--- In [email protected], "Bill (William)Simmons" <[EMAIL PROTECTED]> wrote: > > I cut and pasted this claim from a MUM web page claiming that Where > as little as 1% of population is practising TM "the trend of rising > crime rates is REVERSED. > > > Cities in which one per cent of the population were instructed in the > Transcendental Meditation Programme showed decreased crime rate the > following year, in contrast to matched control cities. Reference: > Journal of Crime and Justice 4: 2545, 1981. > Cities in which one per cent of the population were instructed in > the Transcendental Meditation Programme showed a trend of decreased > crime rate in subsequent years, in contrast to matched control > cities. Reference: Journal of Crime and Justice 4: 2545, 1981. > > > 318. DILLBECK, M.C.; LANDRITH III, G.; and ORME-JOHNSON, D.W. The > Transcendental Meditation Program and crime rate change in a sample > of forty-eight cities. Findings previously published in Journal of > Crime and Justice 4: 25-45, 1981.
"Matched control cities" a methodology often cited in ME studies, including above, and DOJ's analysis of FF crime, have always seemed to be a methodologically difficult issue. "Matched pairs" is an often used resarch method for "creating" a control group from observed data (that is there was not a pre-selection of a control group prior to the "intervention" / dose. Though it is error-prone, such errors and biaseas can be reduced using rigourous statistical methods to match multiple criteria that have been shown to effect the "response" -- the dependent variable. For example, in doing a google search on "matched pair cities" I found no other studies that used matched pairs for comparision of cities. Understandably so given the huge difficulty in rigorously match for the things I cited in my my earlier post -- flaws in DOJ's FF crime anaylsis: "demographic cohorts, temperature, seasonal effects, education levels, % with active religious affiliations, income levels and regional economic trends would be useful if not necessary control / matching factors for a credible analysis." One study that had "matched pairs" and "cities" showed up, but was about matching individuals within a city. Their rigor is of notable contrast to DOJ's and the above study. "Case patients and controls were similar in terms of age, gender, insurance status, median household income, and proportion with an underlying premorbid neurodevelopmental disease (Table 1 [triangle]). Case patients were more likely to be Asian or of Hispanic ethnicity. The odds of Asian children having been involved as a pedestrian in an accident were 5.8 times as high as those for White children (P = .018), and the odds of Latino children having been involved were 4.3 times as high (P = .038)." http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1448312 And they end the paragraph stateing "Admitting diagnoses of controls are available on request from the authors." Hardly an offer made by ME researchers that i have ever seen. This is critical in that the method by which, and the factors used to construct the pairs is critical. Many sets of "unmatched pairs" across ten or more critical variables -- but matching in one, could be constructed. None would be methodologicaly valid. Many small cities, literally 100,000s could be matched for per capital violent crime rates if that, and rough population size "small town" were used. What criteria was used for developing the presumably much smaller set of matched pairs cities. The huge onus is on the resarchers to demonstrate that some random process was used, and not cherry-picking to secure the results "desired/expected by sponsors". The fact that DOJ did not even included a parallel analysis of matched parirs for property crime is highly suspect. Did it not show a "useful" reduction of crime? Another study that came up was forweather -- again not matching cities. http://people.hofstra.edu/geotrans/eng/ch7en/appl7en/ch7a1en.html "Although road safety researchers focus primarily on driver behavior, vehicular defects and road design, there is general agreement that environmental factors, such as weather and darkness, also affect accident risk. Research on weather-related hazards, especially precipitation, has made extensive use of matched sampling. This application of matched sampling first requires weather data that can be used to identify precipitation events. Each event is then matched with a suitable control period (i.e. a period of 'nice' weather). For example, a Monday afternoon rain shower lasting from 1 p.m. until 4 p.m. would be paired with the same three hour period on a Monday afternoon just one or two weeks prior to or following the event. The absence of any kind of adverse weather during the control period is an essential feature of this method. Events without a suitable control are deleted from the sample. It states, the main reason for choosing this quasi-experimental design is related to the degree of control that is implicitly incorporated in the study, i.e. the fact that many variables that have nothing to do with weather, but which do affect accident frequency, such as traffic volume, road conditions and the incidence of impaired driving, are controlled rather effectively. As a result, it seems reasonable to suggest that any deviations from a risk ratio of 1.0 are largely attributable to weather." A third study that appeared discussed its method for creating a post hoc control group for a crime study. it used