Fallen Angels: Twilight's Notes · 11:27am Oct 10th, 2014
Beside finishing the first half of the latest chapter of Fallen Angels, I've been researching (and distracted with the new Rainbow Rocks movie). Hence the delay. For your patience, have some notes.
How can that purple pony do this so quickly?!
I'll tell you how.
Genius.
!(Exclamation and question marks signify personal annotations.)
Previous Literature Reviews
Qualitative literature review
Assumed properties of CSA:
(mixed views and inconsistencies in)
Causality of maladjustment
- confounding of family environments prevents definitive conclusions
- recognizing limitations but likelihood of causality
- strongly implying causation without addressing third party vars.
Pervasiveness
- variable, no inevitable outcome from sexual factors
- effects/correlates prevalent: observed, unobserved, and assumed
Intensity
- varies; normal, slight, and intense psychological impairment
- absent in non-clinical samples
- intensely negative
- gender equivalence
- girls react more negatively, boys more neutral/positive
- equivalent experience
- called by some an "exercise in futility" to determine differences
Limitations of qualitative literature reviews include:
Sampling bias
Clinical and legal samples
- not representative
- extreme/serious cases
- information bias/"effort after meaning"
- investigator transmission of expectancies
- family disruption and CSA highly confounded
- only one review included a majority of nonclinical and nonlegal samples
Subjectivities and imprecision:
- narrative and therefore susceptible to reviewers' subjective interpretations
- confirmation bias, ignoring non-negative outcomes
- failure to use size of difference from control group to assess intensity
Quantitative literature reviews
Advancement in assessing CSA-adjustment relations:
Inclusion of sizable proportion of nonclinical and nonlegal samples
Avoidance of subjectivity and imprecision through meta-analysis
- meta-analysis converts statistics from a set of studies to a common metric to infer association between variables and estimate the strength of that association (small, medium, and large effect size)
- well-suited to examine whether and to what extent there are differences between control and SA respondents' adjustment
- precise gender comparison
Subjectivity and imprecision still apparent through convenience sampling instead of community sampling and unintentionally targeting those with more negative experience.
Method
Meta-analysis appears to address each of its own weaknesses and assumptions thoroughly and straightforwardly.
18 correlates examined:
1. Alcohol problems
2. Anxiety
3. Depression
4. Dissociation -- "depersonalization, memory loss, and not feeling like oneself"
5. Eating disorders
6. Hostility
7. Interpersonal sensitivity -- "uneasiness and marked discomfort when interacting with others, as well as feelings of personal inadequacy and inferiority, especially compared with others"
8. Locus of control -- "the extent to which one feels in control of one's life"
9. Obsessive-compulsive symptomatology -- "unremitting and irresistible thoughts, impulses, and actions that are ego alien or unwanted"
10. Paranoia -- "persistent fear response of an irrational and disproportionate nature to a specific person, place, object, or situation"
11. Psychotic symptoms -- "mental confusion and delusions (i.e., first-rank symptoms of schizophrenia such as hallucinations and thought-broadcasting)"
12. Self-esteem
14. Sexual adjustment -- Romantic and sexual self-esteem, sexual functioning, psychosexual functioning, sexual arousability, Erwin Identity Scale (confidence, sexual identity, and conceptions about body and appearance)
15. Social adjustment
16. Somatization -- "bodily related distress"
17. Suicidal ideation and behavior
18. Wide adjustment -- general well-being
Results
Existing definitions of CSA extremely varied.
Prevalence Rates and Types of CSA
Average prevalence of all types of CSA:
- male 14%
- female 27%
* All results are meta-analytical estimates.
Types (aka. severity/seriousness):
Invitation to do something sexual
- analysis excluded
Exhibitionism (either way)
- College: F=32%; M=22%; Combined=28%
- National: F=%38; M=25%; Combined=33%
Masturbation
Fondling (combined with masturbation, occasionally non-genital)
- College: F=39%; M=51%; Combined=42%
- National: F=67%; M=69%; Combined=68%
Oral sex
- College: F=3%; M=14%; Combined=6%
- National: F=9%; M=22%; Combined=14%
Attempted intercourse
- analysis excluded
Completed intercourse (sometimes combined with attempted)
- College: F=13%; M=33%; Combined=17%
- National: F=16%; M=13%; Combined=15%
Types, sometimes reported combined, divided evenly for analysis.
