Attorneys’ Questions and Children’s Responses Referring to the Nature of Sexual Touch in Child Sexual Abuse Trials

Abstract

Eliciting clear descriptions of sexual body parts and abusive touch in child sexual abuse trials is challenging because of children’s immaturity and embarrassment. This study examined references to sexual body part knowledge and sexual touch in attorneys’ questions and 5- to 10-year-old children’s responses (N = 2,247) in 113 child sexual abuse trials. Regardless of children’s age, attorneys and children primarily used unclear colloquialisms to refer to sexual body parts. Questions asking children to name sexual body parts elicited more uninformative responses than questions about the function of sexual body parts. In turn, questions about the function of sexual body parts were more likely to increase the specificity of body part identifications than questions about the location of sexual body parts. Attorneys predominantly used option-posing (yes-no and forced choice) questions to ask about sexual body part knowledge, the location of touch, the method or manner of touching, skin-to-skin contact, penetration, and how the touching felt. Generally, wh- questions were no more likely than option-posing questions to elicit uninformative responses, and consistently elicited more child-generated information. The results question the legal assumption that children’s uninformative responses when testifying about sexual abuse should be overcome by asking option-posing questions.

Introduction

Children’s ability to describe the location and nature of touch is of key importance in child sexual abuse trials. Children often need to provide detailed descriptions of genital or anal touch to establish that the touch was sexually motivated and to establish the severity of the abuse. Courts have recognized children’s difficulties in specifying the location and nature of touch, but nevertheless expect them to provide details. This study investigated attorneys’ questions and 5- to 10-year-old children’s responses referring to sexual body parts and sexual touch involving female genitalia and the anus in criminal child sexual abuse trials. In the following paragraphs, we review the legal basis for questioning child victims about sexual touch and examine research on children’s descriptions of sexual body parts and on the types of questions attorneys and forensic interviewers use to elicit clear reports of sexual touch.

Professionals who question children often feel compelled by sexual abuse laws to attempt to elicit highly specific information about sexual touch. Although in some states, any sexually motivated touching of a child can be sexual abuse (e.g., California Penal Code Section 288), many states require touching of the child’s sexual parts, which may include the genitalia, inner thighs, buttocks, anus, and breast (Smith, 2012). In any case, the nature and invasiveness of the touch helps distinguish between sexually motivated touch and touch better construed as accidental, playful, or affectionate, and in the case of younger children, as toileting, bathing, or dressing the child. For example, it is more difficult to give a non-sexual justification for genital touch if the touching was skin-to-skin, under the child’s clothing, and penetrative. Furthermore, many states make legal distinctions between different levels of invasiveness, so that more invasive touch is a more serious crime, and distinguish between vaginal and anal touch in defining specific crimes (Smith, 2012). Furthermore, regardless of the specific crime charged, professionals are also likely to ask children to specify the nature and location of touch in order to elicit a more detailed and convincing report.

Children’s Descriptions of Sexual Body Parts

Identifying the body parts involved in the abuse at the required level of specificity can be challenging, because young children rarely use anatomically correct terms such as “vagina” and “anus” (Gordon et al., 1990; Kenny & Wurtele, 2008). Although it may not be legally necessary for children to use anatomically correct terms (Burrows & Powell, 2014), slang terms are often imprecise. Examining transcripts of 130 5- to 12-year-olds testifying about sexual abuse, Sullivan et al. (2021) found that the use of colloquial names for body parts sometimes led to miscommunications, including when children used terms that did not clearly distinguish between the vaginal and anal area (e.g. “the bottom and the top”) or when attorneys used euphemisms (e.g. “private parts”) without establishing children’s understanding.

In a sample of 585 5- to 17-year-old girls who named their sexual parts in a forensic genital examination, Gallion et al. (2016) found that “private” was the most common term used by children under 12 years of age. Examining questions and answers referring to sexual body parts in 161 forensic interviews about sexual abuse with 4 to 17-year-old children that were referred for prosecution, Burrows et al. (2017) found that 48% of children’s initial descriptions failed to distinguish between the vaginal and anal area (e.g., “privates,” “hole”). Interestingly, the proportion of clear terms showed a slight decrease with age, from 50% in the 4–7 age group to 40% among 8 to 10-year-olds, suggesting that the use of unclear terms may have resulted from reluctance or embarrassment in addition to ignorance or immaturity.

