Psychology

Measuring Behavior: Scope of Statistics in Psychological Research

Imagine looking at an abstract painting hanging on the wall of your office. What do you see? What does the other person standing next to you see? What will all other people coming to your office see? The answer to this question is a multifaceted one. People react differently to the same physical stimulus all the time. And this is because our psychological world coexists with our physical world. So whatever we perceive at any instant is integration of the physical stimulus and our own emotional colours and past experiences intermixed together. So now the question is “why measure behavior?” Indeed the reasons are lot many. Behavior is anything and everything that an individual does in response to a stimulus. It is the overt expression of our covert mental processes. Measuring behavior is not only important to understand human nature but also facilitates predicting behavior in a specific situation. For instance, how many people are going to buy a new product just launched in the market? Or considering a road side accident, how many people will actually come ahead to help the victim? Or who among the managers of an organization can make a better leader?

How to measure Behavior? Quantitative or Qualitative measurement:
In the early decades of nineteenth century, studying human experience and interpretation was a popular way of understanding behavior. Earlier attempts to understand behavior and measure it includes the works of William James (1842-1910) and Franz Brentano (1838-1917) who studied human immediate experience and the self, involving systematic and objective techniques. Although, this approach to understand human behavior was later rejected with the emergence of Behaviorism and Cognitivism in the field of Psychology, qualitative approaches to human experience is still being considered to be a very effective method of accessing rich experiential information.

Quantification of human experience, on the other hand, emerged with the groundbreaking work of Gustav Fechner (1801 – 1887), who intended to discover the laws regulating the interconnection of external physical nature of stimulus with internal perceptual experiences. Popularly known as “Fechner’s Law”, the mathematical expression of this connection laid the bed stone for Psychophysics, a subfield of experimental psychology which still has enormous applications in understanding human sensory modalities and brain processes. Towards the beginning of twentieth century, the focus of attention got shifted from psychophysics to psychometric approaches to study behavior. In the earlier attempts to measure human behavior, psychologists, physicians and mathematicians joined hands to develop different kinds of behavioural measures. For instance, behavior during that time was largely understood in terms of intellectual processes. The first standard test of intelligence was developed by Charles Spearman, who proposed a Two-factor theory of Intelligence with a general and a specific factor of intelligence. This was followed by the development of other standardized tests of intelligence through factor analysis, primarily a statistical technique. Apart from Intelligence, attempts were also made to study human personality and accordingly tests based upon factor analytic approach were developed, for example, 16 PF test of personality developed by Raymond Cattell (Fig 1). Needless to mention, quantification of human behavior through statistical approach led to the development of this field to an unfathomable extent.

Fig 1. Cattell’s 16 Personality Factors; Image source: See References

Applications of Statistics in Psychyolog

Statistics is the discipline that is concerned with collection, organization, analysis and interpretation of data. Although theoretical to it’s’ core, statistics has huge applications in all possible fields of scientific research, and understanding human behavior is no exception. In the past few decades, statistics has largely shaped the paradigm of measuring human behavior. In today’s context, statistical techniques form the core of any quantitative psychological research which intends generalization and predictive implications. Be it any sub field of psychology, statistical methods finds its’ applications in every nooks and corner of psychological research, starting from sampling individuals, collecting data, analysis and interpretation of behavior. Apart from the general applications in quantitative analysis as required in experimental research as well as survey studies, application of statistical techniques in the field of psychology also pertains to psychometry involving behavioural measurement through psychological testing. As a matter of fact, some of the statistical techniques that are in existence today were primarily developed to fulfil the needs to construct and validate psychological tests. For instance, to estimate the temporal consistency of a test, the test-retest method of estimating reliability is widely used which is not primarily a mainstream statistical technique. Again, to estimate the construct validity of a test, Confirmatory Factor Analysis is a popular method in use which chiefly attempts to predict latent behavioural variables (e.g. stress) based on observed attributes (for example, loss of sleep, loss of appetite etc.). Such techniques not only enables researcher to obtain an objective measure of behavior, but also produces results that are highly reliable and can be generalized to the maximum.

Recent advances of Statistical Applications in Psychological Research

The last few decades have witnessed the epitome of statistical applications in the field of psychology. With the emergence of different menu driven computer applications like SPSS and STATA, doing statistical analysis of psychological data has become much more sophisticated and precise. Most recent advances in psychological data analysis make use of R-Programming, Python and MATLAB which are rather data-driven applications that are largely in use for psychological data analysis and designing psychological experiments. For instance, the PSYCH package within R-Programming software contains almost all applications that are customized for psychological data analysis. Similarly, PSYCH TOOLBOX application in MATLAB enables cognitive psychologists to design cognitive experiments.

Fig 2: Structural Equation Modeling; Image adapted from  Ye et al., 2018

Statistical applications in Psychological test development also witnessed a drastic change. With the advancement of modern Item Response Theory for test development, test items are now being validated based on logistic models. Prediction based models that are investigated through experimental or survey based methods are now being done through Structural Equation Modeling (Fig 2), which combines mathematical models, computer algorithms and statistical techniques. This includes Measurement Models chiefly required for Confirmatory Factor Analysis, Path Analysis and Latent Growth Models. Newly developed statistical techniques are facilitating researchers to study the change in a psychological attribute over a period of time through Trajectory Analysis. Most of these techniques have in their core the concepts of statistical regression, but applied to different settings.

In the past few years, psychologists and researcher are witnessing a major boom of rich psychological data from people on a daily basis, thanks to the social networking sites and countless online surveys used by different organizations. With the accumulation of this rich source of data, psychological data analysis has reached a new level of sophistication. As a joint effort of Data Scientists and Psychologists, Psychoinformatics as a new discipline of psychology has emerged. This is an interdisciplinary field where Psychology meets Statistics and Machine Intelligence and where data mining techniques are used to predict human behavior long before people themselves can predict their own behavior (Fig 3). Albeit in a nascent phase, such a discipline bears evidence to the fact that the definition of measuring behavior has changed drastically with the increasing conglomeration of Psychological principles with Statistical techniques.

Fig 3: Psychoinformatics; Image Adapted from Markowetze et al., 2014

Conclusive Remarks:
Psychology, while primarily considered to be a Social Science, have come a long way in the past century and has emerged as a fore runner in the field of applied interdisciplinary scientific research. Applications of statistical methods in psychological research have enabled the discipline to overcome the limitations of subjectivity in measurement to a large extent and produce much more objective and statistically sound inferences about human behavior. In a nut shell, Statistics as a discipline, has become one of the cornerstones of Psychology and has equipped it with all the gears to operate most effectively as a field of scientific research.

References:
• Markowetz, A., Blaszkiewicz, K., Montag, C., Switala, C., & Schlaepfer, T.E. (2014). Psycho-informatics: Big Data shaping modern psychometrics. Medical hypotheses, 82 4, 405-11.
• Ye, Jiao & Chen, Jun & Bai, Hua & Yue, Yifan. (2018). Analyzing Transfer Commuting Attitudes Using a Market Segmentation Approach. Sustainability. 10. 2194. 10.3390/su10072194.
• Business Concept Team (2019). 16 PF Personality Test Definition, Importance, Advantages, Disadvantages, Example & Overview. https://www.mbaskool.com/business-concepts/human-resources-hr-terms/17829-16-pf-personality-test.html

Visited 684 times, 1 Visit today