?- How does this affect accuracy?
?- How often are these kept distinct from non-combined? (Sometimes.)
?- Does trimming outliers for homogeneity improve accuracy?
!- All cases are explained and appear to be logical, noting any weaknesses that the combinations might produce.
Prevalence of less severe types likely to be underestimated due to frequent categorization of SA participants exclusively into most "severe"
College men shown to experience equal severity/type of CSA as women in general pop. but general pop. men show less severity. Women show consistent severity regardless of college or not.
Relationship Types:
Wider family (including close family)
- College: F=37%; M=23%; Combined=35%
- National: F=34%; M=13%; Combined=26%
Close family (parent, step-parent, grandparent, sibling)
- College: F=20%; M=8%; Combined=16%
- National: F=15%; M=4%; Combined=11%
College and national prevalence rates noted as "similar," implying college samples' CSA's severity is not less than the general population.
Frequency of CSA:
In both college and national samples was about the same, roughly 50% experiencing more than one CSA episode. Again, similar "severity" in college and general population
?- gender differences?
Force or threat of force:
- College: F=41%; M=23%
Large Standard Deviations (SD) indicate these estimates should be viewed cautiously (26% and 21% respectively).
The estimates above are, however, supported by an additional study's report of combined male and female students experiencing force or threat of force at a rate of 31%.
No national data available in the case of force or threat of force.
Psychological Adjustment
Effect size:
The estimate for 15,912 participants was statistically significant (i.e., SA students were less well adjusted than controls). The difference between SA students and controls was small, however, less than 1%, contradictory to typical assumptions that CSA is associated with intense harm.
- The term "small," used to describe effect size, is relative according to Cohen's original definition, but Rind accounts for adjustment variance, showing less than 1% difference between SA students and the control.
?- Is this variance adequate basis to safely assume the intensity of CSA effects is less than what many commonly purport?
Many of the participants who experienced adjustment issues experienced incest. When these participants were removed, as well as other counter-balancing outliers, the results were the same, however.
?- Incest or some incest correlate may be responsible for maladjustment in SA individuals.
See column r(u) in Table 3 of Rind's meta-analysis.
http://www.helping-people.info/rbt/ma_6.htm#Table%203
Homogeneity:
Of the 18 symptoms, hostility, self-esteem, and sexual adjustment effects were found to be heterogeneous after outliers were removed twice (i.e., they varied among meta-analyzed studies). However, they were still "small."
Moderators
Studies were inconsistent. In two studies, there were 5 common moderator types: force, penetration, duration, frequency, and incest. Several studies examined composites of these 5 types. Only the two were analyzed for individual types with the following results:
Force - associated with more negative reactions and self-reported effects, but unrelated to symptoms. (Magnitude: small)
Incest (close familial CSA) - associated both with more negative reactions/self-reported effects and with symptoms. (Magnitude: small)
Penetration, frequency and duration - unrelated to either reactions/effects or symptoms.
Composite measures used in other studies created too much inconsistency for any clear conclusions to be drawn concerning the relation between moderator types and their relation to reactions/effects and symptoms.
Self-Reported Reactions and Effects
Data reported in terms of individual reactions rather than number of reactions was converted instead into number of reactions for comparison's sake (i.e., one reaction per individual).
Remembered immediate reactions:
Positive: F=11%; M=37%
Neutral: F=18%; M=29%
Negative: F=72%; M=33%
Current reflections:
Positive: F=16%; M=42%
Neutral: F=25%; M=32%
Negative: F=59%; M=26%
- Similar to results for immediate remembered reactions.
- Supported by Haugaard and Emery's (1989) report.