In 83% of the forensic interviews analyzed by Burrows et al. (2017), interviewers attempted to clarify terms by asking if the child knew another name for the body part in question, but the majority of children’s responses were either uninformative or provided similarly unclear terms. In only about a third of the cases were children able to provide either specific anatomical terms (22%) or clear colloquial terms (17%; e.g., “dick,” “willy”). Clarifying questions about the function (e.g., “What do you use your private for?”) or location of sexual body parts (e.g., “Where is your gina?”) were more productive but were asked less frequently. Furthermore, whereas younger children were less likely than older children to provide specific names, there were no age differences in the effectiveness of function and location questions in eliciting informative responses. Hence, questions about function and location appeared superior to questions about the naming of sexual body parts.

Questioning Children About Sexual Body Parts

It is also important to examine how interviewers and attorneys phrase questions about sexual body part knowledge. In sexual abuse trials involving child witnesses, attorneys rely on option-posing questions, including yes-no and forced choice questions (Andrews et al., 2016; Stolzenberg & Lyon, 2014; Stolzenberg et al., 2020). When children appear to have difficulties with describing abuse, courts will allow prosecutors to ask option-posing and even overtly leading questions (United States v. Grassrope, 2003).

Option-posing questions can reduce the likelihood that children will provide “don’t know” and other uninformative responses, compared to wh- questions. This is because yes-no and forced choice questions tap recognition memory, which facilitates retrieval, whereas wh- questions require the child to generate the to-be-remembered information (Ornstein et al., 1997). Furthermore, yes-no questions are more likely than recall questions to elicit information that children are reluctant to disclose, such as transgressions (Lyon et al., 2014), and this may be attributable to children’s aversion to lying when directly asked. For example, examining 5- and 7-year-old girls’ reports of a pediatrician visit, Saywitz et al. (1991) found that yes-no questions elicited true reports of vaginal and anal touch that were omitted from children’s free recall. Although older children’s free recall was superior to younger children’s, they were no more likely to mention vaginal and anal touch without direct prompting. This suggests that embarrassment rather than memory failure partially explained omitted reports of genital touch in free recall.

Conversely, because of the ease with which children can answer recognition questions (by simply choosing a yes-no response or choosing a proffered choice), they are more likely to guess, which increases errors (Waterman et al., 2004). Saywitz et al. (1991) found that recognition questions increased the likelihood that children falsely reported genital touch, and subsequent researchers have emphasized the dangers of option-posing questions in questioning children about touch (Poole & Bruck, 2012). As children’s apparently informative responses to recognition questions may be mere guesses, it is misleading to uniformly label them “informative.” Therefore, we focus in this paper on clearly uninformative answers, such as “don’t know” and off-topic responses. Another problem with option-posing questions is that they are likely to elicit fewer child-generated details than wh- questions. Children exhibit formal reticence, whereby they provide minimally sufficient responses based on the form of the question (Stolzenberg & Lyon, 2017). Yes-no questions tend to elicit unelaborated “yes” and “no” responses, and forced choice questions elicit unelaborated choices (Stolzenberg & Lyon, 2014; 2020), especially from younger children (Verkampt et al., 2019). In contrast, because they require the child to generate the to-be-remembered information, wh- questions elicit child-generated details unless the child fails to respond or answers “I don’t know.”

Although Burrows et al. (2017) recognized the importance of analyzing question type in assessing children’s ability to describe their sexual body parts, they failed to distinguish between recognition questions (yes-no and forced-choice) and recall questions (wh-). Instead, they distinguished among questions they classified as “open-ended,” “specific,” or “leading.” They limited their definition of “open-ended” to questions that “encouraged an elaborate response from the child and did not specify the information the child was required to report” (p. 4). These are normally referred to as “invitations” in the literature on child interviewing, and include only the broadest of recall questions, such as “What happened next?” and “Tell me more about child-generated content” (Lamb et al., 2018). In Burrows and colleagues’ scheme, “specific” questions appeared to include both recall and recognition questions. Only 5% of questions qualified as “open-ended,” and another 8% were classified as “leading,” and the small numbers made it impossible to assess the informativeness and productivity of different question-types. Furthermore, the researchers could not determine whether differences in children’s responses to questions asking them to either name body parts or to describe their function or location might be confounded by question type effects.

