Research: Methodological flaws and syndrome construction - Revision history

Jim Burton: /* Limitations in attempts to address third factors */

2026-04-23T11:21:24Z

Limitations in attempts to address third factors

← Older revision		Revision as of 11:21, 23 April 2026
Line 81:		Line 81:
	* It also did not control for differences in life course development which might contain powerful mediators that explain much of the variance in symptoms.		* It also did not control for differences in life course development which might contain powerful mediators that explain much of the variance in symptoms.

	==== Limitations in attempts to address third factors ====		==== Limitations of attempts to address third factors ====

	Our anthologies of studies contain much data on the results of statistical control of third variables, which often show weak or non-significant correlations of sexual abuse with symptoms. However, these data should be used with great caution and should not be considered sufficient evidence of the absence of harmful effects, because:		Our anthologies of studies contain much data on the results of statistical control of third variables, which often show weak or non-significant correlations of sexual abuse with symptoms. However, these data should be used with great caution and should not be considered sufficient evidence of the absence of harmful effects, because:

Jim Burton: /* Limitations in attempts of consideration of the third factors */

2026-04-23T11:20:13Z

Limitations in attempts of consideration of the third factors

← Older revision		Revision as of 11:20, 23 April 2026
Line 81:		Line 81:
	* It also did not control for differences in life course development which might contain powerful mediators that explain much of the variance in symptoms.		* It also did not control for differences in life course development which might contain powerful mediators that explain much of the variance in symptoms.

	==== Limitations in attempts ~~of consideration of the~~ third factors ====		==== Limitations in attempts to address third factors ====

	Our anthologies of studies contain much data on the results of statistical control of third variables, which often show weak or non-significant correlations of sexual abuse with symptoms. However, these data should be used with great caution and should not be considered sufficient evidence of the absence of harmful effects, because:		Our anthologies of studies contain much data on the results of statistical control of third variables, which often show weak or non-significant correlations of sexual abuse with symptoms. However, these data should be used with great caution and should not be considered sufficient evidence of the absence of harmful effects, because:

Jim Burton: /* Examples of problematic vagueness of the CSA */

2026-04-23T11:16:05Z

Examples of problematic vagueness of the CSA

← Older revision		Revision as of 11:16, 23 April 2026
Line 142:		Line 142:
	*: "[O]ne of the most notable evolutions in how these experiences are treated has been the progressive disappearance and/or the intense redefinition of what earlier researchers called "participating victims," i.e., minors apparently interested in accepting and/or sustaining these relationships.[...] [T]he current ‘‘taboo’’ against the scientific study of these children and their experiences contributes nothing positive, but is instead a serious impediment to any appropriate understand in this area."		*: "[O]ne of the most notable evolutions in how these experiences are treated has been the progressive disappearance and/or the intense redefinition of what earlier researchers called "participating victims," i.e., minors apparently interested in accepting and/or sustaining these relationships.[...] [T]he current ‘‘taboo’’ against the scientific study of these children and their experiences contributes nothing positive, but is instead a serious impediment to any appropriate understand in this area."

	==== Examples of ~~problematic vagueness of the~~ CSA ====		==== Examples of CSA as a vague concept ====
	*'''Wang Y, Sun Y, Missmer S A, Rexrode K M, Roberts A L, Chavarro J E et al. (2023) [https://www.bmj.com/content/381/bmj-2022-073613 Association of early life physical and sexual abuse with premature mortality among female nurses: prospective cohort study] ''BMJ''; 381, doi:10.1136/bmj-2022-073613'''		*'''Wang Y, Sun Y, Missmer S A, Rexrode K M, Roberts A L, Chavarro J E et al. (2023) [https://www.bmj.com/content/381/bmj-2022-073613 Association of early life physical and sexual abuse with premature mortality among female nurses: prospective cohort study] ''BMJ''; 381, doi:10.1136/bmj-2022-073613'''
	*:“67 726 female nurses included in the current analysis” [...] “Sexual abuse was measured using modified questions from the Sexual Experiences Survey: “Were you ever touched in a sexual way” or “forced into any sexual activity” by an adult or an older child?”		*:“67 726 female nurses included in the current analysis” [...] “Sexual abuse was measured using modified questions from the Sexual Experiences Survey: “Were you ever touched in a sexual way” or “forced into any sexual activity” by an adult or an older child?”