Both remembered immediate reactions and current reflections show non-equivalence between male and female:
Female - generally negative
Male - neutral-positive
!- In an unrelated study, about half of young adult college females' first experiences with intercourse is self-reported as negative while all males reported positive. The number of first-intercourse experiences among recorded CSA self-reports is omitted or unavailable in Rind's meta-analysis. However, the factors that make intercourse less positive for adult women their first time may also account for some of the difference between male and female SA student reactions.
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3394901/
Self-reported effects:
Again, notable difference between SA men and women. Severe effects still the minority with men being less affected.
Sex lives
- A minortiy of both reported negative impact from CSA on sex life.
- A large majority reported no impact at all, both male and female.
Lasting negative effects/stress
- SA college women (! Estimates made informally based on mixed data.)
* none (about 5%)
* low (about 50%)
* moderate yet temporary (about 30%)
* persistent or permanent (about 5%)
- SA college men reported lasting effects rarely to any degree
Male vs. female self-report meta-analysis yielded results showing that SA college women report negative impacts from CSA much more frequently than men.
Family environments
Family environments are considered in order to address the issue of causation. Poor family environment is often confounded with negative impact in clinical CSA populations.
Family environment factors:
1. abuse and neglect
2. conflict or pathology
3. family structure
4. traditionalism
5. adaptability
6. support or bonding
(Effect sizes for adaptability and support-bonding were heterogeneous.)
All effect sizes are positive and statistically significant (from .09 to .19, small to medium), implying college students with a history of CSA come from more problematic home environments.
Family environment-symptom relations:
Smaller samples and heterogeneity lead Rind to recommend caution with these estimate results.
Effect size difference between CSA and symptoms (small) and family environment and symptoms (medium) indicate family environment is much more strongly related to symptoms.
"In terms of variance accounted for, family environment outperformed CSA in explaining symptoms by a factor of 9." In other words, family environment is 9 times more likely to cause symptoms than CSA.
Logically, the significance of the relationship between CSA and symptoms is only a step away from being "reduced to nonsignificance." All that is lacking is statistical control for family environment. Rind attempts to do this, but, due to inadequate reporting, only so much could be determined.
Statistical control results:
- before: 34 symptom relations (41%) were significant
- after and excluding possible interdependent symptoms: 14 (17%) are still satistically significant
A study by Wisniewski (1990) supports this reduction in CSA-symptom relation significance, showing no current adjustment (stress) due to CSA for women after controlling for extraneous factors except to a small degree in the case of incest CSA with adult revicitmization subjects. She found instead that other factors, particularly family violence, best explained current adjustment.
"[Results] provide direct evidence that the majority of significant CSA-symptom relations examined in the college samples may have been spurious." (42) They do not disprove symptom relations in the college population, but "they specifically do not support the assumption that CSA causes psychological injury."
Statistical validity problems for family environment relations:
- Unstandardized measures of CSA vs. family environment means lower reliability for CSA measures.
- Dichotomy of CSA (i.e., present or absent) skewed with a strong mode in the absent category results in low base rates.
- Artificial dichotomy of CSA fails to consider its continuous effect.
Validity pluses:
- More than half used in the meta-analysis are modifications of Finkelhor's measure, (reputable and reliable), and some others are proven reliable in their studies.
- 50-50 SA to control subjects is ideal, but 27-73 for females and 14-86 for males leaves room for only small margins of error which, thanks to small effect sizes, factored in, still yeild estimates that lead us to the same conclusion about relations between family environment, CSA, and symptoms. Empirically, low base rates are also shown not to be associated with lower effect sizes but larger ones, possibly supporting the decreased significance of CSA vs. family.
- Wisniewski's study used a continuous rather than dichotomous conceptualization of CSA while controlling for family variables and other factors and yeilded results similar to the meta-analysis of studies which assumed dichotomous CSA. This suggests that any possible false dichotomy does not influence results significantly.
Reliability of the statistical results of the meta-analysis appear unlikely to be a significant problem.