Children’s Descriptions of Sexual Touch

Children’s difficulties may extend to answering questions about other aspects of sexual abuse besides identification of sexual body parts. Of related concern is their ability to describe the location of touch and the method or manner of touch. More specifically, children may struggle with attorneys’ efforts to determine whether there was skin-to-skin contact, penetration, and pain (which supports the inference that penetration occurred; (Rhode Island v. Mensah, 227 A.3d 474, 2020, United States v. Espinoza, 2009)). Although a large body of experimental research has examined the utility of diagrams and dolls in eliciting reports of non-abusive genital touch (e.g., Bruck et al., 2016; Saywitz et al., 1991), and a similarly large body of observational research has examined children’s ability and willingness to disclose abusive touch (e.g., Azzopardi et al., 2019), surprisingly little research has examined how children disclosing sexual abuse describe touch beyond simply affirming or denying that sexual touch occurred (Dupree et al., 2016). In their study examining miscommunications in child sexual abuse trials, Sullivan et al. (2021) noted that “how” questions about the manner of touch were often a source of confusion, though they did not compare “how” questions to other question-types. Similarly, one group of researchers has attempted to study girls’ understanding of penetration by asking girls whether they wiped “inside” after urinating (Gallion et al., 2016; Milam & Nugent, 2017), but they did so using forced choice questions, potentially confounding younger children’s immature knowledge of their genitals with their difficulty answering option-posing questions.

The Present Study

This study investigated attorneys’ questions and 5- to 10-year-old children’s responses regarding sexual body parts (specifically, the vaginal area and the anus) and sexual touch in child sexual abuse trials. With respect to the clarity of references to sexual body parts, we hypothesized that 1) both children and attorneys would frequently use unclear colloquial terms for sexual body parts; 2) the use of clear anatomical terms would increase with children’s age; and 3) questions asking about the names of sexual body parts would elicit more uninformative responses than questions about the function and location of body parts. Additional hypotheses pertained to sexual body part knowledge, the location of touch, the method or manner of touching, skin-to-skin contact, penetration, and how the touching felt. For these topics, we hypothesized that 4) attorneys would rely on option-posing questions to ask about body part knowledge and sexual touch, and that 5) wh- questions would elicit more uninformative responses than option-posing questions, but that 6) wh- questions would elicit more child-generated information than option-posing questions Finally, with respect to age differences, we hypothesized that younger children would give more uninformative responses, and older children more child-generated information. In addition to the quantitative analyses, we also reviewed questions and responses and identified examples of the difficulties attorneys encountered in eliciting details of sexual touch.

Method

Sample

Pursuant to the California Public Records Act (Cal. Government Code 6250, 2016), we obtained information on all felony sexual abuse charges under Section 288 of the California Penal code (sexual abuse of a child under 14 years of age) filed in Los Angeles County from January 2, 1997 to November 20, 2001 (N = 3,622). Of all the cases filed in this time period, 9% went to trial (n = 309). We were able to obtain trial transcripts for 235 of the 309 cases, which included virtually all of the acquittals and mistrials (95% or 53/56) and 71% (182/253) of the convictions. Trial transcripts of acquittals and non-appealed convictions were obtained from court reporters and trial transcripts of convictions were obtained from the appellate court. All cases involving children (88% female) aged between 5 and 10 years (M = 8.53) were selected for the present study (n = 113). All defendants in the sample for the present study were charged with one or more counts of California Penal Code Section 288a, 288b (sexual abuse of a child under 14 years of age using force), or 288.5 (continuous sexual abuse of a child). A small number of defendants (n = 3) were also charged with aggravated sexual assault of a child (Section 269). To illustrate age differences descriptively, children were divided into groups of 5- to 8-year-olds (n = 53) and 9- to 10-year-olds (n = 60). The use of archived interviews for research purposes was approved as exempt by the University of Southern California Institutional Review Board.

