However, there are other groups who are against recognising sex work as a genuine option for women

such as other legitimate

occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
However, there are other groups who are against recognising sex work as a genuine option for women such as other legitimate
occupations as this diminishes the possibility of getting women out and improving their lives.[3]
Not all work done by children should be classified as child labour and needs to be eliminated, and whether or not particular forms of
“work” can be called “child labour” depends on the child’s age, the type and hours of work performed and the conditions under which
it is performed.[6]
She woke up in the motning with the most awful headache and she started talking abou thow everything was conectedMeasurement
of exposures and outcomes
"

2.3. Measurement of exposure

Data on the exposures of interest may be obtained through personal
interviews (either face-to-face or by telephone), self-administered questionnaires,
diaries of behaviour, reference to records, biological measurements
and measurements in the environment. If a subject is too young,
too il, or dead, it is also common to obtain data from a proxy respondent,
usually a member of their family.
The method chosen to collect data depends on many factors: the type
of study; the type and detail of data required; availabilty of existing
records collected for other purposes; lack of knowledge or poor recaIl of
the exposure by subjects; sensitivity of the subjects to questioning about
the exposure; frequency and level of the exposure, and their variabilty
over time; availabilty of physical or chemical methods for measuring the
exposure in the human body or in the environ ment; and the costs of the
various possible methods. Often, more than one approach is used.
Different components of the data often require different collection meth ods,
and using several methods of data collection can help to validate data
and to reduce error in measurement (see Section 2.6).
The information obtained should include details of the exact nature of
the exposure, its amount or dose, and its distribution over time.
2.3.1 Nature of the exposure
The information collected should be as detailed as possible. For
instance, it is better to enquire about different forms of tobacco smoking
separately (cigarettes, pipes, cigars), rather than to enquire simply about
tain
'smoking'. Questions on types of cigarette may also be asked to ob

information on their tar content. Enquiries should also be made about

the route of exposure to the agent (for example, in a study of contraceptives
and breast cancer, it is important to distinguish oral contraceptives
from other types of contraceptive), as weIl as about any behaviour
that may protect against exposure (for example, in an occupational
study, it is important to ask about any behaviour that may have protected
the workers from being exposed to hazards, such as use of protective
cIothing).
2.3.2 Dose
Exposure is seldom simply present or absent. Most exposures of interest
are quantitative variables. Smokers can be classified according to the number
of cigarettes smoked daily; industrial exposures by the extent of exposure
(often achieved by classifying workers according to the duration of
employment and type of job); infections by dose of agent or age at exposure;
breast-feeding by duration; and psychological exposures by sorne
arbitrary scale of severity. Thus the simple situation of two groups, one
exposed and one unexposed, is rare, and the conclusions of a study are
greatly strengthened where there is a trend of increasing disease incidence
with increasing exposure-an exposure-response relationship.
13
She woke up in the motning with the most awful headache and she started talking abou thow everything was conectedMeasurement
of exposures and outcomes
"

2.3. Measurement of exposure

Data on the exposures of interest may be obtained through personal
interviews (either face-to-face or by telephone), self-administered questionnaires,
diaries of behaviour, reference to records, biological measurements
and measurements in the environment. If a subject is too young,
too il, or dead, it is also common to obtain data from a proxy respondent,
usually a member of their family.
The method chosen to collect data depends on many factors: the type
of study; the type and detail of data required; availabilty of existing
records collected for other purposes; lack of knowledge or poor recaIl of
the exposure by subjects; sensitivity of the subjects to questioning about
the exposure; frequency and level of the exposure, and their variabilty
over time; availabilty of physical or chemical methods for measuring the
exposure in the human body or in the environ ment; and the costs of the
various possible methods. Often, more than one approach is used.
Different components of the data often require different collection meth ods,
and using several methods of data collection can help to validate data
and to reduce error in measurement (see Section 2.6).
The information obtained should include details of the exact nature of
the exposure, its amount or dose, and its distribution over time.
2.3.1 Nature of the exposure
The information collected should be as detailed as possible. For
instance, it is better to enquire about different forms of tobacco smoking
separately (cigarettes, pipes, cigars), rather than to enquire simply about
tain
'smoking'. Questions on types of cigarette may also be asked to ob

