Summary QHM :(
,Week 1 – introduction: quantitative historical methods for historians
Concepts
Case (research unit)
Values that are horizontally in the data matrix
Variable
Values that are vertically in the data matrix
Value
The data in every cell
Level of measurement (values)
Nominal (categorical data); (qualitative) different between units
E.g., Amsterdam – Rotterdam – Delft
SPSS (variable view): nominal
We see the differences, but it is purely qualitative. You cannot calculate with
these variables.
Ordinal (categorical data): ranking.
E.g., cities with 5000-9999 inhabitants smaller than cities with 10.000-
19.9999 inhabitants
Words can be ordinal variables too. As long as they are rank able
SPSS (variable view): ranking
Ratio (& interval) (numeric data): exact distance between values
E.g., homicide rates
SPSS (variable view): scale
Scale is the only variable you can actually calculate with
the ‘higher’ variable type, the more statistical possibilities
only possible to convert to ‘lower’ variable type.
You can make ordinal data from your scale variables, but not the other way around.
, So, from scale to ordinal, but not from ordinal to scale.
Nominal is the most basic, you cannot calculate (names, places etc.)
Ordinal data can be ranked (time period)
Scale data is numerical you can calculate with (age, numeric etc.)
Determining the level of measurement
Can you make distinction between values
No useless for statistics
Yes next question
Can you rank the values?
No nominal level/data
Yes next question.
Can you determine the exact difference between two values?
No ordinal level
Yes scale level.
Population
The complete pool of information from which you draw a sample
Sample
Representative selection of the population
How to in SPSS
Sort cases
You can do this by going to the menu bar and clicking on 'Data' and then on 'Sort Cases'.
Or
You can do this faster if you place your cursor on the name of the variable in the 'Data View'
at the top and then right-click it. You can then select 'Sort Ascending' or 'Sort Descending'.
Define missing values
Start by going to the 'Variable View'. In the Missing column, click on the small blue square
that appears after you click on the cell. You then select the 'Discrete missing values' option.
In this option you can specify three codes, which stand for missing values. However, you do
not usually need more than one. So you can fill in a number in the first box which does not
occur in your data and which stands for missing values. Generally, '-9' or '9999' is used.
^
To fill it in fast in SPSS
Data --> Sort Cases --> Ascending (to sort the data). All the missing data will then appear at
the top of your data matrix. Fill in the first research unit (-9) and copy this cell to all other
blank cells.
Variables cannot start with a figure.
A year variable for example will be displayed as Y1980
,Week 1 – introduction: quantitative historical methods for historians
Concepts
Case (research unit)
Values that are horizontally in the data matrix
Variable
Values that are vertically in the data matrix
Value
The data in every cell
Level of measurement (values)
Nominal (categorical data); (qualitative) different between units
E.g., Amsterdam – Rotterdam – Delft
SPSS (variable view): nominal
We see the differences, but it is purely qualitative. You cannot calculate with
these variables.
Ordinal (categorical data): ranking.
E.g., cities with 5000-9999 inhabitants smaller than cities with 10.000-
19.9999 inhabitants
Words can be ordinal variables too. As long as they are rank able
SPSS (variable view): ranking
Ratio (& interval) (numeric data): exact distance between values
E.g., homicide rates
SPSS (variable view): scale
Scale is the only variable you can actually calculate with
the ‘higher’ variable type, the more statistical possibilities
only possible to convert to ‘lower’ variable type.
You can make ordinal data from your scale variables, but not the other way around.
, So, from scale to ordinal, but not from ordinal to scale.
Nominal is the most basic, you cannot calculate (names, places etc.)
Ordinal data can be ranked (time period)
Scale data is numerical you can calculate with (age, numeric etc.)
Determining the level of measurement
Can you make distinction between values
No useless for statistics
Yes next question
Can you rank the values?
No nominal level/data
Yes next question.
Can you determine the exact difference between two values?
No ordinal level
Yes scale level.
Population
The complete pool of information from which you draw a sample
Sample
Representative selection of the population
How to in SPSS
Sort cases
You can do this by going to the menu bar and clicking on 'Data' and then on 'Sort Cases'.
Or
You can do this faster if you place your cursor on the name of the variable in the 'Data View'
at the top and then right-click it. You can then select 'Sort Ascending' or 'Sort Descending'.
Define missing values
Start by going to the 'Variable View'. In the Missing column, click on the small blue square
that appears after you click on the cell. You then select the 'Discrete missing values' option.
In this option you can specify three codes, which stand for missing values. However, you do
not usually need more than one. So you can fill in a number in the first box which does not
occur in your data and which stands for missing values. Generally, '-9' or '9999' is used.
^
To fill it in fast in SPSS
Data --> Sort Cases --> Ascending (to sort the data). All the missing data will then appear at
the top of your data matrix. Fill in the first research unit (-9) and copy this cell to all other
blank cells.
Variables cannot start with a figure.
A year variable for example will be displayed as Y1980
