PSYCHOINFORMATICS IN THE InSPA CONFERENCE, Pondicherry, October, 2016

ORIENTATION WORKSHOP ON PSYCHOINFORMATICS: A NEW PARADIGM OF RESEARCH IN SCHOOL PSYCHOLOGY

LECTURE NOTES OF DEBDULAL DUTTA ROY

PSYCHOLOGY RESEARCH UNIT

INDIAN STATISTICAL INSTITUTE

KOLKATA

School psychology is Professional Psychology for School related crisis intervention. It deals with children, adolescents, family and neighbourhood. 

SCIENTIST - PRACTITIONER FRAMEWORK

Crisis intervention is an integral part of school psychology. School administrators view school psychologists as the school ‘s crisis intervention experts.Crisis intervention affects student’s ability to learn and function effectively. Following recommendation of National Association of School Psychologists (NASP,2007) and the American Psychological Association (APA, 2007), school psychologists adhere to the scientist-practitioner framework and make decisions based on empirical research. Their researches are not necessarily directed toward development of theory rather to systematic inquiry to find effective solutions to real life crisis.

ACTION RESEARCH 

Since school psychologist acts as the change agent or agent for crisis intervention, it follows action research. There are two approaches of action research - hypothesis driven and data driven. Action research does not aim at development of educational theory rather aims at intervention of school related crisis. 

Kurt Lewin (1890–1947) , Organizational Psychologist, has conceptualized action research. Action research is the term which describes the integration of action (implementing a plan) with research (developing an understanding of the effectiveness of this implementation). 

Teachers use action research because: 

1. it deals with their own problems, not someone else’s 
2. it can start now—or whenever they are ready—providing immediate results 
3. action research provides them with opportunities to better understand, and therefore improve, their educational practices
4. as a process, action research promotes the building of stronger relationships among staff 
5. importantly, action research provides educators with alternative ways of viewing and approaching educational questions providing a new way of examining their own practices. Adapted from Mertler, C.A. & Charles, C.M., (2008) Introduction to education research, 6th Edition, Allyn & Bacon, Boston, Mass, page 308.

DIFFERENCE BETWEEN FORMAL AND ACTION RESEARCH

1. Time:   Formal research is extensive and time consuming. Action research consumes less time;
2. Generalization: Findings of formal research  is generalisable to a wider audience. But in action research results is applicable for improving practice in a local situation.
3. Method of problem identification: Method of identifying problems in formal research is review of previous research findings. But in action research, attention is paid to  currently faced problems (bullying, ragging, weapon use) or improvements needed in a set of classrooms or a school.
4. Literature review: In formal research includes extensive enquiry into all research previously conducted on this topic using primary sources. But in Action research, review includes some primary sources but also use of secondary sources plus what practitioners are doing in other schools.
5. Sampling: In formal research, sampling includes random or representative preferably with large populations. But in action research, sampling includes students and/or members of the school community.
6. Research design: It involves rigorous controls over long periods in formal research. But in action research, design is more flexible, quick time frame.
7. Research reasoning: Research approach includes  deductive reasoning – theory to hypothesis to data to confirmation in formal research. But in action research,  inductive reasoning – observations, patterns, interpretations, recommendations.
8. Data analysis: In formal research, analysis of data includes tests leading to statistical significance. But in action research it includes grouping of raw data using descriptive statistics.
9. Research significance: In formal research, findings have  theoretical significance. But in action research, results have  practical significance. 

Si

S

Ref: https://www.det.nsw.edu.au/proflearn/docs/pdf/actreguide.pdf

When working through your action research remember that:

•  it is cyclical and progress is made in small chunks
•  it is based heavily on critical reflection
• you can use a wide range of methods for collecting data but it may be advisable tolimit these to a manageable number
• participants should have meaningful roles in the collection and presentation of data
• because of the flexibility of the process and the constant reflection, not every cycle will be complete. There may be times when it is advisable to stop mid stream and start a new cycle.