information on their tar content. Enquiries should also be made about

the route of exposure to the agent (for example, in a study of contraceptives
and breast cancer, it is important to distinguish oral contraceptives
from other types of contraceptive), as weIl as about any behaviour
that may protect against exposure (for example, in an occupational
study, it is important to ask about any behaviour that may have protected
the workers from being exposed to hazards, such as use of protective
cIothing).
2.3.2 Dose
Exposure is seldom simply present or absent. Most exposures of interest
are quantitative variables. Smokers can be classified according to the number
of cigarettes smoked daily; industrial exposures by the extent of exposure
(often achieved by classifying workers according to the duration of
employment and type of job); infections by dose of agent or age at exposure;
breast-feeding by duration; and psychological exposures by sorne
arbitrary scale of severity. Thus the simple situation of two groups, one
exposed and one unexposed, is rare, and the conclusions of a study are
greatly strengthened where there is a trend of increasing disease incidence
with increasing exposure-an exposure-response relationship.
13
She woke up in the motning with the most awful headache and she started talking abou thow everything was conectedMeasurement
of exposures and outcomes
"

2.3. Measurement of exposure

Data on the exposures of interest may be obtained through personal
interviews (either face-to-face or by telephone), self-administered questionnaires,
diaries of behaviour, reference to records, biological measurements
and measurements in the environment. If a subject is too young,
too il, or dead, it is also common to obtain data from a proxy respondent,
usually a member of their family.
The method chosen to collect data depends on many factors: the type
of study; the type and detail of data required; availabilty of existing
records collected for other purposes; lack of knowledge or poor recaIl of
the exposure by subjects; sensitivity of the subjects to questioning about
the exposure; frequency and level of the exposure, and their variabilty
over time; availabilty of physical or chemical methods for measuring the
exposure in the human body or in the environ ment; and the costs of the
various possible methods. Often, more than one approach is used.
Different components of the data often require different collection meth ods,
and using several methods of data collection can help to validate data
and to reduce error in measurement (see Section 2.6).
The information obtained should include details of the exact nature of
the exposure, its amount or dose, and its distribution over time.
2.3.1 Nature of the exposure
The information collected should be as detailed as possible. For
instance, it is better to enquire about different forms of tobacco smoking
separately (cigarettes, pipes, cigars), rather than to enquire simply about
tain
'smoking'. Questions on types of cigarette may also be asked to ob

information on their tar content. Enquiries should also be made about

the route of exposure to the agent (for example, in a study of contraceptives
and breast cancer, it is important to distinguish oral contraceptives
from other types of contraceptive), as weIl as about any behaviour
that may protect against exposure (for example, in an occupational
study, it is important to ask about any behaviour that may have protected
the workers from being exposed to hazards, such as use of protective
cIothing).
2.3.2 Dose
Exposure is seldom simply present or absent. Most exposures of interest
are quantitative variables. Smokers can be classified according to the number
of cigarettes smoked daily; industrial exposures by the extent of exposure
(often achieved by classifying workers according to the duration of
employment and type of job); infections by dose of agent or age at exposure;
breast-feeding by duration; and psychological exposures by sorne
arbitrary scale of severity. Thus the simple situation of two groups, one
exposed and one unexposed, is rare, and the conclusions of a study are
greatly strengthened where there is a trend of increasing disease incidence
with increasing exposure-an exposure-response relationship.
13
She woke up in the motning with the most awful headache and she started talking abou thow everything was conectedMeasurement
of exposures and outcomes
"

