Week 1
Meehl (1920 – 2003)
Paul Meehl (1920-2003) was een klinisch psycholoog die ook veel heeft geschreven over
wetenschapsfilosofische onderwerpen. Hij was zeer kritisch over de manier waarop psychologen hun
theorieën toetsen en niet optimistisch over de kans dat dat zou verbeteren. In dit artikel stelt hij de
vraag: waarom gaat de psychologie zo langzaam vooruit (cumulatie)? Met name in de 'zachte'
psychologie, zoals klinische en sociale psychologie, is er volgens Meehl weinig sprake van
wetenschappelijke vooruitgang. Hij bespreekt twee oorzaken: de aard van het object van onderzoek,
en het gebruik van nulhypothese significatietoetsen.
In de bijvoorbeeld de sociale psychologie wordt een theorie totaal geaccepteerd en op een gegeven
moment verliezen mensen hun interesse hierin in plaats van dat de theorie aangepast wordt o.i.d. Er
vindt dus minder cumulatie van wetenschap plaats. In dit artikel worden er twintig moeilijkheden van
psychologie besproken:
o Probleem van het classificeren van antwoorden
o Probleem van de situatie taxonomie
o Eenheid van meting
o Individuele verschillen
o Polygene erfelijkheid
o Divergente causaliteit
o Idiografisch probleem
o Onbekende kritieke gebeurtenissen
o Hinderlijke variabelen
o Feedback loops
o Autokatalytische processen
o Random walk; de levensloop van iemand
o Een groot aantal variabelen
o Belang van culturele factoren
o Context afhankelijke stochasten
o Open concepten
o Intentionaliteit, bedoeling en betekenis
o Regel bestuur
o Unieke menselijke gebeurtenissen en power
o Ethische beperkingen in onderzoek
Soft psychology is the term for the areas in psychology such as clinical psychology, counselling, social,
personality and school psychology that lack in cumulative scientific knowledge. The scientific findings
in these areas tend to not be disproved nor validated. Instead it seems that theories in soft
psychology come and go. This leads to the lack of cumulative scientific knowledge, which is not true
for disciplines such as astronomy, molecular biology and genetics. The author of this article came up
with 20 reasons for why there is a lack of cumulative scientific knowledge that leads to that these
areas in psychology are called soft psychology. Each of the difficulties presented is debatable.
Twenty difficulties in conducting research in certain areas of psychology
1. The response-class problem
This difficulty or problem is about that it is difficult to slice up behaviour in measurable units. For
example, what is important to getting a degree? Is it sitting an exam, writing a sentence, writing a
,letter or is the firing of a neuron that is responsible for the movement of the arm and hand
important? This problem is already evident in the experiment of the Skinner box and becomes even
more problematic when studying human behaviour.
2. Situation-taxonomy problem
This is kind of the same as the response-class problem, but this time it is applied to stimuli or to
situations. For example, every theory of personality defines behaviour as a function of the person
and the situation. So, the situation should be defined: is it the whole culture, the region a person
lives in, a town, a house, a room?
3. Unit of measurement
This is about the questions in rating scales and in psychometrics. This is a problem, because there
often is no agreement in whether, for example, an interval or a ratio scale should be used.
4. Individual differences
Individuals differ in their dispositions and even more problematic, in the way their dispositions are
shaped and organised. This makes it difficult for researchers to conduct research and compare
individuals on the results.
5. Polygenic heredity
Most of the research done in the soft areas of psychology are things that are influenced by polygenic
systems. This means that there are many different influences on for example, introversion. In other
words, introversion is probably the result of a confluence of different polygenic contributors such as
anxiety, dominance, need for affiliation, etc.
6. Divergent causality
This difficulty is about that a small event can lead to many different (divergent) outcomes. For
example, many students marry other students. However, the way that they get to know these other
students varies from individual to individual and it depends on things such as to which classroom
they get allocated.
7. Idiographic problem
There are many factors which influence certain things, for example personality. These factors also
interact and therefore every individual is quite unique. This is a problem for research, because the
research methods should therefore be extremely large and complex. This is the same for the
statistical analysis, which should also be sophisticated enough to analyse all of the factors and their
numerous interactions.
8. Unknown critical events
This problem is related to divergent causality and the idiographic understanding. This problem is
about that it is hard to know which events in the development of personality have been critical.
Sometimes these critical events are unknown to the researchers as well as to the individual. These
events can also be inner events such as fantasies.