Coding

Machine coding was used to identify question-answer pairs referring to genitals (e.g., “butt,” “private,” “peepee”), verbs describing potentially abusive touch (e.g., “put,” “rub,” “touch”), and prepositions that may refer to the invasiveness of touch (e.g., “inside,” “outside”). Each question-answer pair was then coded by a human coder to ascertain that it referred to the vaginal or anal area. All questions had been question-type coded for prior research, distinguishing between option-posing questions (yes-no and forced choice) and wh- questions. Because they were virtually non-existent (n = 12), questions that could be classified as “invitations” (Lamb et al., 2018) were classified as wh- questions. Furthermore, “Do you know” questions with an embedded wh-word (e.g., “Do you know where your butt is?”) were coded as wh- questions, as they requested information beyond an unelaborated “yes” or “no.”

Question Topics

For all relevant lines (N = 2,247), questions were classified as asking about 1) sexual body part knowledge; 2) the location of sexual touch; 3) the method or manner of sexual touch (e.g., “what did he touch you with?”); 4) whether there was skin-to-skin contact; 5) whether penetration occurred; or 6) the child’s feelings during abuse. Sexual body part knowledge questions were subcategorized according to whether they asked about the name of the body part (naming questions; e.g., Q: “What do you call that part?”), the location of the body part (location questions; e.g., Q: “Is that the area between your legs?”), or the excretory function of the body part (function questions; e.g., Q: “Is that where Number two comes from?”). Sexual body part knowledge questions about location and questions about the location of sexual touch were subcategorized into requests for verbal responses and requests to point on the child’s body, a doll or a diagram.

Clarity of References to Sexual Body Parts

We categorized the terms used by attorneys and children as clear if they clearly referred to the vaginal area (e.g., “vagina,” “peepee”) or anal area (e.g. “butt,” “bottom”), or unclear, if it was unclear whether they referred to the vaginal or anal area (e.g. “private,” “down there”).

Children’s Responses

Researchers coded whether the response was uninformative or informative, and also whether the child generated information in the response. Uninformative responses included no response, don’t know or don’t remember responses, requests for clarification, off-topic responses, and underinformative responses (e.g., Q: “Do you know where your private parts are?” A: “Yes.”; Evans et al., 2017). Informative responses provided an on-topic answer. Responses were coded as including child-generated information when the child spontaneously provided details, as opposed to giving an unelaborated “yes” or “no” to a yes-no question or an unelaborated choice in response to a forced choice question. Children’s responses to sexual body part knowledge questions were also coded to indicate whether their response provided more specific information about which body part was involved. Children’s answers to clarifying questions did not increase specificity when their answers left it vague which body part was involved (e.g., Q: “When you say your private, where is that?” A: “Down there.”), and when the body part was already specified to the extent that a layperson would understand what the child was referring to before the clarification (e.g., Q: “Butt means your behind?” A: “Yes.”). In order to assess whether penetration was alleged, attorneys’ questions and children’s responses were also coded according to whether they referred to touching as occurring “in” or “inside.”

Inter-Rater Reliability

A random selection of 20% of the question-answer pairs across 20% of transcripts were independently re-coded. Cohen’s κ was used to assess agreement between coders. Agreement was high regarding whether question-answer pairs referred to sexual body parts (K = .97), the clarity of terms for genitalia used by attorneys (K = .93) and children (K = .88), question topics (K = .86), and whether children’s responses were uninformative (K = .87).

Analysis Plan

Analyses were conducted using generalized linear mixed effect models (GLMMs). GLMMs combine the properties of linear mixed models (which incorporate random effects) and generalized linear models (which handle non-normal data) and are preferable to traditional analysis of variance (ANOVA) models because they have fewer assumptions, handle binary response variables, and maximize power while simultaneously estimating between-subject variance (Bates et al., 2015). A by-subject random effect was included to control for differences in the number and type of questions in each interview and the individual response characteristics of each child. Fixed effects included 1) question-type (wh-, option-posing) and 2) age (continuous). Binary dependent variables included 1) response informativeness and 2) child-generated information. For the sexual body part questions, we also included a fixed effect for question subcategory (function, location, naming), and the binary dependent variables also included 1) unclear terminology and 2) increased specificity.