Jim Burton: /* Limitations in attempts of consideration of the third factors */

2026-04-23T11:15:16Z

Limitations in attempts of consideration of the third factors

← Older revision		Revision as of 11:15, 23 April 2026
Line 92:		Line 92:
	* '''Vachon DD, et al. (2015) Assessment of the Harmful Psychiatric and Behavioral Effects of Different Forms of Child Maltreatment.'''		* '''Vachon DD, et al. (2015) Assessment of the Harmful Psychiatric and Behavioral Effects of Different Forms of Child Maltreatment.'''
	::"[C]hild sexual abuse is an infrequent event that is almost always accompanied by other types of CM[child maltreatment]. This pattern of rarity and lopsided co-occurrence has several consequences.[...] [T]here is no practical way to understand the specific consequences of sexual CM because its correlates may be attributed to other forms of co-occurring CM. Statistically controlling for co-occurring CM removes what little covariation is left after the attenuation caused by unidirectional redundancy, further gutting the variance in the sexual abuse variable and producing unreliable parameter estimates. Alternatively, cases of “pure” sexual abuse (without co-occurring CM) are extremely rare and unrepresentative. This intractable issue may explain why research on sexual CM produces mixed results."		::"[C]hild sexual abuse is an infrequent event that is almost always accompanied by other types of CM[child maltreatment]. This pattern of rarity and lopsided co-occurrence has several consequences.[...] [T]here is no practical way to understand the specific consequences of sexual CM because its correlates may be attributed to other forms of co-occurring CM. Statistically controlling for co-occurring CM removes what little covariation is left after the attenuation caused by unidirectional redundancy, further gutting the variance in the sexual abuse variable and producing unreliable parameter estimates. Alternatively, cases of “pure” sexual abuse (without co-occurring CM) are extremely rare and unrepresentative. This intractable issue may explain why research on sexual CM produces mixed results."


	When discussing the effects of sexual abuse/contact, it is vital to bring in additional evidence and not make definitive conclusions to avoid appearing ignorant and biased. Additional evidence could be:		When discussing the effects of sexual abuse/contact, it is vital to bring in additional evidence and not make definitive conclusions to avoid appearing ignorant and biased. Additional evidence could be:

Jim Burton: /* Limitations in attempts of consideration of the third factors */

2026-04-23T11:14:54Z

Limitations in attempts of consideration of the third factors

← Older revision		Revision as of 11:14, 23 April 2026
Line 82:		Line 82:

	==== Limitations in attempts of consideration of the third factors ====		==== Limitations in attempts of consideration of the third factors ====
	Our anthologies of studies contain ~~many~~ data on the results of statistical control of third variables, which often show weak or non-significant correlations of sexual abuse with symptoms. However, these data should be used with great caution and should not be considered sufficient evidence of the absence of harmful effects, because:
			Our anthologies of studies contain much data on the results of statistical control of third variables, which often show weak or non-significant correlations of sexual abuse with symptoms. However, these data should be used with great caution and should not be considered sufficient evidence of the absence of harmful effects, because:

	*'''Briere J. (1988) Controlling for family variables in abuse effects research: a critique of the ‘‘partialling’’ approach. J ''Interpers Violence'' ;3:80–89.'''		*'''Briere J. (1988) Controlling for family variables in abuse effects research: a critique of the ‘‘partialling’’ approach. J ''Interpers Violence'' ;3:80–89.'''

Jim Burton: /* Lack of consideration of third factors */

2026-04-23T11:09:25Z

Lack of consideration of third factors

← Older revision		Revision as of 11:09, 23 April 2026
Line 44:		Line 44:
	*Harmful secondary reactions (i.e. exaggerated parental reactions, invasive investigations and family disruption due to legal threats, the trauma myth in therapy, media-induced phobias)		*Harmful secondary reactions (i.e. exaggerated parental reactions, invasive investigations and family disruption due to legal threats, the trauma myth in therapy, media-induced phobias)

			See examples in the articles:

	~~See examples in the articles:~~
	*[[Research: Association or Causation]]		*[[Research: Association or Causation]]
	*[[Research: Family Environment]]		*[[Research: Family Environment]]

Thorn: /* Example */

2025-10-09T17:40:28Z

Example

← Older revision		Revision as of 17:40, 9 October 2025
Line 71:		Line 71:

	==== Example ====		==== Example ====
	For instance, the study by ''Nelson et al. (2002), "Association Between Self-reported Childhood Sexual Abuse and Adverse Psychosocial Outcomes''," used a twin design, which gives it quasi-experimental status and makes it one of the most methodologically rigorous approaches available. Because twins share a common genetics and family environment, this design effectively controls for these confounds. Despite this, the CSA “effect” persisted (however significantly diminished), lending more plausibility to the idea of a genuine traumatic impact. However:		For instance, the study by ''Nelson et al. (2002), "Association Between Self-reported Childhood Sexual Abuse and Adverse Psychosocial Outcomes''," used a twin design, which gives it quasi-experimental status and makes it one of the most methodologically rigorous approaches available. Because twins share common genetics and a family environment, this design effectively controls for these confounds. Despite this, the CSA “effect” persisted (however significantly diminished), lending more plausibility to the idea of a genuine traumatic impact. However:

	*The definition of CSA in the study included rape, coercion, and any sexual contact whatsoever, thereby collapsing very different levels of severity into a single category.		*The definition of CSA in the study included rape, coercion, and any sexual contact whatsoever, thereby collapsing very different levels of severity into a single category.