Discussion
"Commonly expressed opinions, both lay and professional, have implied that CSA possesses four basic properties: causality (it causes harm), pervasiveness (most SA persons are affected), intensity (harm is typically severe), and gender equivalence (boys and girls are affected equally)." (42)
Assumed Properties of CSA Revisited
1. Gender Equivalence
- "It is unsurprising that men and women should show similar differences in their reactions to CSA," (43)
2. Causality (harm)
- "Despite these caveats, the current results imply that the claim that CSA inevitably or usually produces harm is not justified," (43).
3. Pervasiveness (of effects)
- "In short, the self-reported effects data do not support the assumption of wide-scale psychological harm from CSA," (43).
4. intensity (severe)
- "This low intensity finding for generally negative CSA experiences is inconsistent with an expectation of intense harm from nonnegative CSA experiences," (43).
Moderators
"The intensity of the relationship between CSA and adjustment varied reliably as a function of gender, level of consent, and the interaction of these two factors," (44) but did not rely on contact, providing no support for the common assumption that contact sex is especially harmful, even for girls.
Only force (more so) and incest (less so) moderated outcomes and not penetration, duration, or frequency. Near-zero correlation between penetration and outcome is consistent with analysis finding that contact sex did not moderate adjustment. "This result provides empirical support for Finkelhor's (1979 , p. 103) observation that our society's view of intercourse as the most damaging form of CSA is 'a well-ingrained prejudice' unsupported by research," (45).
Usefulness of CSA as a Construct
"In most studies examined in the current review, CSA was defined based on legal and moral, rather than empirical and phenomenological, criteria. This approach may form a defensible rationale for legal restrictions of these behaviors, but is inadequate and may be invalid in the context of scientific inquiry ( Okami, 1994 ). In science, abuse implies that particular actions or inactions of an intentional nature are likely to cause harm to an individual (cf. Kilpatrick, 1987 ; Money & Weinrich, 1983 ). Classifying a behavior as abuse simply because it is generally viewed as immoral or defined as illegal is problematic, because such a classification may obscure the true nature of the behavior and its actual causes and effects," (45).
Rind is not the first to question the definition of sexual abuse, nor is child sex abuse the first behavior to be questioned historically for the validity of it's coining as "abuse" (e.g., masturbation). As currently defined, CSA is not a useful, predictive scientific construct.
!- It creates iatrogenic victims.
"The foregoing discussion does not imply that the construct CSA should be abandoned, but only that it should be used less indiscriminately to achieve better scientific validity. Its use is more scientifically valid when early sexual episodes are unwanted and experienced negatively-a combination commonly reported, for example, in father-daughter incest," (46) or in instances of force.
Due to reported positive reactions and reflections and lack of symptoms and self-labeling as victims in most sex abused men and some sex abused women, "it is important for researchers to be cautious in defining abuse for both men and women in attempts to validly examine the antecedents and effects of these experiences," (46).
Conclusion
Meta-analysis results do not support present assumptions about the properties of CSA. "The finding that college samples closely parallel national samples with regard to prevalence of CSA, types of experiences, self-perceived effects, and CSA-symptom relations strengthens the conclusion that CSA is not a propertied phenomenon and supports Constantine's (1981) conclusion that CSA has no inbuilt or inevitable outcome or set of emotional reactions," (46).
"Overinclusive definitions, ..., produce poor predictive validity, ..., [A] more thoughtful approach is needed by researchers when labeling and categorizing events that have heretofore been defined sociolegally as CSA," (46) possibly by focusing on a young person's perceived willingness and reactions, by labeling willing, positive encounters with a value-neutral term like "adult-child sex," and reserving the term "child sexual abuse" for unwilling or negatively perceived encounters. The term "child," also, would better be reserved for nonadolescent children, and "adult-adolescent sex" termed separately from "adult-child sex" and from "adolescent sexual abuse."