We first calculated descriptive statistics regarding the percentage of questions about different topics, the percentage of children’s spontaneous references, the clarity of references to sexual body parts, and body part question subcategories. For each question topic, we calculated the proportion of option-posing questions, the rate of uninformative responses and the rate of child-generated information. Analyses were performed using the glmer function in the R package lme4 with the bobyqa optimizer (Bates et al., 2015). The anova function in the R stats package (R Core Team, 2013) was used to test whether interactions between fixed effects improved the model fit. Pairwise comparisons with Tukey HSD and estimated marginal means were computed using the emmeans function in the R package emmeans (Lenth, 2020). Age was added to the models as a continuous variable, but means are reported for a younger (5–8 years) and older (9–10 years) age group to illustrate comparisons. The results of the best-fit models are reported alongside the unstandardized fixed effects estimates (B), standard errors of the estimates (SE) and estimates of statistical significance (Z and p values). The results of interactions are reported only if the best fit model included an interaction term.

Results

Of 2,247 question-answer pairs, we identified 312 questions asking about sexual body part knowledge (14%), 542 questions about the location of sexual touch (24%), 568 questions about the method or manner of sexual touch (25%), 200 questions about skin-to-skin-contact (9%), 238 questions about penetration (11%), and 150 questions about feelings during abuse (7%). We also identified 237 spontaneous descriptions of sexual touch provided in response to questions that did not specifically ask about this topic (10%). Table 1 presents the percentages of uninformative and child-generated responses to wh- and option-posing questions on different topics.

Table 1.

Percentage of uninformative and child-generated responses to different question topics, comparing wh- and option-posing questions.

	Uninformative responses		Child-generated responses
	Wh-	Option-posing	Wh-	Option-posing
Sexual body part knowledge (n = 312)	27%	24%	72%*	2%
Location of touch (n = 478)	12%	9%	87%*	17%
Method and manner (n = 568)	14%	12%	87%*	16%
Skin-to-skin contact (n = 200)	N/A	15%	N/A	8%
Penetration (n = 238)	23%1	9%	74%*	5%
Feelings (n = 150)	24%*1	4%	75%*	18%

Notes:

^*denotes a significant difference between wh- and option posing questions at p < .05. The percentages are the estimated marginal means (adjusted for children’s age and sexual body part question subtype).

¹The main effect of question-type was qualified by an interaction between question-type and age.

Clarity of References to Sexual Body Parts

When attorneys referred to children’s sexual body parts (n = 944), they usually relied on unclear terms (59%) such as “private(s)” (42%). There was no significant difference in the proportion of unclear terms used by attorneys when talking to older and younger children. Children named their sexual body parts either in spontaneous descriptions (n = 207) or in response to questions about body part knowledge or the nature of touch (n = 311). Unclear colloquial terms dominated in both children’s spontaneous descriptions (70%) and names elicited by attorneys’ questions (58%). Analyses examined the use of unclear terms in relation to spontaneity and children’s age. Children were more likely to use unclear terms in spontaneous mentions than in response to attorneys’ questions, B = 1.07, SE = .36, Z = 2.98, p = .003. Younger children were non-significantly more likely to use unclear terms (44%) than older children (32%).

Sexual Body Part Knowledge

Attorneys’ questions about children’s knowledge of sexual body parts (n = 312) focused on naming (44%) and location (37%), with function questions used less frequently (19%). Among location questions, requests for verbal descriptions were most common (42%), followed by requests to point on the child’s own body (32%), a diagram (21%) or a doll (6%). Because the transcripts did not allow us to determine whether children’s non-verbal responses were uninformative, questions requesting that the child indicate their sexual body parts by pointing on their own body, a diagram or doll (n = 66) were excluded from further analyses (leaving a final sample of n = 246).

Sexual Body Part Knowledge: Factors Related to Uninformative Responses

Overall, 57% of questions about sexual body part knowledge were option-posing. Thirty-four percent of children’s responses were uninformative. We examined the frequency of uninformative responses in relation to body part subcategory (function, location, naming), question type (wh-, option-posing) and age (continuous). Naming questions were more likely to elicit uninformative responses (43%) than function questions (15%; B = 1.42, SE = .44, Z = 3.21, p = .001). There was no difference in the rate of uninformative responses to location questions (23%) compared to function and naming questions. Wh- questions were no more likely to elicit uninformative responses (27%) than option-posing questions (24%). Younger children were more likely to give uninformative responses (34%) than older children (33%; B = .32, SE = .14, Z = 2.29, p = .02).

