Quantitative Historical Methods Summary
Week 1: Introduction
Qualitative vs. Quantitative
Diversity * Similarities
Holistic * Standardized
Content analysis * Statistical analysis
Flexible method * Rigid design database/software
Detailed * generalities
Interpretation * prediction
Values levels of measurement
- Nominal (categorical data): (qualitative) difference between units
e.g., Amsterdam-Rotterdam-Delft
- Ordinal (categorical data): ranking
e.g., cities with 5.000-9.999 inhabitants smaller than cities with 10.000-19.999 habitants
- Ratio (& interval) (numeric data): exact distance between values
e.g., homicide rates
the higher the scale, the more statistical possibilities
Soft- and hardware development of
last 5 decades made largescale
quantitative studies feasible
Major danger is the use of online
data by historians with too little
training in statistical techniques
Variable: any feature or any
attribute of a person, object or
situation that may vary from person
to person, object to object or
situation to situation.
Excel is not able to carry out statistical processing as accurately as SPSS can.
How to determine measurement level for a variable:
1. Does one value have a different significance to the other one? E.g., is a family with one child
different to a family with three children?
No: this variable cannot be used in statistics
Yes: go to question 2
2. Could you say that one value is greater or lesser than the other one? Are the values in any
kind of order? E.g., is a family with one child greater or lesser than a family with three
children?
No: this variable is measured at nominal level (use String in SPSS)
Yes: go to question 3
3. Can you determine how much difference there is between two values? E.g., can you say that
there is a difference of two children between a family with one child and a family with three
children?
No: this variable is measured at ordinal measurement level (Numeric)
Yes: this variable is measured at cardinal measurement level (Scale)
,Difference sample and population data:
Population: entire group that you want to draw conclusions about
Sample: specific group that you will collect data from
Advantages of quantitative history
1. From elite to the masses broadly based quantitative studies can provide reliable evidence
2. Descriptive statistics refer to reorganization or display of data in order to convey information
more efficiently
3. Relationshiops and extrapolation
4. Sampoling
5. Common language of science
Interpretation of figures
- Ultimate objective is to understand and explain past events better
- Figures have to be interpreted in a logical manner and backed up by facts
- We can deduce from these numbers (draw a logical conclusion)
Categorical nominal and ordinal
Numeric scale
Week 2
Crafts literature
Weighted index numbers: gives more weight to certain securities
Unweighted index numbers: gives equal allocation (verdeling) to all securities within the index
1997: veranderingen in o.a. levensverwachting en vrije tijd
2002: nieuwe BBp-gegevens en nieuwe HDI-formule
Logarithms: propertions rather than absolute differences.
HDI = (E+I+L)/3
E = education
I = income
L= longevity (lang leven?)
Calculating growth for periods
Growth: end value – start value
Average growth: (end value – start value) / period
, Growth rate: (end value – start value) / start value * 100 relative development
Logarithmic transformation SPSS compute variable: LN (variable) or LOG10 (variable)
Are being used to making this easier to grasp based on proportions on values. The distance
between for example 100 and 1000 is the same as 10 and 1000; 10. Used to transform skewed data
to approximately conform to normality
the exponent or power to which a base must be raised to yield a given number
Absolute figures:
Growth
Average growth
Relative figures:
Growth rate (percentage compared to start value)
Logarithmic (proportional)
essential differences between absolute and relative development
Relative figures often more informative
Index numbers: the value of a variable expressed as a percentage
- Historical trends and comparison:
o Relative change in time series throughout certain period (base year (often start year)
always equals 100!)
o Difference with growth rate:
Index numbers always compared to (one and the same) base period
o Formula: current period / base period * 100
- Geographical comparison
o Variation on scale
E.g., 0 1 (HDI), or 0 100 (KOF globalization index)
o Calculation based on minimum and maximum value of used variable
o Formula: (value variable-minimum)/(maximum-minimum)
w index numbers:
Formula: I = (variable / value base year) * 100
Some rules:
- Round off figure to integer
- Base period = 100
- Always mention the base period!
Choice of base year:
- Not always earliest case in time series
- Rule: use “normal” period, instead of “outlier” (e.g., war, plague, economic crisis etc)
- Sometimes average of number of years
Composite index numbers
Evolution of different variables in one number
Why?
- Evolution often quite complex (combination of components):
o Economic growth (productivity different sectors)
o Cost of living (price different products)
- Essential for historical research (evolutions)
- Very important in contemporary society (e.g. adaptation wages, rent prices etc.)