Thorn: /* Example */

2025-09-09T13:43:47Z

Example

← Older revision		Revision as of 13:43, 9 September 2025
Line 75:		Line 75:
	*The definition of CSA in the study included rape, coercion, and any sexual contact whatsoever, thereby collapsing very different levels of severity into a single category.		*The definition of CSA in the study included rape, coercion, and any sexual contact whatsoever, thereby collapsing very different levels of severity into a single category.

	*The study did not control for differentially negative treatment between the twins—for example, differing levels of emotional abuse or neglect—which, as shown in research on [[Research:_Prevalence_of_Harm_and_Negative_Outcomes#Poly_Victimizations_and_Comorbidities\|polyvictimization]], tend to have stronger and more stable effects than CSA alone.		*The study did not control for differentially negative treatment between the twins—for example, differing levels of emotional abuse or neglect—which, as shown in research on [[Research:_Prevalence_of_Harm_and_Negative_Outcomes#Poly_Victimizations_and_Comorbidities\|polyvictimization]], tend to have stronger and more stable effects than CSA alone. One of the basic and frequently replicated findings from behavioral genetics is that ''"Most environmental effects are not shared by children growing up in the same family"''.<ref>Plomin, R., DeFries, J. C., Knopik, V. S., & Neiderhiser, J. M. (2016). [https://scottbarrykaufman.com/wp-content/uploads/2016/05/2016-plomin.pdf Top 10 Replicated Findings From Behavioral Genetics.] ''Perspectives on psychological science : a journal of the Association for Psychological Science'', 11(1), 3–23. https://doi.org/10.1177/1745691615617439</ref>

	*It did not account for differences in personality traits or sexuality between the twins. Even monozygotic twins can differ significantly in these domains. For instance, concordance rates for sexual orientation among monozygotic twins have been estimated at only 20–50% ~~(according to ''~~Bailey et al. (2000) Genetic and environmental influences on sexual orientation and its correlates in an Australian twin sample~~'')~~. Similarly, personality trait correlations among monozygotic twins are only in the range of r = 0.28 to 0.54 ~~(according to ''~~Susan C. South et al. (2018) Sex differences in the Big Five model personality traits: A behavior genetics exploration~~''.)~~		*It did not account for differences in personality traits or sexuality between the twins. Even monozygotic twins can differ significantly in these domains. For instance, concordance rates for sexual orientation among monozygotic twins have been estimated at only 20–50% <ref>Bailey et al. (2000) [https://pubmed.ncbi.nlm.nih.gov/10743878/ Genetic and environmental influences on sexual orientation and its correlates in an Australian twin sample]</ref>. Similarly, personality trait correlations among monozygotic twins are only in the range of r = 0.28 to 0.54.<ref>Susan C. South et al. (2018) [https://www.sciencedirect.com/science/article/abs/pii/S0092656618300321 Sex differences in the Big Five model personality traits: A behavior genetics exploration]</ref>

	* It also did not control for differences in life course development which might contain powerful mediators that explain much of the variance in symptoms.		* It also did not control for differences in life course development which might contain powerful mediators that explain much of the variance in symptoms.

Thorn: /* Limitations in attempts of consideration of the third factors */

2025-08-03T07:58:09Z

Limitations in attempts of consideration of the third factors

← Older revision		Revision as of 07:58, 3 August 2025
Line 99:		Line 99:
	*The correction for the rarity of abuse in Rind et al. (1998) samples (the “50–50 split”) increases the effect size by only a negligible amount (≤ .03 in r), so low base rate alone cannot account for small raw effects (though this doesn’t resolve the problem with partialling).		*The correction for the rarity of abuse in Rind et al. (1998) samples (the “50–50 split”) increases the effect size by only a negligible amount (≤ .03 in r), so low base rate alone cannot account for small raw effects (though this doesn’t resolve the problem with partialling).

	* Many studies reporting raw (unadjusted) correlations across maltreatment types show that sexual abuse (under a broad legal definition) has the smallest association among maltreatment forms even before controlling for co-occurrence (see [[Research:_Prevalence_of_Harm_and_Negative_Outcomes]]).		* Many studies reporting raw (unadjusted) correlations across maltreatment types show that sexual abuse (under a broad legal definition) has the smallest association among maltreatment forms even before controlling for co-occurrence (see [[Research:_Prevalence_of_Harm_and_Negative_Outcomes\|Prevalence of Harm and Negative Outcomes]]).