Sexual Body Part Knowledge: Factors Related to Child-Generated Information

About a third (33%) of children’s responses to questions about sexual body part knowledge included child-generated information. Examining the effects of body part question subcategory, question type, and age, we found a non-significantly higher rate of child-generated responses for function questions (27%) than terminology (18%) and location (10%) questions. Child-generated information was more common in response to wh- questions (72%) than option-posing questions (2%, B = 5.03, SE = 1.09, Z = 4.60, p < .001). Older children were non-significantly more likely to provide child-generated information (37%) about sexual body part knowledge than younger children (28%).

Sexual Body Part Knowledge: Factors Related to Specificity of Responses

Thirty-one percent of questions about sexual body part knowledge elicited responses that increased the specificity of body part identifications. Examining the effects of body part question subcategory, question type, and age, we found that function questions were significantly more likely to increase the specificity of body part descriptions (40%) than location questions (21%, B = .94, SE = .47, Z = 1.99, p = .04). There was no difference in the rates of specificity-increasing responses to terminology questions (27%) compared to function and location questions. Wh- questions (31%) and option-posing questions (27%) increased the specificity of body part identifications at similar rates. Similarly, there was no difference in rates of specificity-increasing responses to sexual body part questions provided by older (31%) and younger children (31%).

Location of Touch: Factors Related to Uninformative Responses

Overall, 52% of questions about the location of genital touch (n = 542) were option-posing. Thirteen percent of children’s responses were uninformative. The majority of questions about location of touch asked the child to verbally describe where they were touched (88%), with questions asking the child to point on their own body (6%), a diagram (5%) or a doll (1%) used less frequently. As noted above, questions asking for non-verbal responses could not be coded for uninformativeness (n = 64) and were excluded from subsequent analyses. Analyses examined the frequency of uninformative responses to location of touch questions (n = 478) in relation to question type and age. Wh- questions and option-posing questions elicited comparable percentages of uninformative responses (12% and 9%, respectively). Younger children were more likely to give uninformative responses to questions about location of touch (19%) than older children (8%; B = .31, SE = .12, Z = 2.53, p = .01).

Location of Touch: Factors Related to Child-Generated Information

Half (50%) of responses to questions about the location of touch contained child-generated information. Wh- questions elicited significantly more child-generated information (87%) than option-posing questions (17%; B = 3.50, SE = .33, Z = 10.52, p < .001). Older children provided significantly more child-generated information (55%) than younger children (45%; B = .24, SE = .11, Z = 2.15, p = .03).

Method or Manner of Touch: Factors Related to Uninformative Responses

Overall, 47% of questions about the method or manner of touch (n = 568) were option-posing. Fifteen percent of children’s responses were uninformative. Wh- questions about the method or manner of touch were no more likely to elicit uninformative responses (14%) than option-posing questions (12%). Younger children provided uninformative responses at a marginally significant higher rate (20%) than older children (11%; B = .21, SE = .11, Z = 1.92, p = .05).

Method or Manner of Touch: Factors Related to Child-Generated Information

About half (53%) of responses to questions about the method or manner of touch contained child-generated information. Wh- questions elicited a significantly higher proportion of child-generated information (87%) than option-posing questions (16%; B = 3.50, SE = .30, Z = 11.69, p < .001). The difference in the rate of child-generated information provided by older children (57%) and younger children (46%) in response to questions about the method or manner of touch was not significant.

Skin-to-Skin Contact: Factors Related to Uninformative Responses

Almost all questions about skin-to-skin contact (n = 200) were option-posing (99%) and were often phrased as asking whether the child was touched “over” or “under” clothing. Fifteen percent of responses were uninformative. Analyses examined the frequency of uninformative responses to skin-to-skin contact questions in relation to children’s age. Due to the small number of wh- skin-to-skin contact questions (n = 3), only option-posing questions were included (n = 197). Older children were just as likely to give uninformative responses (15%) as younger children (15%). Only 8% of responses contained child-generated information.

Penetration

The number of questions asking about penetration (n = 238) was somewhat surprising given that 96% of children testified in trials where the only charges were filed under California Penal Code Section 288 (sexual abuse of a child under 14), which does not require proving penetration. Among the three cases that involved aggravated sexual abuse (Cal. Penal Code Section 269), one was an attempted rape charge, and one was an oral copulation charge, leaving only a single case where proving penetration was legally required. In a small number of cases (n = 8), primarily involving younger children (M = 7.25 years), attorneys checked children’s understanding of prepositions, including “on top,” “inside” and “underneath,” before asking children about penetration. Children affirmed penetration in response to 67% of questions about penetration. In the other third of responses, children either denied penetration (17%) or provided unclear responses about whether penetration occurred or not (16%). Although frequencies were too low for statistical analyses, younger children were more likely to provide unclear responses (22%) than older children (13%).