2.3. Measurement of exposure

Data on the exposures of interest may be obtained through personal
interviews (either face-to-face or by telephone), self-administered questionnaires,
diaries of behaviour, reference to records, biological measurements
and measurements in the environment. If a subject is too young,
too il, or dead, it is also common to obtain data from a proxy respondent,
usually a member of their family.
The method chosen to collect data depends on many factors: the type
of study; the type and detail of data required; availabilty of existing
records collected for other purposes; lack of knowledge or poor recaIl of
the exposure by subjects; sensitivity of the subjects to questioning about
the exposure; frequency and level of the exposure, and their variabilty
over time; availabilty of physical or chemical methods for measuring the
exposure in the human body or in the environ ment; and the costs of the
various possible methods. Often, more than one approach is used.
Different components of the data often require different collection meth ods,
and using several methods of data collection can help to validate data
and to reduce error in measurement (see Section 2.6).
The information obtained should include details of the exact nature of
the exposure, its amount or dose, and its distribution over time.
2.3.1 Nature of the exposure
The information collected should be as detailed as possible. For
instance, it is better to enquire about different forms of tobacco smoking
separately (cigarettes, pipes, cigars), rather than to enquire simply about
tain
'smoking'. Questions on types of cigarette may also be asked to ob

information on their tar content. Enquiries should also be made about

the route of exposure to the agent (for example, in a study of contraceptives
and breast cancer, it is important to distinguish oral contraceptives
from other types of contraceptive), as weIl as about any behaviour
that may protect against exposure (for example, in an occupational
study, it is important to ask about any behaviour that may have protected
the workers from being exposed to hazards, such as use of protective
cIothing).
2.3.2 Dose
Exposure is seldom simply present or absent. Most exposures of interest
are quantitative variables. Smokers can be classified according to the number
of cigarettes smoked daily; industrial exposures by the extent of exposure
(often achieved by classifying workers according to the duration of
employment and type of job); infections by dose of agent or age at exposure;
breast-feeding by duration; and psychological exposures by sorne
arbitrary scale of severity. Thus the simple situation of two groups, one
exposed and one unexposed, is rare, and the conclusions of a study are
greatly strengthened where there is a trend of increasing disease incidence
with increasing exposure-an exposure-response relationship.
13
She woke up in the motning with the most awful headache and she started talking abou thow everything was conectedMeasurement
of exposures and outcomes
"

2.3. Measurement of exposure

Data on the exposures of interest may be obtained through personal
interviews (either face-to-face or by telephone), self-administered questionnaires,
diaries of behaviour, reference to records, biological measurements
and measurements in the environment. If a subject is too young,
too il, or dead, it is also common to obtain data from a proxy respondent,
usually a member of their family.
The method chosen to collect data depends on many factors: the type
of study; the type and detail of data required; availabilty of existing
records collected for other purposes; lack of knowledge or poor recaIl of
the exposure by subjects; sensitivity of the subjects to questioning about
the exposure; frequency and level of the exposure, and their variabilty
over time; availabilty of physical or chemical methods for measuring the
exposure in the human body or in the environ ment; and the costs of the
various possible methods. Often, more than one approach is used.
Different components of the data often require different collection methods,
and using several methods of data collection can help to validate data
and to reduce error in measurement (see Section 2.6).
The information obtained should include details of the exact nature of
the exposure, its amount or dose, and its distribution over time.
2.3.1 Nature of the exposure
The information collected should be as detailed as possible. For
instance, it is better to enquire about different forms of tobacco smoking
separately (cigarettes, pipes, cigars), rather than to enquire simply about
tain
'smoking'. Questions on types of cigarette may also be asked to ob