"Finally, it is important to consider implications of the current review for moral and legal positions on CSA. If it is true that wrongfulness in sexual matters does not imply harmfulness ( Money, 1979 ), then it is also true that lack of harmfulness does not imply lack of wrongfulness. Moral codes of a society with respect to sexual behavior need not be, and often have not been, based on considerations of psychological harmfulness or health (cf. Finkelhor, 1984 ). Similarly, legal codes may be, and have often been, unconnected to such considerations ( Kinsey et al., 1948 ). In this sense, the findings of the current review do not imply that moral or legal definitions of or views on behaviors currently classified as CSA should be abandoned or even altered. The current findings are relevant to moral and legal positions only to the extent that these positions are based on the presumption of psychological harm," (47).
References and original article
http://www.helping-people.info/rbt/metaana.htm
http://eric.ed.gov/?id=EJ220805
http://en.wikipedia.org/wiki/Homogeneity_(statistics)
http://en.wikipedia.org/wiki/Effect_size
http://en.wikipedia.org/wiki/Analysis_of_variance
ohgodthat'salotofwords.
i've been up reading "Hard Reset" and "Anthropology" for the last.. 13 hours.. i should probably sleep soon, THEN i'll read whatever this post is about(i've a rough idea).
2522993
Sleep well!
one word
devotion
2523101 data2.whicdn.com/images/77960759/thumb.jpg
~reads through and nods~ mmmhmm I understand some of these words.
Geez you weren't kidding when you said you were researching haha
Very cool!
I've never really known what to say about the Rind study... It's kind of funny that it was officially condemned by Congress, the only peer-reviewed study ever to have the honor. Worth noting that its results were later replicated by another team of researchers.
2538242 I'd actually really like any more information about the Rind study that you might be able to give me. I've been given some reviews that make it their business to tear down the Rind study, but I haven't gotten through more than their abstracts yet. It would really help me understand some of the strengths of the Rind study if I could look at another that replicates its results.
2538554 this is Heather Ulrich's replication of the study, I can't find a pdf of it though. If you're looking at NARTH's rebuttal to the analysis (the main one I've run into in the past), well, consider that their acronym stands for National Association for Research & Therapy of Homosexuality. Not exactly a slam-dunk in Rind's favor, but might suggest they've got a certain bias of their own.
On the other hand, the site you're looking at the study on, helping-people.info, seems to be associated with IPCE, the International Pedophile and Child Emancipation organization. That's my guess at least - they seem to have the same web designer but IPCE is a much larger site. It also happens to host an extensive set of comments back and forth between the Rind authors and their critics, if you have the patience to leaf through it all. All I know is the confluence of moral outrage and heated discussion regarding p-values makes my eyes glaze over.
My personal takeaway from all of it is that sexual interactions between adults and children pose a risk of harm to the child that is significantly higher than in the case of two adults, but that risk is also variable based on certain factors in the relationship. Theoretically that should be good news, not to mention more or less intuitive, but it calls into question cultural truths that are more deep-seated than they seem at first glance.
2542725 Nice links! Thanks. I'll probably look for the supporting article you mentioned.
Yeah, the site on which Rind's article is hosted and several other things I've read in defense of adult-child relationships is, at least as far as I can tell, a clearly pro-pedo site. I have more than one rebuttal to the Rind analysis that I'm going to have to sift through, but I'll be glad to put anything by NARTH on the back-burner or wait until I need a more biased science-y thing to use in my writing, so thanks for letting me know. I have read one response from Rind to their critics that seems to cover a number of issues, but I'm glad to have the additional responses you provided.
Thanks again.
2542725 In case you were curious, these are the articles I've been given to look at that respond to the Rind analysis.
http://psycnet.apa.org/psycinfo/2001-05308-001
http://psycnet.apa.org/psycinfo/2001-05308-002
http://www.tandfonline.com/doi/abs/10.1080/00223980109603677#.VEPMA_l4pwl
Also, is this the study by Ulrich you mentioned? It looks like it, but it's from 2008 I believe:
http://etd.lib.umt.edu/theses/available/etd-09202007-123302/unrestricted/umi-umt-1028.pdf