Penetration: Factors Related to Uninformative Responses

Overall, 90% of questions about penetration were option-posing. Fourteen percent of children’s responses were uninformative. Although there were no significant effects due to question type or age, there was a significant interaction between question type and children’s age (B = 1.25, SE = .61, Z = 2.06, p = .04). The uninformativeness of wh- questions about penetration increased with age, whereas the uninformativeness of option-posing questions did not vary by age. Because the total number of wh- questions about penetration was very small (n = 23), the decreased productivity of wh- questions with age might be attributable to a series of incomprehensible questions posed to a 10-year-old in a single case (e.g., Q: “How does his private inside your private?” A: “Yes.” Ramirez, 1998 [Italicized names refer to individual cases in the data]).

Penetration: Factors Related to Child-Generated Information

Thirteen percent of responses to questions about penetration contained child-generated information. Wh- questions elicited a higher rate of child-generated information (74%) than option-posing questions (5%; B = 3.90, SE = .75, Z = 5.19, p < .001). The rate of child-generated information in response to questions about penetration did not significantly differ between younger (18%) and older children (10%).

Feelings Questions: Factors Related to Uninformative Responses

Overall, 60% of questions about feelings were option-posing (n = 150). Thirteen percent of children’s responses were uninformative. There were significant effects for both question type (B = 16.31, SE = 5.82, Z = 2.81, p = .005) and age (B = .66, SE = .31, Z = 2.11, p = .04), which were qualified by an interaction between question type and age (B = 1.65, SE = .63, Z = 2.64, p = .008). Whereas the rate of uninformative responses in response to option-posing questions about feelings varied little with age, wh- questions elicited higher rates of uninformative responses from younger children (36%) than older children (16%) (Figure 1).

Figure 1.

The interaction between children’s age and question type on the rate of uninformative responses to questions about feelings.

Feelings Questions: Factors Related to Child-Generated Information

Thirty-nine percent of responses to questions about feelings contained child-generated information. Wh- questions elicited more child-generated information (75%) than option-posing questions (18%; B = 2.66, SE = .42, Z = 6.29, p < .001). The difference in the rate of child-generated information provided by older children (44%) and younger children (32%) in response to questions about feelings was marginally significant (B = .37, SE = .19, Z = 1.96, p = .05).

Qualitative Analysis of Attorneys’ Questions and Children’s Responses

We identified two difficulties that expand on our quantitative results: 1) attorneys’ attempts to elicit information through highly specific option-posing questions about the genitalia and the anus often appeared to result in acquiescent responses that were subject to impeachment by defense attorneys, and 2) conversely, attempts to specify sexual body parts with wh- questions often failed to elicit sufficiently specific responses from children.

Highly Specific Option-Posing Questions About the Genitalia and Anus

Attorneys’ attempts to elicit descriptions of specific parts of the vagina or anus often led to unelaborated affirmations that failed to demonstrate children’s understanding of the question. For example, attorneys attempted to specify where the “hole” is inside the vagina in a case involving a 9-year-old female (Q: “Is that the part where the hole – where the pee comes out?” A: “Yes.” Acinelli, 2000) and in the anus in a case involving an 8-year-old boy (Q: “When you say go in your behind, is there a hole area in your behind?” A: “Yes.” Johnson, 1998). Attorneys also asked questions about the “inside” and “outside” of genitals, potentially to prepare children for questions about penetration. One prosecutor asked a 7-year-old girl about her awareness of the “inside” of her vagina (Q: “Do you know that you have an outside and an inside to your body?” A: “Uhhuh.” Bruce, 2000), and another prosecutor asked an 8-year-old boy about his concept of “up inside” the anal area (Q: “Okay. Now in connection with the red X that you made on the little boy’s butt that is a private area would you say?” A: “Yes.” Q: “Okay! And of course up inside of the little boy’s butt that would be an even more private area. Is that right?” A: “Yes.” Jewsbury, 1998). Prosecutors sometimes asked highly complex questions about penetration with yes-no questions, such as in a trial involving a 6-year-old girl (Q: “And you understand that little girls, like all women - okay - where their pee-pee comes out there is skin to either side, right?” A: “Yes.” Q: “Now, I want you to tell the court the touching that Jose did, was it on top of, did it go in between those two pieces of skin?” A: “Yes.” Camarena, 1999).

