Goals of Science and Scientific Methods

Goals of Scientific Method
The scientific method has four goals: description, prediction, explanation, and
application.
Figure: 1.02 Goals of Scientific Method

 Description
Social scientists try to find and describe the events and relationship among the factors
causing the events. Researchers mostly use nomothetic approach (Nomothetic approach
includes wide range of subjects or participant in the study and determined their
performance on average basis using the normal distribution, measurement of central
tendency, and measurement of desperation) and quantitative analysis. Description is the
process which researchers use to explain, define, and categorize the events and
relationships among the factors. For example Clinical psychologists use the DSM-V to
classify the behavior and disorders on the basis of symptoms.
 Prediction
Co relational relationship permit researchers to predict about behavior and events but do
not allows them to deduce that what the causes of this relationship are?
Prediction is the declaration or set of statements about what will be observed
prior to it actually present. It is the complicated operation involving comprehensive
knowledge of phenomena under deliberation, including related facts and vital principles.
Description of events and their relationship frequently provide the foundation to
prediction.
For example: does children of broken family or separated family are more
vulnerable to develop emotional maladjustment? Do the illiteracy and poverty are the causes of terrorism? These information not-only provide an immense knowledge in the
fields of social sciences but also helpful in prediction, action, and prevention.
 Explanation
There are three conditions to assert the causal inference: co variation, time-order
relationship, elimination of plausible alternative causes.
A simple example will help you to understand these three conditions. Suppose
you walk for a mile and feel pain in your legs. Presumably you would infer that walking
for a mile causes pain in legs. This is the first condition of causal inference, that waking
for a mile co varies with leg’s pain. When one condition changes this automatically
creates changes to other.
Time-order relationship is the second condition for causal inference. It is also
known as contingency. In first condition pain occurred after walking for a mile. If the
pain occurs before you walk, then you would not infer that walking for a mile causes pain
in legs.
Finally, causal explanations are accepted only when other possible causes have
been removed—when plausible alternative causes have been eliminated.
Threats to the studies of psychology occur when there are deficiencies
(confounding) occur during the conducting of psychological studies. These threats are
dangers to validity of psychological research or studies. In this chapter we will discuss
that how and from where these threats come and how can we control or minimize these
threats.
 Application
In applied research, researchers apply their knowledge and findings to improve people’s
lives. Application is generalizability (external validity) of results (scientifically obtained
usually from experiments) to other or real situation obtained from a part of population
called sample and sample is the proportion of population consider the representative of
whole population of interest. External validity is the degree/criteria to which research
conclusion provides an exact explanation and depiction of what normally happening in
the real world. A research has high external validity if it has high internal validity so
external validity dependent on internal validity and when External Validity is high, the
generalizations are truthful and can represent the external world as real world related to
the study. External Validity becomes very important when findings come from statistical
analysis of the statistical research because you sure that you have an accurate
representation of the population and when external validity is low, the integrity of your
research comes into uncertainty. You can reduce the threats of external validity by
making sure that sample should be select randomly, sample size should be large, and
strictly follow and use the random assignment then application will become more
reliable.