Ask Kazzy #2



One man. Two boys. Twelve kids.

Related Post Roulette

15 Responses

  1. Avatar Art Deco says:

    And the fact is that black people are not more violent than white people.

    You do not spend much time studying crime statistics. You may wish to argue about the reasons observed reality takes the form it does or argue that this particular observed reality does not manifest itself in domestic life or workplaces (you’d be out of luck arguing it does not in secondary schools). There is no point in arguing the observations are not what they are.Report

    • Avatar Kazzy in reply to Art Deco says:

      Please take your racist, trolling drivel elsewhere. This will be the last response I offer you here.Report

    • Avatar Art Deco in reply to Art Deco says:

      Racist trolling drivel????

      Kazzy, it does not matter whether I am a person of good character or bad, or whether I have it in for some ascribed group or not, or whether I subscribe to some approved narrative of social relations or I do not. The social statistics which describe the world in which we live are what they are. You can argue over the interpretation of those statistics or the implication of those statistics, but there is not point in pretending that what is is something other than what is and no point in making personal assaults on people who collect the data or know what the data says. Deal with social reality and stop being an ass.Report

      • Avatar kenB in reply to Art Deco says:

        This comes down to a question of exactly what is meant by the original (hypothetical) statement. If it’s meant as shorthand for “In the U.S., the percentage of black people who’ve committed a violent crime is higher than the equivalent percentage for white people”, then it’s a statement of fact. But as phrased, it sounds like the speaker is claiming that black people are inherently more violent, which is probably why Kazzy reacted the way he did.

        But in any case, this is a threadjack — the specific content of that hypo is not the subject of the post.Report

        • Avatar Art Deco in reply to kenB says:

          I will point out that my remark above contains this phrase:

          You may wish to argue about the reasons observed reality takes the form it does or argue that this particular observed reality does not manifest itself in,,

          which rather discredits that charitable construction of Kazzy’s remarks (unless Kazzy has a severe reading comprehension issue – a problem in a schoolteacher).

          As far as I can recall, the last time I saw in public print a remark as aggressively and obnoxiously naive as Kazzy’s was around about 1984. Makes me more suspicious of the ed schools than I already am.

          Where I grew up (the Genesee Valley), you have a complex of disagreeable inner city neighborhoods, agreeable inner city neighborhoods, tract suburbs, small towns, and countryside. The disagreeable inner city neighborhoods have a homicide rate of 35 per 100,000 in a typical year and comprehend about 10% of the population. The remainder has a mean homicide rate of 2.5 per 100,000, and I doubt you could find a neighborhood in the remainder that had it worse than the national metropolitan mean of 6.5 per 100,000. An aspect of that is wretched understaffing of the police and fragmentation of the police in a half-dozen different departments. Both of these can be feasibly addressed and the experience of New York City suggests a homicide rate of 14 per 100,000 is a realistic goal. But you cannot get to first base on quality-of-life improvements in slum neighborhoods if you insist on adhering to social fictions.Report

    • Avatar j r in reply to Art Deco says:

      You should probably learn the difference between an observation and a conclision. While you are at it, you may want to bolster your knowledge of statistical methods by learning the difference between a descriptive statistic and a predictive statistic and also learn what a unit of observation is.Report

      • Avatar Art Deco in reply to j r says:

        I know perfectly well what the difference is between a descriptive statistic and fruits of statistical analysis. It has been years since I made use of SAS and Statview, but I was skilled in that at one point, albeit dealing with economic data, not social data.

        I assume you can put together a massive bibliography of regression analyses of the influences on observed criminal behavior. That would certainly be helpful for the formulation of public policy. It certainly helps generally to sort out what you see when you hold other influences constant. That is a rather different exercise than lying to elementary school kids or in verbal assaults on fat middle aged men telling them they are repulsive characters for pointing out the world has the contours it does.Report

        • Avatar BlaiseP in reply to Art Deco says:

          Describe the difference, if you can. I’ll check your work.Report

          • Avatar Art Deco in reply to BlaiseP says:

            A descriptive statistic is a metric derived from survey research. In the analysis, you develop a model derived from theoretical discourse and test the model. There can be bivariate and multivariate analyses. You can test for between group differences with t-tests and ANOVA or you can undertake a regression analysis and attempt to locate the variables that have a statistically significant relationship with your outcome variable. There are phenomena which invalidate your analyses. There are also metrics which assess how much observed variation in a phenomenon can be absorbed by the model you have constructed.Report

            • Avatar BlaiseP in reply to Art Deco says:

              … and what sort of statistic would we call an analysis of the reasons observed reality takes the form it does?Report

              • Avatar Art Deco in reply to BlaiseP says:

                I used statistics as a tool. I am not a student of mathematical statistics and never was. I never really dealt with data that was not normally distributed.

                It will depend on what your discipline is and what you are attempting to determine. Multivariate regression analyses are modal in economic literature. I think psychologists tend to favor t-tests and ANOVA, but that’s outside anything I ever studied.Report

        • Avatar j r in reply to Art Deco says:

          Now incorporate the concept of a unit of observation and you should be able to see your errors.Report

        • Avatar George Turner in reply to Art Deco says:

          I’d point out that whites in Chicago only commit 5.3% of the city’s homicides (despite the city being a melting pot of every organized criminal element from Europe), and that among only those, Chicago has a lower number of homicides and a lower homicide rate than London England. But then, whites don’t commit half the crime in London, either. The British write about the matter quite extensively, and it’s probably not because they lack the proper statistical professorships.

          But if I pointed that out, I’d be tasked with deriving an entire branch of mathematics from first principles, because as Eric Holder pointed out, we’re all cowards, and we’d rather check under the refrigerator for dust bunnies than dare mention some uncomfortable data that might subject us to public scorn and sanction.Report

  2. Avatar zic says:

    Wonderful response, Kazzy. Because one of the most important things all authority figures do is model authority; and part of that is that folks often disagree, are frequently wrong, and we all need to develop something called judgment; not only to evaluate others, but to evaluate our own positions.Report