Archives
Contribute
| Maharishi Effect Research Assessment Of Causality |
Martha B. Colby
07/24/2008
A Methodological Review - Part I
When confronted with unorthodox research findings such as those on the Maharishi Effect, there is a tendency to dismiss them as the result of faulty research. Some might say they go to show that statistics can be made to prove anything. In fact, these studies were rigorously conducted. Their publication in leading journals, such as Journal of Conflict Resolution, Social Indicators Research, and Journal of Mind and Behavior, indicates that they have met the highest standards for social science research. This is particularly true because paradigm-breaking research is always subjected to closer methodological scrutiny than standard research. The review that follows addresses in layman’s terms the basic issues that arise in trying to prove causality in sociological research, and discusses how the research on the Maharishi Effect has addressed these issues
Controlling for Alternative Hypotheses
How do we know that the TM and TM-Sidhi program is responsible for observed reductions in crime rate? How do we know that the changes in society are not due to some other influence?
In many cases, a number of other variables that could potentially influence the social indicators under study were known from prior research. For example, studies have shown that crime rate is influenced by such factors as the proportion of young adult males in the population, percentage of families with incomes below poverty level, and median years education. In such cases where other relevant variables were known, research on the Maharishi Effect has controlled for these variables by taking their influence into account.
Taken as a whole, the 22 studies on the influence of the Maharishi Effect on crime have statistically controlled for the influence of all variables known to influence crime before assessing the effects of the TM and TM-Sidhi program. These studies have specifically controlled for population, college population, population density, geographic region, percent of persons aged 15–29, ratio of police to population, police coverage, neighborhood watch programs, median years of education, family income, per capita income, percentage unemployed, percentage of families with incomes below poverty level, percent in same residence after 5 years, and seasonal effects.
Time Series Analysis
Many social indicators are influenced by seasonal cycles. Crime rate, for example, decreases in the cold winter months and increases in the hot summer months. Weekends and major holidays also influence many indices of human behavior. Therefore, in studies of the Maharishi Effect, all such seasonal cycles are taken into account before assessing the contribution of the TM and TM-Sidhi program. In addition, there may be upward or downward endogenous trends. Whereas trends represent a systematic change in the level of the process (an overall increase or decline), some data sets may simply drift randomly around some mean level, such as the stock market. Studies on the Maharishi Effect must demonstrate that improvement did not occur at the time of the experiment because of cycles, trends, or drifts.
The methodology for taking cycles, trends, and drifts into account is called time series analysis. A time series is a sequence of measurements over equal periods of time, such as days, months, or years. Time series analysis identifies the time-dependent regularities in the data and then calculates a mathematical model that best describes them. Using this mathematical description of the regularities in the data, time series analysis then statistically removes their influence before assessing the possible effects of other variables. The variable that is believed to affect the process is called the exogenous or independent variable.
In studies on the Maharishi Effect, the TM and TM-Sidhi program is the independent variable. These studies have examined the influence of change in number of TM and TM-Sidhi participants in the population on various social indicators. Twenty-eight studies on the Maharishi Effect have used time series methodology to show that quality of life improved in a way that could not have been predicted by time-dependent regularities in the data. Furthermore, by removing these regularities from the data, this methodology, in principle, controls for any unknown variable systematically influencing the series.
Also, time series analysis allows the researcher to control for other exogenous variables before estimating the impact of the Maharishi Effect. For example, studies of the influence of TM and TM-Sidhi participants on inflation and unemployment in the U.S. used this method to control for monetary growth, change in crude materials prices, industrial productivity, and a measure of the money supply. Time series analysis also provides a precise estimation of the size of the statistical effect.
An important issue in statistical modeling is what constitutes the best model of the data. Some experts argue that the best model is the simplest model and that one should only include components that can be easily interpreted, such as weekly or yearly cycles, while other researchers prefer strictly mathematical criteria for determining the best model. Research on the Maharishi Effect has used both criteria to demonstrate that the effect is robust no matter how one defines the “best†model.
To be continued.. The second part will deal with Causality and
Empirical Confirmation of Mechanisms
|
You may also access this article through our web-site http://www.lokvani.com/