information on their tar content. Enquiries should also be made about

the route of exposure to the agent (for example, in a study of contraceptives
and breast cancer, it is important to distinguish oral contraceptives
from other types of contraceptive), as weIl as about any behaviour
that may protect against exposure (for example, in an occupational
study, it is important to ask about any behaviour that may have protected
the workers from being exposed to hazards, such as use of protective
cIothing).
2.3.2 Dose
Exposure is seldom simply present or absent. Most exposures of interest
are quantitative variables. Smokers can be classified according to the number
of cigarettes smoked daily; industrial exposures by the extent of exposure
(often achieved by classifying workers according to the duration of
employment and type of job); infections by dose of agent or age at exposure;
breast-feeding by duration; and psychological exposures by sorne
arbitrary scale of severity. Thus the simple situation of two groups, one
exposed and one unexposed, is rare, and the conclusions of a study are
greatly strengthened where there is a trend of increasing disease incidence
with increasing exposure-an exposure-response relationship.
13
She woke up in the motning with the most awful headache and she started talking abou thow everything was conectedMeasurement
of exposures and outcomes
"

2.3. Measurement of exposure

Data on the exposures of interest may be obtained through personal
interviews (either face-to-face or by telephone), self-administered questionnaires,
diaries of behaviour, reference to records, biological measurements
and measurements in the environment. If a subject is too young,
too il, or dead, it is also common to obtain data from a proxy respondent,
usually a member of their family.
The method chosen to collect data depends on many factors: the type
of study; the type and detail of data required; availabilty of existing
records collected for other purposes; lack of knowledge or poor recaIl of
the exposure by subjects; sensitivity of the subjects to questioning about
the exposure; frequency and level of the exposure, and their variabilty
over time; availabilty of physical or chemical methods for measuring the
exposure in the human body or in the environ ment; and the costs of the
various possible methods. Often, more than one approach is used.
Different components of the data often require different collection methods,
and using several methods of data collection can help to validate data
and to reduce error in measurement (see Section 2.6).
The information obtained should include details of the exact nature of
the exposure, its amount or dose, and its distribution over time.
2.3.1 Nature of the exposure
The information collected should be as detailed as possible. For
instance, it is better to enquire about different forms of tobacco smoking
separately (cigarettes, pipes, cigars), rather than to enquire simply about
tain
'smoking'. Questions on types of cigarette may also be asked to ob

information on their tar content. Enquiries should also be made about

the route of exposure to the agent (for example, in a study of contraceptives
and breast cancer, it is important to distinguish oral contraceptives
from other types of contraceptive), as weIl as about any behaviour
that may protect against exposure (for example, in an occupational
study, it is important to ask about any behaviour that may have protected
the workers from being exposed to hazards, such as use of protective
cIothing).
2.3.2 Dose
Exposure is seldom simply present or absent. Most exposures of interest
are quantitative variables. Smokers can be classified according to the number
of cigarettes smoked daily; industrial exposures by the extent of exposure
(often achieved by classifying workers according to the duration of
employment and type of job); infections by dose of agent or age at exposure;
breast-feeding by duration; and psychological exposures by sorne
arbitrary scale of severity. Thus the simple situation of two groups, one
exposed and one unexposed, is rare, and the conclusions of a study are
greatly strengthened where there is a trend of increasing disease incidence
with increasing exposure-an exposure-response relationship.
13
She woke up in the motning with the most awful headache and she started talking abou thow everything was conectedMeasurement
of exposures and outcomes
"

2.3. Measurement of exposure

Data on the exposures of interest may be obtained through personal
interviews (either face-to-face or by telephone), self-administered questionnaires,
diaries of behaviour, reference to records, biological measurements
and measurements in the environment. If a subject is too young,
too il, or dead, it is also common to obtain data from a proxy respondent,
usually a member of their family.
The method chosen to collect data depends on many factors: the type
of study; the type and detail of data required; availabilty of existing
records collected for other purposes; lack of knowledge or poor recaIl of
the exposure by subjects; sensitivity of the subjects to questioning about
the exposure; frequency and level of the exposure, and their variabilty
over time; availabilty of physical or chemical methods for measuring the
exposure in the human body or in the environ ment; and the costs of the
various possible methods. Often, more than one approach is used.
Different components of the data often require different collection meth ods,
and using several methods of data collection can help to validate data
and to reduce error in measurement (see Section 2.6).
The information obtained should include details of the exact nature of
the exposure, its amount or dose, and its distribution over time.
2.3.1 Nature of the exposure
The information collected should be as detailed as possible. For
instance, it is better to enquire about different forms of tobacco smoking
separately (cigarettes, pipes, cigars), rather than to enquire simply about
tain
'smoking'. Questions on types of cigarette may also be asked to ob