control variables including: Demographic variables in Part 1 include gender, educational background, and occupation. Other variables include educational ambitions, time spent after school on a variety of activities, reading interests, opinions about labor, extracurricular activities, type of ideal job, relationships with teachers, parents' occupations, parents' expectations of children, family discussions, family's material well-being, relationships with friends, dating relationships, descriptions of self, number of delinquent acts, and reasons for delinquency. Demographic variables in Part 2 include gender, birthdate, birthplace, religion, education level, marital status, military status, and occupation. Other variables in this file are dates when subjects moved into and out of the address in their residential file, number of public safety violations, types and dates of public safety violations, number of criminal offenses, types and dates of criminal offenses, and penalties for criminal offenses. http://webapp.icpsr.umich.edu/cocoon/NACJD-STUDY/03751.xml The picture is clear, for valid matched pair studies, a large number of control variables, those that may effect the dependent variable -- the "resposnse" variable, such as crime in the DOJ FF crime analysis -- need to be examined and tested for their effects on the response variable. And those that do need to be incorporated into the matching criteria or models. The DOJ FF crime analysis, and other "matched city" studies sponsored by MUM or done by MUM faculty appear to be quite methologically deficient in using adquate controls for matching cities -- or deficient in stating their methods. (The latter being improbable, given the importance of such methods to the findings.) "Matched pair cities" should raise a huge red flag for any ME study citing it. One reaason DOJ probably didn't attempt, or report, rigorous matching of cities is that it would be difficult if not virtually impossible to match FF/Parsons to any other small town. Matching demographic cohorts is for paired analysis would be critical in crime studies since a majority of crime has been shown to be committed by male 18-26 year olds. Approximately 35% of FF/Parsons population fit this profile. And apparently, on some nights, an influx of simililar aged/gender "flockers" "invaded' FF - perhaps raising the ratio to near 50%. How many other small towns have such a demographic characteristic? Maybe some all male college towns. How many of those towns had students not from the community, or living at home (the profile of FF/Parsons students -- away from home and no roots in the community). Using this second variable for matching is critical, and probably reduces the number of viable towns to a handful. That these two vaiables were apparently not used for matching diminishes the validity of the study substantially - -to perhaps near zero. it would be like comparing crime to a wild, preominantly male, college town to a retirement community. Its simply irrelevant. The issue could be addressed by developing a series of multi-variate crime models, one for each major clustering of crime types, at a minimum separate models of violent and property crimes, for 30-100 small towns, including variables such as those mentioned above. And add a variable for a constructed "ME Units" index.* This variable would be near zero for most towns (low but not zero for some college towns with TM programs), zero for FF/parsons years, and increase ovr time as FF grew in meditators and YFs. Such a study design would essentially test the effect of ME units on different types of crime, while controling for factors such as demographic cohorts, temperature, seasonal effects, education levels, % with active religious affiliations, income levels and regional economic trends, etc. This would be a far more rigorous and plausible study design than using matched control cities. And all of the problems of matching towns, and subsequent methodological issues raised in the matching, would disappear. And larger towns could be included, thus increasing the variation in the sample, and adding more TM towns, by adding a population size variable to the model (probably a dummy variable for 3-4 discrete population ranges). In this way, ME effects could be "captured' in other TM towns other than just FF. This would be a critical step in demonstrating causality of ME, not justssome random, spurious correlation effects. Particularly given that the ME units series would not be highly correleated. If the TMO, or any researchers, were serious about the ME effect, they would sponsor, or seek funding for studies such as described above. Several studies do appear to examine ME effects across multiple cities. I will lok into those. Hopefully they are more rigorous and met hodologically sound than DOJ's or othr "matched control city" studies. -- *ME units is an index based on the presumption that TM 20/2x has less effect than rounding, and YF. Several indexes could be cosntructed and tested in the general model, that is for each MEer, 1=TM 20/2x, 4 = rounding, 6 = YF 2x, 15=YF 4-6x. To subscribe, send a message to: [EMAIL PROTECTED] Or go to: http://groups.yahoo.com/group/FairfieldLife/ and click 'Join This Group!' Yahoo! Groups Links <*> To visit your group on the web, go to: http://groups.yahoo.com/group/FairfieldLife/ <*> To unsubscribe from this group, send an email to: [EMAIL PROTECTED] <*> Your use of Yahoo! Groups is subject to: http://docs.yahoo.com/info/terms/