	*Twin studies provide a strongly reliable adjustment for shared genetics and family environment. They indicate these factors indeed (though not completely) account for a large portion of the CSA–symptom association, so FE confounding is not mere speculation. At the same time, even good twin studies do not adjust for all other factors by which twins might be discordant (see the example above).		*Twin studies provide a strongly reliable adjustment for shared genetics and family environment. They indicate these factors indeed (though not completely) account for a large portion of the CSA–symptom association, so FE confounding is not mere speculation. At the same time, even good twin studies do not adjust for all other factors by which twins might be discordant (see the example above).
Line 105:		Line 105:
	*Enumerate all plausible third factors (see above). Although exact unique effects aren’t available, a hypothetical model that adjusts for all confounders and mediators leaves little room for an “implicit harm” property: differential emotional abuse/neglect, other maltreatment, parental psychopathology, SES, temperament/personality, sexual orientation/behavioral tendencies, early sexual debut, dysfunctional socialization, substance use, school/neighborhood adversity, etc.		*Enumerate all plausible third factors (see above). Although exact unique effects aren’t available, a hypothetical model that adjusts for all confounders and mediators leaves little room for an “implicit harm” property: differential emotional abuse/neglect, other maltreatment, parental psychopathology, SES, temperament/personality, sexual orientation/behavioral tendencies, early sexual debut, dysfunctional socialization, substance use, school/neighborhood adversity, etc.

	*Studies showing absence or inconsistency of dose–response effect for abuse (e.g., Browning & Laumann (1997; 2003); Rind et al. (1998)). "CSA severity" doesn't necessarily associated with worse consequences. This weakens claims about an intrinsic harmful property of sex per se.		*Studies showing absence or inconsistency of dose–response effect for abuse (e.g., Browning & Laumann (1997; 2003); Rind et al. (1998)). "CSA severity" doesn't necessarily associated with worse consequences. This weakens claims about an intrinsic harmful property of sex per se. (see [[Research:_Prevalence_of_Harm_and_Negative_Outcomes#Outcomes_and_%22severity%22\|Outcomes and "severity"]])

	*Studies showing that peer sexual contacts have effect sizes similar to adult–minor contacts. That pattern is difficult to reconcile with an “inherent inter-generational harm” hypothesis. <ref name="Rind2001" />		*Studies showing that peer sexual contacts have effect sizes similar to adult–minor contacts. That pattern is difficult to reconcile with an “inherent inter-generational harm” hypothesis. <ref name="Rind2001" />

Thorn: /* Limitations in attempts of consideration of the third factors */

2025-08-03T07:42:02Z

Limitations in attempts of consideration of the third factors

← Older revision		Revision as of 07:42, 3 August 2025
Line 95:		Line 95:
	When discussing the effects of sexual abuse/contact, it is vital to bring in additional evidence and not make definitive conclusions to avoid appearing ignorant and biased. Additional evidence could be:		When discussing the effects of sexual abuse/contact, it is vital to bring in additional evidence and not make definitive conclusions to avoid appearing ignorant and biased. Additional evidence could be:

	*Causal structure matters for validity of partialling. Whether sexual abuse itself leads to family dysfunction (third factor), or vice versa, affects the validity of partialling. You can emphasize studies on extra-familial abuse, where CSA→FE is less plausible.<ref name="Rind2001">Rind, B., Tromovitch, P., & Bauserman, R. (2001). The validity and appropriateness of methods, analyses, and conclusions in Rind et al. (1998): A rebuttal of victimological critique from Ondersma et al. (2001) and Dallam et al. (2001). Archives of Sexual Behavior, 30(5), 463–499. ~~https~~:~~//doi.org/~~10.1023/A:1010271214506</ref>		*Causal structure matters for validity of partialling. Whether sexual abuse itself leads to family dysfunction (third factor), or vice versa, affects the validity of partialling. You can emphasize studies on extra-familial abuse, where CSA→FE is less plausible.<ref name="Rind2001">Rind, B., Tromovitch, P., & Bauserman, R. (2001). The validity and appropriateness of methods, analyses, and conclusions in Rind et al. (1998): A rebuttal of victimological critique from Ondersma et al. (2001) and Dallam et al. (2001). Archives of Sexual Behavior, 30(5), 463–499. Doi:10.1023/A:1010271214506</ref>

	*The correction for the rarity of abuse in Rind et al. (1998) samples (the “50–50 split”) increases the effect size by only a negligible amount (≤ .03 in r), so low base rate alone cannot account for small raw effects (though this doesn’t resolve the problem with partialling).		*The correction for the rarity of abuse in Rind et al. (1998) samples (the “50–50 split”) increases the effect size by only a negligible amount (≤ .03 in r), so low base rate alone cannot account for small raw effects (though this doesn’t resolve the problem with partialling).