information on their tar content. Enquiries should also be made about

the route of exposure to the agent (for example, in a study of contraceptives
and breast cancer, it is important to distinguish oral contraceptives
from other types of contraceptive), as weIl as about any behaviour
that may protect against exposure (for example, in an occupational
study, it is important to ask about any behaviour that may have pro tected
the workers from being exposed to hazards, such as use of protective
cIothing).
2.3.2 Dose
Exposure is seldom simply present or absent. Most exposures of interest
are quantitative variables. Smokers can be classified according to the number
of cigarettes smoked daily; industrial exposures by the extent of exposure
(often achieved by classifying workers according to the duration of
employment and type of job); infections by dose of agent or age at exposure;
breast-feeding by duration; and psychological exposures by sorne
arbitrary scale of severity. Thus the simple situation of two groups, one
exposed and one unexposed, is rare, and the conclusions of a study are
greatly strengthened where there is a trend of increasing disease incidence
with increasing exposure-an exposure-response relationship.
13
She woke up in the motning with the most awful headache and she started talking abou thow everything was conectedMeasurement
of exposures and outcomes
"

2.3. Measurement of exposure

Data on the exposures of interest may be obtained through personal
interviews (either face-to-face or by telephone), self-administered questionnaires,
diaries of behaviour, reference to records, biological measurements
and measurements in the environment. If a subject is too young,
too il, or dead, it is also common to obtain data from a proxy respondent,
usually a member of their family.
The method chosen to collect data depends on many factors: the type
of study; the type and detail of data required; availabilty of existing
records collected for other purposes; lack of knowledge or poor recaIl of
the exposure by subjects; sensitivity of the subjects to questioning about
the exposure; frequency and level of the exposure, and their variabilty
over time; availabilty of physical or chemical methods for measuring the
exposure in the human body or in the environ ment; and the costs of the
various possible methods. Often, more than one approach is used.
Different components of the data often require different collection meth ods,
and using several methods of data collection can help to validate data
and to reduce error in measurement (see Section 2.6).
The information obtained should include details of the exact nature of
the exposure, its amount or dose, and its distribution over time.
2.3.1 Nature of the exposure
The information collected should be as detailed as possible. For
instance, it is better to enquire about different forms of tobacco smoking
separately (cigarettes, pipes, cigars), rather than to enquire simply about
tain
'smoking'. Questions on types of cigarette may also be asked to ob

information on their tar content. Enquiries should also be made about

the route of exposure to the agent (for example, in a study of contraceptives
and breast cancer, it is important to distinguish oral contraceptives
from other types of contraceptive), as weIl as about any behaviour
that may protect against exposure (for example, in an occupational
study, it is important to ask about any behaviour that may have protected
the workers from being exposed to hazards, such as use of protective
cIothing).
2.3.2 Dose
Exposure is seldom simply present or absent. Most exposures of interest
are quantitative variables. Smokers can be classified according to the number
of cigarettes smoked daily; industrial exposures by the extent of exposure
(often achieved by classifying workers according to the duration of
employment and type of job); infections by dose of agent or age at exposure;
breast-feeding by duration; and psychological exposures by sorne
arbitrary scale of severity. Thus the simple situation of two groups, one
exposed and one unexposed, is rare, and the conclusions of a study are
greatly strengthened where there is a trend of increasing disease incidence
with increasing exposure-an exposure-response relationship.
13
She woke up in the motning with the most awful headache and she started talking abou thow everything was conectedMeasurement
of exposures and outcomes
"

