Research: Methodological flaws and syndrome construction - Revision history

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).

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

2025-08-02T14:40:12Z

Limitations in attempts of consideration of the third factors

← Older revision		Revision as of 14:40, 2 August 2025
Line 93:		Line 93:


	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.		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>

			*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]]).

			*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).

			*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 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 highlighting the importance of life-course deviation mechanisms in explaining the CSA–symptom link (e.g., Browning & Laumann (1997; 2003); <ref name="Rind2001" />

	=== Absence of comparison group ===		=== Absence of comparison group ===

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

2025-08-02T13:45:45Z

Limitations in attempts of consideration of the third factors

← Older revision		Revision as of 13:45, 2 August 2025
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 a sufficient evidence of the absence of harmful effects, because:		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:

	*'''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.'''

Thorn: /* Lack of consideration of third factors */

2025-08-02T13:18:38Z

Lack of consideration of third factors

← Older revision		Revision as of 13:18, 2 August 2025
Line 41:		Line 41:
	*Accompanied maltreatment (i.e. emotional, physical)		*Accompanied maltreatment (i.e. emotional, physical)
	*Socioeconomic context (i.e. possibilities of education and employment, environmental criminality)		*Socioeconomic context (i.e. possibilities of education and employment, environmental criminality)
	*Later life stressors, life course development (i.e. job loss, death of relative, socialization ~~within~~ subcultures)		*Later life stressors, life course development (i.e. job loss, death of relative, socialization in dysfunctional subcultures)
			*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:

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

2025-08-02T10:54:00Z

Limitations in attempts of consideration of the third factors

← Older revision		Revision as of 10:54, 2 August 2025
Line 86:		Line 86:

	::"In summary, it appears that the use of partialling control procedures can be problematic when the goal is to determine the relative importance of sexual abuse (or other variables) in the etiology of adult symptomatology. [...] Additionally, the algebra of the semipartial correlation ensures a conservative test: In order to be significant, abuse must correlate with symptoms after all variance shared with the control variable (e.g., family background) has been removed. Thus studies that show that abuse continues to predict later adjustment after “control’”’ variance has been extracted (e.g., Bagley & McDonald, 1984; Bagley & Ramsay, 1985; Finkelhor, 1984; Peters, 1984) should be considered quite seriously, whereas those with negative findings (e.g., Fromuth, 1986) do not necessarily indicate that sexual abuse has no psychological impact."		::"In summary, it appears that the use of partialling control procedures can be problematic when the goal is to determine the relative importance of sexual abuse (or other variables) in the etiology of adult symptomatology. [...] Additionally, the algebra of the semipartial correlation ensures a conservative test: In order to be significant, abuse must correlate with symptoms after all variance shared with the control variable (e.g., family background) has been removed. Thus studies that show that abuse continues to predict later adjustment after “control’”’ variance has been extracted (e.g., Bagley & McDonald, 1984; Bagley & Ramsay, 1985; Finkelhor, 1984; Peters, 1984) should be considered quite seriously, whereas those with negative findings (e.g., Fromuth, 1986) do not necessarily indicate that sexual abuse has no psychological impact."

			* '''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."


	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.		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.

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

2025-08-02T10:40:15Z

Limitations with attempts of consideration of the third factors

← Older revision		Revision as of 10:40, 2 August 2025
Line 79:		Line 79:
	* 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 ~~with~~ 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 a sufficient evidence of the absence of harmful effects, because:		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 a sufficient evidence of the absence of harmful effects, because:

