in Quantitative Methods for Historians.
  Cambridge Cambridge University Press.

• Lectures/computer classes
• Normally the first hour will consist of a lecture
  introducing one of the topics
• The second hour will be devoted to a computer
  class where we look at applications of the new
  topics, using STATA

• Weekly homework
• Readings (relevant chapters in the textbook and
  articles containing applications of the topics
• Problem sets ? To be handed in by noon on Mondays

6
Course Arrangements (II)

• Mock exam
• At the end of the course there will be a simple
  take-away examination to test your understanding
  of the various concepts and procedures covered
  during the course

• All information is available on the courses
  website
• http//www.nuff.ox.ac.uk/users/studer/teaching.htm
  
• including slides, problem sets and data sets

7
Your Input

• Any questions so far?
• Any comments?
• Short introduction of participants
• Whats your background and how much do you know
  already about statistics and quantitative
  methods?
• Why are you taking the course?
• Are you planning to use quantitative methods in
  your research?
• Is there anything specific you want to learn that
  is not on the program?

8
Lecture 1 Definitions Concepts

• Cases, variables, and values
• A data set consists of a series of cases each of
  which has one or more characteristics, known as
  variables. For each variable there is a sequence
  of varying observations, each with its own
  particular value
• Cases are the basic unit of measurement, and they
  can be individuals, households, firms, towns,
  countries, etc.
• Example English Poor Law data set
• 311 English parishes Cases? Variables?
  Values?
• Relief expenditure of each parish Cases? Variabl
  es? Values?
• 20.4, 16.3 Cases? Variables? Values?

9
Cross-section and time-series variables

• Time-series variables
• A set of measurements that applies to a single
  case at different periods of time is referred to
  as a time series
• Example Real wages in London, 1500-1800

Source RC Allen (2001), The Great Divergence
10
Cross-section and time-series variables (II)

• Cross-section variables
• A set of measurement that applies to different
  cases at a single point in time is referred to as
  a cross-section
• Example Real wages across Europe in 1750

Source RC Allen (2001), The Great Divergence
11
Cross-section and time-series variables (III)

• Panel data
• Time series are used to study changes over time
  and cross-sections to study differences between
  cases
• It is also possible to combine cross-section and
  time series data to create a panel or
  longitudinal data set
• Example Real wages across Europe, 1500-1800

Source RC Allen (2001), The Great Divergence
12
Levels of Measurement

• Nominal measurement
• Each value defines a distinct category but gives
  no information other than the label or name
  (hence nominal level)
• This is the lowest level of measurement
• Example Cases Migrants
• Variables Birthplace of migrants
• Values London, Oxford, London

• Ordinal measurement
• This applies when it is possible to order or rank
  all the categories according to some criterion
  without being able to specify the exact size of
  the interval between any two categories
• Example Cases Migrants
• Variables Skill level of migrants
• Values unskilled, unskilled, semi-skilled

13
Levels of Measurement

• Interval or ratio measurement
• These measurements have all the properties of an
  ordinal scale. In addition, it is now possible to
  measure the exact distance between any pair of
  values
• This level of measurement is truly quantitative
  and any statistical procedure can be applied to
  values on an interval or ratio scale
• Values are either continuous or discrete
• Example 1 Cases Migrants
• Variables Number of children
• Values 0, 2, 7
• ? Discrete variable! Values can take only a
  number of pre-determined values
• Example 2 Cases Migrants
• Variables Height of migrants
• Values 1.55m, 1.62m, 1,58m
• ? Continuous variable is measured in units
  that can (theoretically) be reduced in
  size to an infinite degree

14
Populations and Samples

• The population
• Refers to all possible observations. In the Poor
  Law example that would be all parishes in the
  whole country

• The sample
• Is a subset of the population, e.g. all the
  parishes starting with O, like Oxford.
• Normally we only have samples, either because not
  all cases are available or because it would take
  far too much time to study all cases
• This is not a big problem, as we have tools to
  make valid inferences from a sample
• A crucial feature of any sample is whether it is
  random or not
• In Social Science there are clear procedures how
  to construct and deal with samples
• As historical data is a lot harder to obtain, you
  sometimes just have to use what you are able to
  get

15
Dummy variables

• Variables that cannot be measured can still be
  used for quantitative work by assigning values
  (mostly 0 and 1) representing the (two or more)
  categories
• Example 1 Cases Migrants
• Variable Sex
• Values 0 (man), 0, 1 (woman)

16
Homework

• Readings
• Feinstein Thomas, Ch. 1-2, Appendix A.1
  (describing the data set we use next time)
• Get a textbook, if you still dont have one
• Also, register with IT if you havent done so
  already
• No problem set this week