2.3. Measurement of exposure

Data on the exposures of interest may be obtained through personal
interviews (either face-to-face or by telephone), self-administered questionnaires,
diaries of behaviour, reference to records, biological measurements
and measurements in the environment. If a subject is too young,
too il, or dead, it is also common to obtain data from a proxy respondent,
usually a member of their family.
The method chosen to collect data depends on many factors: the type
of study; the type and detail of data required; availabilty of existing
records collected for other purposes; lack of knowledge or poor recaIl of
the exposure by subjects; sensitivity of the subjects to questioning about
the exposure; frequency and level of the exposure, and their variabilty
over time; availabilty of physical or chemical methods for measuring the
exposure in the human body or in the environ ment; and the costs of the
various possible methods. Often, more than one approach is used.
Different components of the data often require different collection meth ods,
and using several methods of data collection can help to validate data
and to reduce error in measurement (see Section 2.6).
The information obtained should include details of the exact nature of
the exposure, its amount or dose, and its distribution over time.
2.3.1 Nature of the exposure
The information collected should be as detailed as possible. For
instance, it is better to enquire about different forms of tobacco smoking
separately (cigarettes, pipes, cigars), rather than to enquire simply about
 tain
'smoking'. Questions on types of cigarette may also be asked to ob

information on their tar content. Enquiries should also be made about

the route of exposure to the agent (for example, in a study of contraceptives
and breast cancer, it is important to distinguish oral contraceptives
from other types of contraceptive), as weIl as about any behaviour
that may protect against exposure (for example, in an occupational
study, it is important to ask about any behaviour that may have protected
the workers from being exposed to hazards, such as use of protective
cIothing).
2.3.2 Dose
Exposure is seldom simply present or absent. Most exposures of interest
are quantitative variables. Smokers can be classified according to the number
of cigarettes smoked daily; industrial exposures by the extent of exposure
(often achieved by classifying workers according to the duration of
employment and type of job); infections by dose of agent or age at exposure;
breast-feeding by duration; and psychological exposures by sorne
arbitrary scale of severity. Thus the simple situation of two groups, one
exposed and one unexposed, is rare, and the conclusions of a study are
greatly strengthened where there is a trend of increasing disease incidence
with increasing exposure-an exposure-response relationship.
13
She woke up in the motning with the most awful headache and she started talking abou thow everything was conectedMeasurement
of exposures and outcomes
"

2.3. Measurement of exposure

Data on the exposures of interest may be obtained through personal
interviews (either face-to-face or by telephone), self-administered questionnaires,
diaries of behaviour, reference to records, biological measurements
and measurements in the environment. If a subject is too young,
too il, or dead, it is also common to obtain data from a proxy respondent,
usually a member of their family.
The method chosen to collect data depends on many factors: the type
of study; the type and detail of data required; availabilty of existing
records collected for other purposes; lack of knowledge or poor recaIl of
the exposure by subjects; sensitivity of the subjects to questioning about
the exposure; frequency and level of the exposure, and their variabilty
over time; availabilty of physical or chemical methods for measuring the
exposure in the human body or in the environ ment; and the costs of the
various possible methods. Often, more than one approach is used.
Different components of the data often require different collection meth ods,
and using several methods of data collection can help to validate data
and to reduce error in measurement (see Section 2.6).
The information obtained should include details of the exact nature of
the exposure, its amount or dose, and its distribution over time.
2.3.1 Nature of the exposure
The information collected should be as detailed as possible. For
instance, it is better to enquire about different forms of tobacco smoking
separately (cigarettes, pipes, cigars), rather than to enquire simply about
tain
'smoking'. Questions on types of cigarette may also be asked to ob