Thorn: /* Lack of consideration of third factors */

2025-08-02T10:38:56Z

Lack of consideration of third factors

← Older revision		Revision as of 10:38, 2 August 2025
Line 61:		Line 61:
	::"Although the results of this systematic review provide clear evidence that the relationship between child sexual abuse and health problems does exist, the presence of confounding variables and the generally poor quality of the studies included in each review do not allow for causal inferences to be made, thus findings must be interpreted with caution. First, studies included in each review were generally methodologically limited (e.g., poor sampling methods, absence of matched comparison groups, poor description of methods of eliciting abuse histories, inadequate control for effect modifiers and confounders, or designs inappropriate to prove causality). Consequently, associations were confounded and causal inferences not feasible"		::"Although the results of this systematic review provide clear evidence that the relationship between child sexual abuse and health problems does exist, the presence of confounding variables and the generally poor quality of the studies included in each review do not allow for causal inferences to be made, thus findings must be interpreted with caution. First, studies included in each review were generally methodologically limited (e.g., poor sampling methods, absence of matched comparison groups, poor description of methods of eliciting abuse histories, inadequate control for effect modifiers and confounders, or designs inappropriate to prove causality). Consequently, associations were confounded and causal inferences not feasible"

	~~Some other studies that focus on this issue and~~ infer the same:		Other meta-reviews infer the same:
	*'''Maniglio, R. (2010) – Child sexual abuse in the etiology of depression: A systematic review of reviews. ''Depression and Anxiety'', 27, 631-642.'''		*'''Maniglio, R. (2010) – Child sexual abuse in the etiology of depression: A systematic review of reviews. ''Depression and Anxiety'', 27, 631-642.'''
	*'''Maniglio, R. (2011a) – The role of child sexual abuse in the etiology of substance-related disorders. ''Journal of Addictive Diseases'', 30, 216-228.'''		*'''Maniglio, R. (2011a) – The role of child sexual abuse in the etiology of substance-related disorders. ''Journal of Addictive Diseases'', 30, 216-228.'''
	*'''Maniglio, R. (2011b) – The role of child sexual abuse in the etiology of suicide and non-suicidal self-injury. ''Acta Psychiatrica'', 124, 30-41.'''		*'''Maniglio, R. (2011b) – The role of child sexual abuse in the etiology of suicide and non-suicidal self-injury. ''Acta Psychiatrica'', 124, 30-41.'''
	*'''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, Elliot DM. (1993) Sexual abuse, family environment, and psychological symptoms: on the validity of statistical control. J ''Consult Clin Psychol''; 61:284–288.'''

	Even studies attempting to account for third variables—even those with strong baseline methodology—often fail to address all relevant categories of third-variable confounding.		Even studies attempting to account for third variables—even those with strong baseline methodology—often fail to address all relevant categories of third-variable confounding.
Line 80:		Line 78:

	* 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 with 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 a 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.'''
			::"If two or more predictor variables are intercorrelated (a phenomenon sometimes referred to as ‘‘multicolinearity’’), and each are correlated with the criterion variable, it is quite difficult to determine the unique (nonredundant) contribution of any given predictor to the variance in the criterion. [...] Stated in the current context, since negative family variables are often associated with sexual abuse (e.g., Finkelhor, 1980; Fromuth, 1986), and since family problems and sexual abuse are each associated with later psychological difficulties (e.g., Bagley & McDonald, 1984; Briere & Runtz, 1986, 1987; Browne & Finkelhor, 1986; Dorpat, 1965), there is no easy way to determine how much of adult symptomatology is “due’’ to abuse as opposed to family influences."

			::"In summary, it appears that the use of partialling control procedures can be problematic when the goal is to determine the relative importance of sexual abuse (or other variables) in the etiology of adult symptomatology. [...] Additionally, the algebra of the semipartial correlation ensures a conservative test: In order to be significant, abuse must correlate with symptoms after all variance shared with the control variable (e.g., family background) has been removed. Thus studies that show that abuse continues to predict later adjustment after “control’”’ variance has been extracted (e.g., Bagley & McDonald, 1984; Bagley & Ramsay, 1985; Finkelhor, 1984; Peters, 1984) should be considered quite seriously, whereas those with negative findings (e.g., Fromuth, 1986) do not necessarily indicate that sexual abuse has no psychological impact."

			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.

	=== Absence of comparison group ===		=== Absence of comparison group ===