Social Causation & Mental Illness

View Segments Segment :

  • Citations
  • Add to My List
  • Embed
  • Link
  • Help
  • Citations
  • Add to My List
  • Embed
  • Link
  • Help
Successfully saved clip
Find all your clips in My Lists
Failed to save clip
  • Transcript
  • Transcript

    Auto-Scroll: ONOFF 
    • 00:01

      I'm Dr. Brent Shea, Professor of Sociology at Sweet BriarCollege and author of the entry titledSocial Causation in Cultural Sociology of MentalIllness, edited by Andrew Skull, published by Sage in 2014.Social causation goes beyond description of social facts,such as those available in population census

    • 00:23

      data, proposing to offer explanations instead.These explanations typically are based on statistical analysisof representative samples of populations,a cost-effective and practical wayto study large, diverse populations.Social causation, like any form of causation,

    • 00:44

      must meet three criteria-- co-variation, temporalpriority, and controls for alternative explanations.In the simplest case involving only two variables,the two variables must co-vary sothat a change in one variable is accompaniedby a change in the other variable.This change can be positive or negative.

    • 01:06

      A positive relationship between two variablesmeans that as one variable increases, the other does, too.A negative relationship means that as one variable increases,the other decreases.Whether the variable is positive or negative,its size or strength typically is measured by a correlation

    • 01:28

      coefficient, with larger coefficients indicatingstronger relationships.Regardless of the size of the relationship between twovariables, the likelihood that a relationship of that sizecould occur by chance alone also must be assessed.This assessment is done by means of a testof statistical significance with higher values indicating

    • 01:51

      a greater likelihood that a relationship between the twovariables of a given strength could occur by chance alone.Most tests of statistical significanceare based on the assumption of a linear relationshipbetween variables-- that is, as one increases,the other increases or decreases.

    • 02:11

      The standard threshold for statistical significance is 05,which indicates a 5% probability that a relationship between anytwo variables of a given size could have occurred by chancealone.When a relationship between two variables resultsin a statistical significance level higher than 05,

    • 02:34

      it is regarded as too likely to have been produced by chancealone to be regarded as a real relationship between the twovariables.Generally, the more strongly the variables are related,the greater the probability is that the relationship willbe statistically significant.In addition to co-variation, a second criterion of causation

    • 02:56

      is the temporal priority of the two variables.Which variable came first?Which is the cause or the independent variable,and which is the outcome or effect, the dependent variable?In order to investigate that question empirically,a hypothesized relationship is formulated

    • 03:17

      specifying which variable is temporally prior to the other.In addition to co-variation and temporal priority,a third criterion for causation is the demonstrated attemptto control for alternative explanationsfor a statistically significant relationship.

    • 03:37

      This seeks to discover whether the hypothesized causeof a particular effect or event is indeed the only cause.This is accomplished by identifying and measuringvariables that are implied by alternative explanationsfor the same event.Alternative explanations potentiallychallenge any relationship that posits

    • 03:59

      a single cause for any outcome.Controlling for multiple potential causal influenceson a variable is necessary, because itis possible to conclude incorrectly that it may havebeen due to only one factor.In the simplest case, potential multiple causes

    • 04:22

      of the same outcome contribute to explaining that outcome.When combined, that is, these variablesaccount for some or all of the reasonswhy a particular outcome occurs.A more complex causal sequence involves one or more variablesthat intervene between the causal variable and the outcomevariable.

    • 04:43

      These intervening variables make possible an interpretationof the original relationship between two variables.They help explain why these two variables, whichco-vary at a statistically significant level,are related to each other.The identification of one or more intervening variablesindicates one or more reasons why the original relationship

    • 05:06

      exists in the first place.Even in the absence of the original causal variable,the intervening variable or variablesare interpreted as causes of the original outcome variable.A much more complex causal orderingthan that provided by either contributing or interveningvariables is provided by antecedent variables.

    • 05:28

      An antecedent variable explains whythe original statistically significant co-variationoccurred.Indeed, it makes the causal interpretationof the original two variables spurious or suspect.The reason they are related is notbecause one causes the other, but because both are causedby a third prior variable.

    • 05:50

      This is confirmed if the original relationshipis weakened by the introduction of an antecedent variable,especially if this weakening rendersthe original relationship non-significant.Where the variables are contributing, intervening,or antecedent, the possibility of a statistical interactionis present.By that, it is meant that the effect

    • 06:12

      of one variable on another is dependent on the levelof a third variable.For example, the effect of alcohol on consciousnesscan depend on the level of barbiturateswhich interact with alcohol in a way thatcan result in sleep or coma.Causation is established only when

    • 06:33

      all three criteria-- co-variation, temporalpriority, and controlling for alternative explanations--are met.Guided by these tenants of scientific research,experimental scientists are able to establish causation.Most social scientists, however, do not conduct experiments.The reason for not conducting experimentsinclude ethical as well as practical and financial

    • 06:56

      concerns.It also is not necessary to experimentally manipulatesocial situations.Variables in causes of outcomes of interestto social scientists occur in society anyway.Inequalities in socioeconomic status as well as inequalitiesin the prevalence of mental illness, for example,already exist.

    • 07:17

      These are among the inequalities thatengage social researchers who develop their systematic bodyof knowledge by identifying, measuring, and monitoringsocially-patterned differences-- that is, differencesin outcomes that are related to the templateof social structures and processesthat social researchers have developed theoretically,

    • 07:38

      then analyzed empirically.This analysis is accomplished by manipulating data statisticallyrather than by manipulating peoplewhose cultural, social, and economic conditionsalready differ.Data sets that are used to investigate causal patternsare primarily either survey or aggregate data.

    • 07:60

      Survey data makes it possible to study independent variablescharacterizing particular respondents in relationto dependent, attitudinal, or behavioral variablesof that individual.Aggregate data, collected to characterize aerial unitslike cities, states, or nations, sometimesare misinterpreted as survey data.

    • 08:22

      This is known as the ecological fallacy, which is the attemptto impute causal linkages between data collectedat the aggregate level to data at the individual level.For example, whether high unemployment rateshave a causal influence on smokingrequires more than a high correlation between the twoat the aggregate level.

    • 08:43

      It requires individual level survey datathat links high rates of smoking to unemployed persons,including data on whether smokingbegan before or after being unemployed.Social causation may be understoodas essentially probabilistic.Once an outcome such as a psychiatric diagnosis based

    • 09:07

      on sociodemographic characteristics is identified,this almost never implies that all individualswith those characteristics have that disorder,or even that all individuals with that disorderhave those characteristics.It may mean that most do, or it maymean that it is an outcome relatedto a particular disorder with a given probability.

    • 09:29

      This lack of certainty that particular risk factors arepredictive of disorders, or protective factors protectionfrom disorders, may make diagnosis somewhat imprecise.

Social Causation & Mental Illness

View Segments Segment :

Abstract

This film examines social causation—when a social factor causes a change in another social factor—and the relationship between social causation and mental illness.

Social Causation & Mental Illness

This film examines social causation—when a social factor causes a change in another social factor—and the relationship between social causation and mental illness.

Copy and paste the following HTML into your website

Back to Top