information on their tar content. Enquiries should also be made about

the route of exposure to the agent (for example, in a study of contraceptives
and breast cancer, it is important to distinguish oral contraceptives
from other types of contraceptive), as weIl as about any behaviour
that may protect against exposure (for example, in an occupational
study, it is important to ask about any behaviour that may have protected
the workers from being exposed to hazards, such as use of protective
cIothing).
2.3.2 Dose
Exposure is seldom simply present or absent. Most exposures of interest
are quantitative variables. Smokers can be classified according to the number
of cigarettes smoked daily; industrial exposures by the extent of exposure
(often achieved by classifying workers according to the duration of
employment and type of job); infections by dose of agent or age at exposure;
breast-feeding by duration; and psychological exposures by sorne
arbitrary scale of severity. Thus the simple situation of two groups, one
exposed and one unexposed, is rare, and the conclusions of a study are
greatly strengthened where there is a trend of increasing disease incidence
with increasing exposure-an exposure-response relationship.
13
She woke up in the motning with the most awful headache and she started talking abou thow everything was conectedMeasurement
of exposures and outcomes
"

2.3. Measurement of exposure

Data on the exposures of interest may be obtained through personal
interviews (either face-to-face or by telephone), self-administered questionnaires,
diaries of behaviour, reference to records, biological measurements
and measurements in the environment. If a subject is too young,
too il, or dead, it is also common to obtain data from a proxy respondent,
usually a member of their family.
The method chosen to collect data depends on many factors: the type
of study; the type and detail of data required; availabilty of existing
records collected for other purposes; lack of knowledge or poor recaIl of
the exposure by subjects; sensitivity of the subjects to questioning about
the exposure; frequency and level of the exposure, and their variabilty
over time; availabilty of physical or chemical methods for measuring the
exposure in the human body or in the environ ment; and the costs of the
various possible methods. Often, more than one approach is used.
Different components of the data often require different collection meth ods,
and using several methods of data collection can help to validate data
and to reduce error in measurement (see Section 2.6).
The information obtained should include details of the exact nature of
the exposure, its amount or dose, and its distribution over time.
2.3.1 Nature of the exposure
The information collected should be as detailed as possible. For
instance, it is better to enquire about different forms of tobacco smoking
separately (cigarettes, pipes, cigars), rather than to enquire simply about
tain
'smoking'. Questions on types of cigarette may also be asked to ob

information on their tar content. Enquiries should also be made about

the route of exposure to the agent (for example, in a study of contraceptives
and breast cancer, it is important to distinguish oral contraceptives
from other types of contraceptive), as weIl as about any behaviour
that may protect against exposure (for example, in an occupational
study, it is important to ask about any behaviour that may have protected
the workers from being exposed to hazards, such as use of protective
cIothing).
2.3.2 Dose
Exposure is seldom simply present or absent. Most exposures of interest
are quantitative variables. Smokers can be classified according to the number
of cigarettes smoked daily; industrial exposures by the extent of exposure
(often achieved by classifying workers according to the duration of
employment and type of job); infections by dose of agent or age at exposure;
breast-feeding by duration; and psychological exposures by sorne
arbitrary scale of severity. Thus the simple situation of two groups, one
exposed and one unexposed, is rare, and the conclusions of a study are
greatly strengthened where there is a trend of increasing disease incidence
with increasing exposure-an exposure-response relationship.
13
She woke up in the motning with the most awful headache and she started talking abou thow everything was conectedMeasurement
of exposures and outcomes
"

2.3. Measurement of exposure

Data on the exposures of interest may be obtained through personal
interviews (either face-to-face or by telephone), self-administered questionnaires,
diaries of behaviour, reference to records, biological measurements
and measurements in the environment. If a subject is too young,
too il, or dead, it is also common to obtain data from a proxy respondent,
usually a member of their family.
The method chosen to collect data depends on many factors: the type
of study; the type and detail of data required; availabilty of existing
records collected for other purposes; lack of knowledge or poor recaIl of
the exposure by subjects; sensitivity of the subjects to questioning about
the exposure; frequency and level of the exposure, and their variabilty
over time; availabilty of physical or chemical methods for measuring the
exposure in the human body or in the environ ment; and the costs of the
various possible methods. Often, more than one approach is used.
Different components of the data often require different collection methods,
and using several methods of data collection can help to validate data
and to reduce error in measurement (see Section 2.6).
The information obtained should include details of the exact nature of
the exposure, its amount or dose, and its distribution over time.
2.3.1 Nature of the exposure
The information collected should be as detailed as possible. For
instance, it is better to enquire about different forms of tobacco smoking
separately (cigarettes, pipes, cigars), rather than to enquire simply about
tain
'smoking'. Questions on types of cigarette may also be asked to ob