9. Nuisance variables
For research it is good to sort variables in three classes: variables which we manipulate, variables
which we do not manipulate but can control (or keep constant to reduce the influence of them on
the results) and variables that are ‘quasi random’, which means that they only contribute to the
measurement error or the standard deviation of a static. However, there are also variables which are
not random but which are systematic and are influenced by other variables. The problem is that it is
not always easy to determine the size and the sign of these relationships. One of the biggest of these
,‘nuisance variables’ is socioeconomic status, which is very hard to control for, because this variable
has a big influence on many aspects of people’s lives.
10. Feedback loops
In psychology it is known that an individual’s behaviour affects the behaviour of other people. This
fact also makes that conducting research on behaviour can be very complex.
11. Autocatalytic processes
Autocatalysis means that one of the end products leads to catalysing the process itself. In
psychology, examples of these processes are depression and anxiety as affects or economic failure as
a social impact. It is very difficult to get a clear view of situations in which autocatalytic processes are
involved.
12. Random walk
Things happen randomly sometimes, so the quantification of this behaviour becomes difficult. For
instance, luck is one of the biggest reasons for individual differences. Social scientists seem to neglect
this in their research sometimes.
13. Sheer number of variables
There are many variables which contribute to a psychological disorder or to a personality trait.
Because of this, there are also a lot of nuisance variables and all of these variables also lead to
idiographic development through divergent causality.
14. The importance of cultural factors
Sometimes researchers forget to include cultural factors which lead to certain disorders or diseases.
For example, a physician will treat someone with a stomach ache in the same way, regardless of their
social class. This is a problem, because sometimes a problem can be solved or prevented by looking
at cultural factors (e.g. someone belongs to the lowest class regarding SES and therefore he or she
has many deficits in his or her diet which lead to stomach ache).
15. Context-dependent stochastotogicals
In psychology, the term ‘nomological network’ is used to represent the relations between constructs.
The constructs are ‘nodes’ which are linked to other constructs through strands. Meehl states that
‘nomological’ in psychology is a synonym for ‘meaning’ and it could even be misleading. He puts it
like this, because the relationships are not as strict as suggested by calling it a nomological network.
In psychology, instead of strict relationships, there are correlations, tendencies, statistical clustering,
increments of probabilities and stochastic dispositions. The author therefore suggests a
“stochastological network”. With this he tries to make clear that the results obtained in psychological
experiments are dependent on the context in which the measurements were obtained, because it is
hard to determine what influences what and how big this influence is. So, in soft psychology, the
context is very important. However, we do not know the complete list of the contextual influences,
the function form of context dependency for the known influences and the numerical values of
parameters in those functions and lastly, the values of the context variables.
16. Open concepts
It is often hard to come up with operational definitions for theoretical concepts. Even if researchers
come up with operational definitions, these seem to lack theoretical background and technologies to
measure them. This means that there are a lot of ‘open concepts’ in psychology, which further
complicates research.
17. Intentionality, purpose and meaning
Humans think, plan and intend way more than any other species. This makes them stand out
, compared to other species, but again leads to complexity in researching their behaviour.
18. Rule governance
People or human beings do not only do things out of inner motivation, they also do things because
they think that this is what is expected from them. They feel obligated to follow certain rules. In
research there is no clear definition of when a rule can be considered as a rule. So when conducting
research, it is sometimes difficult to determine whether one performs some kind of behaviour
because of real, internal motivation or because he or she is following a certain rule.
19. Uniquely human events and powers
Because many of the research in psychology, such as digestion and reproduction, comes from studies
with animals, some things are very difficult to research. For example, Richard M. Elliot stated that the
reason for the lack of high quality research on humour is that humans are the only species that laugh.
So, because some things are uniquely human, it makes it difficult to conduct research. Except for
laughing, other examples that are unique to humans are seeking revenge for things that happened
years earlier or have discussions on discussions.
20. Ethical constraints on research
There are many ethical constraints in research in psychology. For example, it is unethical to let a
group of children from as old as 10 years smoke and then compare them to children who have never
smoked. Meehl names five noble traditions in psychology that have lead to great insights. In the
times of these traditions, the ethical constraints were a lot less than that they are now. The five
noble traditions are: descriptive clinical psychiatry, psychometric assessment, behaviour genetics,
behaviour modification and psychodynamics. He states that Freud did not perform any t-tests, but
that he would choose Freud’s clinical observations over most of the researchers’ t-tests. All of these
traditions have not focused on statistical significance testing.