Week 1: Introduction
Qualitative vs. Quantitative
Diversity * Similarities
Holistic * Standardized
Content analysis * Statistical analysis
Flexible method * Rigid design database/software
Detailed * generalities
Interpretation * prediction
Values levels of measurement
- Nominal (categorical data): (qualitative) difference between units
e.g., Amsterdam-Rotterdam-Delft
- Ordinal (categorical data): ranking
e.g., cities with 5.000-9.999 inhabitants smaller than cities with 10.000-19.999 habitants
- Ratio (& interval) (numeric data): exact distance between values
e.g., homicide rates
the higher the scale, the more statistical possibilities
Soft- and hardware development of
last 5 decades made largescale
quantitative studies feasible
Major danger is the use of online
data by historians with too little
training in statistical techniques
Variable: any feature or any
attribute of a person, object or
situation that may vary from person
to person, object to object or
situation to situation.
Excel is not able to carry out statistical processing as accurately as SPSS can.
How to determine measurement level for a variable:
1. Does one value have a different significance to the other one? E.g., is a family with one child
different to a family with three children?
No: this variable cannot be used in statistics
Yes: go to question 2
2. Could you say that one value is greater or lesser than the other one? Are the values in any
kind of order? E.g., is a family with one child greater or lesser than a family with three
children?
No: this variable is measured at nominal level (use String in SPSS)
Yes: go to question 3
3. Can you determine how much difference there is between two values? E.g., can you say that
there is a difference of two children between a family with one child and a family with three
children?
No: this variable is measured at ordinal measurement level (Numeric)
Yes: this variable is measured at cardinal measurement level (Scale)
,Difference sample and population data:
Population: entire group that you want to draw conclusions about
Sample: specific group that you will collect data from
Advantages of quantitative history
1. From elite to the masses broadly based quantitative studies can provide reliable evidence
2. Descriptive statistics refer to reorganization or display of data in order to convey information
more efficiently
3. Relationshiops and extrapolation
4. Sampoling
5. Common language of science
Interpretation of figures
- Ultimate objective is to understand and explain past events better
- Figures have to be interpreted in a logical manner and backed up by facts
- We can deduce from these numbers (draw a logical conclusion)
Categorical nominal and ordinal
Numeric scale
Week 2
Crafts literature
Weighted index numbers: gives more weight to certain securities
Unweighted index numbers: gives equal allocation (verdeling) to all securities within the index
1997: veranderingen in o.a. levensverwachting en vrije tijd
2002: nieuwe BBp-gegevens en nieuwe HDI-formule
Logarithms: propertions rather than absolute differences.
HDI = (E+I+L)/3
E = education
I = income
L= longevity (lang leven?)
Calculating growth for periods
Growth: end value – start value
Average growth: (end value – start value) / period
, Growth rate: (end value – start value) / start value * 100 relative development
Logarithmic transformation SPSS compute variable: LN (variable) or LOG10 (variable)
Are being used to making this easier to grasp based on proportions on values. The distance
between for example 100 and 1000 is the same as 10 and 1000; 10. Used to transform skewed data
to approximately conform to normality
the exponent or power to which a base must be raised to yield a given number
Absolute figures:
Growth
Average growth
Relative figures:
Growth rate (percentage compared to start value)
Logarithmic (proportional)
essential differences between absolute and relative development
Relative figures often more informative
Index numbers: the value of a variable expressed as a percentage
- Historical trends and comparison:
o Relative change in time series throughout certain period (base year (often start year)
always equals 100!)
o Difference with growth rate:
Index numbers always compared to (one and the same) base period
o Formula: current period / base period * 100
- Geographical comparison
o Variation on scale
E.g., 0 1 (HDI), or 0 100 (KOF globalization index)
o Calculation based on minimum and maximum value of used variable
o Formula: (value variable-minimum)/(maximum-minimum)
w index numbers:
Formula: I = (variable / value base year) * 100
Some rules:
- Round off figure to integer
- Base period = 100
- Always mention the base period!
Choice of base year:
- Not always earliest case in time series
- Rule: use “normal” period, instead of “outlier” (e.g., war, plague, economic crisis etc)
- Sometimes average of number of years
Composite index numbers
Evolution of different variables in one number
Why?
- Evolution often quite complex (combination of components):
o Economic growth (productivity different sectors)
o Cost of living (price different products)
- Essential for historical research (evolutions)
- Very important in contemporary society (e.g. adaptation wages, rent prices etc.)