information on their tar content. Enquiries should also be made about

the route of exposure to the agent (for example, in a study of contraceptives
and breast cancer, it is important to distinguish oral contraceptives
from other types of contraceptive), as weIl as about any behaviour
that may protect against exposure (for example, in an occupational
study, it is important to ask about any behaviour that may have protected
the workers from being exposed to hazards, such as use of protective
cIothing).
2.3.2 Dose
Exposure is seldom simply present or absent. Most exposures of interest
are quantitative variables. Smokers can be classified according to the number
of cigarettes smoked daily; industrial exposures by the extent of exposure
(often achieved by classifying workers according to the duration of
employment and type of job); infections by dose of agent or age at exposure;
breast-feeding by duration; and psychological exposures by sorne
arbitrary scale of severity. Thus the simple situation of two groups, one
exposed and one unexposed, is rare, and the conclusions of a study are
greatly strengthened where there is a trend of increasing disease incidence
with increasing exposure-an exposure-response relationship.
13
She woke up in the motning with the most awful headache and she started talking abou thow everything was conectedMeasurement
of exposures and outcomes
"

2.3. Measurement of exposure

Data on the exposures of interest may be obtained through personal
interviews (either face-to-face or by telephone), self-administered questionnaires,
diaries of behaviour, reference to records, biological measurements
and measurements in the environment. If a subject is too young,
too il, or dead, it is also common to obtain data from a proxy respondent,
usually a member of their family.
The method chosen to collect data depends on many factors: the type
of study; the type and detail of data required; availabilty of existing
records collected for other purposes; lack of knowledge or poor recaIl of
the exposure by subjects; sensitivity of the subjects to questioning about
the exposure; frequency and level of the exposure, and their variabilty
over time; availabilty of physical or chemical methods for measuring the
exposure in the human body or in the environ ment; and the costs of the
various possible methods. Often, more than one approach is used.
Different components of the data often require different collection meth ods,
and using several methods of data collection can help to validate data
and to reduce error in measurement (see Section 2.6).
The information obtained should include details of the exact nature of
the exposure, its amount or dose, and its distribution over time.
2.3.1 Nature of the exposure
The information collected should be as detailed as possible. For
instance, it is better to enquire about different forms of tobacco smoking
separately (cigarettes, pipes, cigars), rather than to enquire simply about
tain
'smoking'. Questions on types of cigarette may also be asked to ob

information on their tar content. Enquiries should also be made about

the route of exposure to the agent (for example, in a study of contraceptives
and breast cancer, it is important to distinguish oral contraceptives
from other types of contraceptive), as weIl as about any behaviour
that may protect against exposure (for example, in an occupational
study, it is important to ask about any behaviour that may have protected
the workers from being exposed to hazards, such as use of protective
cIothing).
2.3.2 Dose
Exposure is seldom simply present or absent. Most exposures of interest
are quantitative variables. Smokers can be classified according to the number
of cigarettes smoked daily; industrial exposures by the extent of exposure
(often achieved by classifying workers according to the duration of
employment and type of job); infections by dose of agent or age at exposure;
breast-feeding by duration; and psychological exposures by sorne
arbitrary scale of severity. Thus the simple situation of two groups, one
exposed and one unexposed, is rare, and the conclusions of a study are
greatly strengthened where there is a trend of increasing disease incidence
with increasing exposure-an exposure-response relationship.
13