Both formal and action research follow hypothesis driven approach.

Hypothesis: Hypothesis is a tentative assumption drawn from knowledge and theory to serve in the investigation of other unknown facts and theory. Hypothesis enables obtaining requisite data, application of specific statistics for the analysis of the result, drawing conclusions and generalization and thus serves in the advancement of knowledge and research. There are few limitations of hypothesis driven research.

1.Hypothesis formulation depends upon prior theories and observation. Both of them are situation and time specific. Therefore, sometimes collected data do not satisfy the hypothesis causing difficulty in interpreting results. 
2. Hypothesis governs scaling of variables as it states whether variables will be metric or non-metric.  
2. Hypothesis controls resarcher's freedom of data collection. The statement suggests whether you will follow experimental or survey method.. 
3. Hypothesis is directive. It gives specific direction of data analysis.It tells whether researcher will use differences in central tendency or use  correlation statistics. 


PSYCHOINFORMATICS

Psychoinformatics is paradigm shift from classical hypothesis driven approach to data driven research approach. With rapid, randomized digitalized and non digitalized information explosion, the problems of psychology are moving from bounded psychology arena to unbounded psychology. Knowing psychology through psychological testing limits our knowledge to pre-assumed psychological traits. This causes serious problem to gauge all the determinants of individual differences in behaviour.Psycho-informatics  is  a new science through which we can mine data in any form and can build model to understand psychological traits. It uses computer databases to store, retrieve and assist in understanding psychological information. Below are two studies of Psycho-informatics.

  Psychoinformatics  is  a technique through which we can mine data in any form and can develop pattern based on  relations among data. The pattern finally reflects specific psychological phenomena.  Sometimes, researcher finds unique phenomenon of psychology as this approach follows here and now rather earlier notions. It uses computer databases to store, retrieve and assist in understanding psychological phenomenon. Psychoinformatics  entails the creation and advancement of databases, algorithms, computational and statistical techniques and theory to solve formal and practical problems arising from the management and analysis of  psychological  data. There are five stages of Psychoinformatics:

Data Warehouse:  It is the reservoir of school related data It includes both digitized and non-digitized data. Digitized data is the informal electronic chat among students, among teachers and students over social media -facebook, google group, Whatsapp etc. Non-digitized data are the text data like presence and absence, complaints of students, answer sheets of students, minutes of the meeting etc. 

Data retrieving: Based on the research questions, specific types of data are to be collected from the data warehouse. In this case, it is considered that data is represented in a structured way, and there is no ambiguity in data. In order to retrieve the desired data the user present a set of criteria by a query.  Data retrieval poses special challenges for archival storage systems, which must provide the capacity and reliability to retain data for many years, protect that data against unauthorized changes and quickly locate fragments of data from the archive on demand. 












Data warehouse, data retrieving, data mining, pattern recognition  and discovery of knowledge are the five stages of psychoinformatics. 







.

Data warehouse :    It is the reservoir of data. Answer sheets of selection service board examinations, case history records of psychiatric patients, absenteeism records of armed personnel, foot falls in the war zones, records of different cultural programs in army tents are the data warehouse. These data are usually in non-digitalized form. The digitalized warehouses are social network groups in virtual world. Yahoo groups, Orkut, Facebook, message boards, blogs and Youtube are the digitalized warehouses. In the data warehouse, data are stored in multiple forms like text, numeric, sound, and pictures. Basic assumption of psychoinformatics is that data warehouse is formed by random generated data. There are some cross data base search engines like Entrez in the internet. The address of Entrez is ‘http://www.ncbi.nlm.nih.gov/sites/gquery’. It is maintained by NCBI or National Center for Biotechnology Information. As it is cross data base warehouse, one can get number of articles published in different data warehouse ranged from Pubmed (Biomedical literature citations and abstracts), to genome project (genome project information).  So, one can do literature survey by analysing number of citations through the above cross and non cross data base search engines. Ganguly and Dutta Roy (2010) did content analysis of web data to explore categorization of  articles of Entrepreneurial psychology published in the Science direct.com.

