Summary I: Clinical Psychology: A Modern
Health Profession
Chapter 1: Being a Clinical Psychologist
Clinical psychologists can be seen as scientist-practitioners, because they
research the questions they deal with in clinical practice, and vice-versa.
The chapter describes what the average days of a psychology student and two
clinical psychologists look like.
Chapter 2: Becoming a Clinical Psychologist
The chapter describes how to become a clinical psychologist (grades needed,
reference letters, etc) and raises some questions about which type of field might
be best for what kind of person.
Chapter 3: Methods for Research and Evaluation
Properties of Psychological Tests
Reliability
Reliability (r): A test’s ability to produce the same results over its repeated
administrations.
Inter-rater reliability (r, or kappa): If researchers have trained their observers
to extract the same information, this will be high. Ideally, behavior is videotaped
so it can later be observed. It is acceptable when all behavior is observed by one
rater, and then later 10-20% of the observations is re-assessed by a second rater.
Internal consistency (Cronbach’s alpha): Determines how useful each item is
for measuring the overall construct.
Split-half reliability: The number of items measuring the same construct is
split in two equal halves and reliability is compared.
Test-retest reliability: Comparing reliability at two different time points.
Validity
Validity: When a test measures what it is supposed to measure.
Face validity: When it is very obvious what a test is trying to measure. It can be
good or bad when test takers immediately know what the test’s aim is.
Content validity: When a test is based on information obtained from experts in
the field.
Concurrent validity: Results of a (new) test are similar to those of an already
existing test.
Construct validity: When a test measures the construct it is supposed to
measure (it needs to show predicted links with other aspects of anxiety, for
example, and needs to show connections to other aspects related to distress).
Criterion validity: The test scores represent criteria that can be used to make a
decision (for example, they indicate whether someone is at risk).
, Predictive validity: When a test is able to make an accurate prediction about
the future.
When a new test is developed, it is important to evaluate all aspects of validity
and reliability. For example, the WAIS is predictive of performance from
elementary to high school, but not later in life, and therefore has limited
predictive validity.
Methods Used to Learn about Therapy Outcome
It is important to measure whether a form of therapy is effective. There are
multiple ways:
Case studies: This can be in the form of a narrative about an individual or
via conducting an experiment with one individual. When the latter is done,
it is often in the form of an ABAB experiment: A four-phase protocol
where baseline and treatment phases alternate twice: Baseline normal
behavior – treatment – no treatment – second treatment stage;
Therapy outcome research: Results can be compromised by:
o Confounding treatments: E.g., participants start taking a drug as
well as have a psychological treatment;
o Placebo effect/history effect randomized controlled trial
(RCT): participants are randomly divided in a control and in an
experimental group. The control group can be a wait-list group:
Appealing for participants (they are promised treatment at some
point) and for researchers (because it is ethical). The idea that all
patients should receive the same treatment is called the equipoise
principle.
The study can also be single-blind: The participant does not know
whether s/he receives treatment (e.g., placebo control). Double
blind: Both researcher and participant do not know whether the
participant receives treatment.
In psychotherapy, it is sometimes possible to have a placebo-
control; however, it is often more important to know whether the
treatment is better than pre-existing treatments (treatment effect
specificity).
Results can be significant, without being clinically significant: the
result is not significant to society.
Within studies, the greatest value is usually placed on RCTs and
combinations of RCTs in meta-analyses (even though these also
know some limitations).
o Qualitative research: It is not hypothesis-driven and therefore
allows for new insights. One idea now is that different parts of the
research (interviews, analysis) should be done by different people.
Problems with qualitative research: problems with reliability,
reliability of extraction methods, subjectivity of the interpretation
process;
o Program evaluation: Programs that are meant to reflect and guide
clinical practice, and the evaluation is meant to find out whether it is
needed, by whom, etc. Steps:
1. Assessment of needs: Is there a need for a certain
program?
Health Profession
Chapter 1: Being a Clinical Psychologist
Clinical psychologists can be seen as scientist-practitioners, because they
research the questions they deal with in clinical practice, and vice-versa.
The chapter describes what the average days of a psychology student and two
clinical psychologists look like.
Chapter 2: Becoming a Clinical Psychologist
The chapter describes how to become a clinical psychologist (grades needed,
reference letters, etc) and raises some questions about which type of field might
be best for what kind of person.
Chapter 3: Methods for Research and Evaluation
Properties of Psychological Tests
Reliability
Reliability (r): A test’s ability to produce the same results over its repeated
administrations.
Inter-rater reliability (r, or kappa): If researchers have trained their observers
to extract the same information, this will be high. Ideally, behavior is videotaped
so it can later be observed. It is acceptable when all behavior is observed by one
rater, and then later 10-20% of the observations is re-assessed by a second rater.
Internal consistency (Cronbach’s alpha): Determines how useful each item is
for measuring the overall construct.
Split-half reliability: The number of items measuring the same construct is
split in two equal halves and reliability is compared.
Test-retest reliability: Comparing reliability at two different time points.
Validity
Validity: When a test measures what it is supposed to measure.
Face validity: When it is very obvious what a test is trying to measure. It can be
good or bad when test takers immediately know what the test’s aim is.
Content validity: When a test is based on information obtained from experts in
the field.
Concurrent validity: Results of a (new) test are similar to those of an already
existing test.
Construct validity: When a test measures the construct it is supposed to
measure (it needs to show predicted links with other aspects of anxiety, for
example, and needs to show connections to other aspects related to distress).
Criterion validity: The test scores represent criteria that can be used to make a
decision (for example, they indicate whether someone is at risk).
, Predictive validity: When a test is able to make an accurate prediction about
the future.
When a new test is developed, it is important to evaluate all aspects of validity
and reliability. For example, the WAIS is predictive of performance from
elementary to high school, but not later in life, and therefore has limited
predictive validity.
Methods Used to Learn about Therapy Outcome
It is important to measure whether a form of therapy is effective. There are
multiple ways:
Case studies: This can be in the form of a narrative about an individual or
via conducting an experiment with one individual. When the latter is done,
it is often in the form of an ABAB experiment: A four-phase protocol
where baseline and treatment phases alternate twice: Baseline normal
behavior – treatment – no treatment – second treatment stage;
Therapy outcome research: Results can be compromised by:
o Confounding treatments: E.g., participants start taking a drug as
well as have a psychological treatment;
o Placebo effect/history effect randomized controlled trial
(RCT): participants are randomly divided in a control and in an
experimental group. The control group can be a wait-list group:
Appealing for participants (they are promised treatment at some
point) and for researchers (because it is ethical). The idea that all
patients should receive the same treatment is called the equipoise
principle.
The study can also be single-blind: The participant does not know
whether s/he receives treatment (e.g., placebo control). Double
blind: Both researcher and participant do not know whether the
participant receives treatment.
In psychotherapy, it is sometimes possible to have a placebo-
control; however, it is often more important to know whether the
treatment is better than pre-existing treatments (treatment effect
specificity).
Results can be significant, without being clinically significant: the
result is not significant to society.
Within studies, the greatest value is usually placed on RCTs and
combinations of RCTs in meta-analyses (even though these also
know some limitations).
o Qualitative research: It is not hypothesis-driven and therefore
allows for new insights. One idea now is that different parts of the
research (interviews, analysis) should be done by different people.
Problems with qualitative research: problems with reliability,
reliability of extraction methods, subjectivity of the interpretation
process;
o Program evaluation: Programs that are meant to reflect and guide
clinical practice, and the evaluation is meant to find out whether it is
needed, by whom, etc. Steps:
1. Assessment of needs: Is there a need for a certain
program?