Data retrieving: Using well defined criteria, specific data will be retrieved from the data warehouse. When the data are in digitalized form, computer algorithm should be used to retrieve the data. In web analysis, search engines require query from the user. User provides query in the selected space. The search engines using computer algorithm extract data matched with query. In web content analysis of Entrepreneurial psychology, 7 keywords as cognition and entrepreneurship, attitudes of entrepreneurs, characteristics of corporate entrepreneurs, application of psychology in entrepreneurship, personality characteristics of corporate  entrepreneurs, behavior of entrepreneurs were used. And 60 articles were extracted. These are finally categorized into 5 categories as general personality, thinking, leadership, demography and others. Most of the articles (44%) focused on leadership traits.    Table 1 shows number of citations of different psychiatric disorders available in Pubmed.

Table 1: Number of citations on psychiatric disorders available in Pubmed

Names	Number of citations	Proportion
Anxiety disorder	2588781	0.90
Depression	235173	0.08
Conversion reaction	419	0.00
Obsessive compulsive disorder	3247	0.00
Hypochondriasis	962	0.00
Somatoform disorder	2456	0.00
Schizophrenia	11128	0.00
Manic depressive	11382	0.00
Paranoid disorder	6593	0.00
Autism	14624	0.01

Sometimes, data warehouse includes large database and it is important to select only those data used for data mining. These digitalized data can be retrieved by using computer algorithm. Those who are not aware of algorithms can use available algorithms of different application software. Ms-Excel has several functions to retrieve text data or string data as LEFT(text, num_chars), MID(text,start_num,num_chars), RIGHT(text, num_chars),  SEARCH(find_text,within_text,start_num) , LEN(text).

Data mining:  Data mining is an analytic process designed to explore data  in search of consistent patterns and or systematic relationships between variables, and then to validate the findings by applying the detected patterns to new subsets of data (Fayyad, Girnistein and Wierse, 2002).  Classification (involves finding rules that partition a given data set into disjoint classes)  and clustering (conceptual groups in data on the basis of similarity), prediction are common models (Andrusiewicz and Orlowska, 1997).

Pattern recognition:    Psychological data are complex as they possess multiple attributes. In a group of randomized psychological data sets, attributes of one or more variables are related to one or more variables of others and finally they form certain pattern. Later on that pattern can be recognized using certain model.  Model is  prototype of complete and consistent set of verbal arguments, mathematical equations or computational rules which are thought to correspond to some observable entity in the World. For example, one uses cluster analysis for data mining and finally notes dendogram.  

Knowledge discovery: It is the ultimate solution of psychoinformatics. Here, new knowledge about relationship among the variables or the process of analysis is discovered. For example, dendogram provides knowledge about classification of variables and researcher assigns new name to each cluster.

Case-study: Psychoinformatics based on digitalized data warehouse

With rapid change in web technology, training method gradually shifts from manual to web based computer adaptive training. Here, each trainee is not exposed to same stimuli. Stimulus exposure depends upon individual differences in proficiency level accounted by the computer. Therefore, usual validity estimating statistical tools (ANOVA with repeated measurement) are not applicable. This method can be applied into armed forces in order to speed up the training process. Dutta Roy (2008) conducted one study to examine criterion related validity of one web based computer adaptive training program. This is discussed below to understand principles of psychoinformatics.

Data warehouse:     Responses of individuals (N=25) to one Internet-based training program (Fast ForWord®) were stored in progress tracker.  Progress tracker is an on-line service to download the results of each trainee. It is the warehouse of digitalized data. Training program helps individuals to build oral language comprehension and other critical skills necessary for learning to read or becoming a better reader. The product is based on the adaptive training techniques such as frequency, reward, intensity and motivation allow for more rapid learning. It includes seven modules. All modules were classified into two as sound (Circus sequence, old McDonald flying farm) and word (phonic match, phonic words, language comprehension builder and block commander) exercises. Here, progress tracker is a data warehouse.

Data retrieving:  Data of each trainee  were downloaded from the progress tracker after the training session and kept them in one Microsoft Excel sheet.  Percentages of success of each  trainee  across each training module were extracted using formula of Ms-Excel.

Data mining:    As each response is random and each participant is exposed to different types of stimuli, available statistical tools like ANOVA or MANOVA can not be used. Therefore, for data visualization, Box whisker plot analysis was made to compare the results across treatments or trials. The boxplot was invented in 1977 by American statistician John Tuckey. A box plot or box and whiskers plot is a graphical way of representing the salient features of a distribution (Figure 1). It can be used with either Gaussian or non-Gaussian distributions. The box plot shows a rectangle stretching from the first to the third quartile of the distribution, these quartiles, the edges of the box, are called "hinges". The box displays in a pictorial fashion the variability in the data. A line inside the box shows the approximate position of the median. From the median, one can determine the central tendency or location. From the length of box, one can determine spread, or variability of observation. If the median is not in the middle of the box the distribution is skewed. The further the median is from the middle, the more skewed is the distribution. If the median is closer to the bottom of the box than to the top, the data are positively skewed. If the median is closer to the top of the box than to the bottom, the distribution is negatively skewed. Cases with values that are between 1.5 and 3 box lengths from the upper or lower edge of the box are called outliers and are designated with circle.

Several parameters of box plots (Outliers, box size, location of median in the box, location of  upper and lower whisker, location of hinges, and fluctuation of box size after achieving the target) were identified as useful to estimate the validity of training modules and to understand pattern of cognitive difficulties during training.

Module A

Module  B

Figure 2: Box-whisker plot of two training modules.

In Figure 2, Module A shows systematic increase in median and decrease in box sizes over training periods suggesting high criterion related validity in comparison with Module B. Fluctuation of box size in module A indicates cognitive difficulties of respondents after achieving the criterion. On the other hand no systematic learning curve is noticed in case of Module B. This suggests poor criterion related validity.

Box plot results suggest possible ability contamination in changing performance level. Therefore, in order to control contamination of ability on validity assessment, high and low ability groups were identified based on initial training performance. Learning graph profiles of both groups were compared finally. Results identified very poor validity in two training modules out of seven. By scrutinizing the contents of these two modules, it is  noted that both are meant for developing proficiency in western pronunciation and accent of English words. Though the participants of the study were English speaking  but they could not follow the English accent used in the module as revealed in the results.

Discovery of knowledge:  When the responses are randomised and data were collected from adaptive testing procedures, Box whisker plot is useful analytical technique. Second, after achieving the target efficiency, concentration difficulty occurs.

References:

1. Tutorial: Psycho-informatics: model for measuring randomized behaviour : Measuring behaviour, 2012, Utrecht, Netherlands.http://www.measuringbehavior.org/mb2012/tutorial-psycho-informatics-model-measuring-randomized-behaviour

2. Dutta Roy,D.(2010). Psychoinformatics: Innovation in mining randomized data. PSYBER NEWS: International Psychology Research Publication,1,1,23-31.
3. Dutta Roy,D.(2006). Development of picture drawing test to assess consciousness layers of tribal children of Tripura, Journal of the Indian Academy of Applied Psychology,32, 1, 20-25
4. Dutta Roy, D. and Paul, M. (2002). Reading motivation of children in grades III and IV,Indian Educational Review, 38,1,43-51.

are five basic principles ofpsychoinformatics. This tutorial session will focus on following topics:

1. Research paradigms of school psychology
2. Paradigm shift from hypothesis to data driven model
3. Five principles of Psychoinformatics and its application in school psychology.

4.  

Psychoinformatics is free to group