Highly specific option-posing questions about penetration might have elicited the desired answers, but they left children open to defense challenges, as illustrated by the cross-examination of an 8-year-old girl (Q: “Okay. When you were asked some questions about if he put his weenie in you, what did that mean to you, putting his weenie in you?” A: “What do you mean?” Q: “Have you ever seen a picture of what a girl looks like down there?” A: “No.” Q: “Do you know how many holes or openings you have down there?” A: (no answer) Q: “So when – when – when this attorney or this man over here was asking you if you felt something inside of you, did you know what that meant?” A: “No.” Marin, 1998). Defense attorneys also challenged the credibility of children’s testimony on the basis of apparent contradictions about penetration, such as in a trial involving a 7-year-old girl (Q: “Do you remember if you told the policemen that he put it on your butt?” A: “Yes.” Q: “Not in your butt?” A: “In.” Q: “So you told the policemen that he put it on your butt; is that right?” A: “No.” Lopez, 2000). Defense attorneys highlighted contradictory statements about genital pain, as in a trial involving a 9-year-old girl (Q: “Let me be clear. Your privates weren’t hurting?” A: “Yes.” Q: “Your privates were hurting too?” A: “No.” Q: “But you told the police – you told the police your privates were hurting, didn’t you?” A: “No.” Brown, 1998).

Attempts to Specify Sexual Body Parts with Wh- Questions

Children sometimes struggled with providing clear details at the level of specificity attorneys requested in response to wh- questions. Younger children often provided vague responses to sexual body part knowledge questions (Q: “And where does the pee-pee come from?” A: “Underneath the clothes.” Farmer, 1998), but older children’s responses were not always specific either (Q: “When you say front part, what part of the body are you talking about?” A: “The front.” Whitehurst, 2000). Although an 8-year-old female answered a body part function question informatively, her response failed to distinguish between the vagina and the anus (Q: “Do you do anything with that part of your body?” A: “Go to the bathroom.” Durado, 1998). Some children, including an 8-year-old female, expressed reluctance to answer questions about body part knowledge (Q: “What do you call that part of your body?” A: “I don’t really talk about it.” Olvera, 1999).

Discussion

This study examined references to sexual body parts and sexual touch in attorneys’ questions and 5- to 10-year-old children’s responses in child sexual abuse trials. A major challenge for attorneys is to elicit clear descriptions of sexual body parts, given children’s immaturity and embarrassment. We found that most of attorneys’ and children’s spontaneous references to sexual body parts were unclear, and this did not vary by the age of the child. Attorneys’ attempts to inquire into children’s sexual body part knowledge led to uninformative answers about a third of the time, and children were particularly likely to be uninformative when asked for alternative names. Questions about function were less likely to fail than questions about naming body parts, and more likely to increase the specificity of body part identifications than questions about location. However, attorneys asked function questions the least often.

We also compared the tendency for option-posing (yes-no and forced choice) and wh- questions to elicit uninformative responses and child-generated information, examining questions about sexual body part knowledge, the location of touch, the method or manner of touching, whether there was skin-to-skin contact, whether penetration occurred, and how the touching felt. Across question-types, 47–99% of questions were option-posing, and 13–15% of children’s responses were uninformative. Although we predicted that wh- questions would elicit more uninformative responses than option-posing questions, this only occurred for questions about feelings, and only for younger children. Hence, we found little evidence that wh- questions were more uninformative than option-posing questions.

Conversely, there was clear support for our prediction that children would produce more child-generated information in response to wh- questions. This was because children’s responses to option-posing questions were usually unelaborated. Yes-no and forced choice questions not only limit children’s productivity (Stolzenberg & Lyon, 2017), but also compromise the accuracy of their responses (Waterman et al., 2004). Our qualitative analysis revealed that highly specific option-posing questions about the genitalia and anus frequently elicited unelaborated affirmations. These unthinking and potentially inaccurate responses were often subject to impeachment by defense attorneys.

Younger children were often more likely to give uninformative responses and less likely to generate information than older children, consistent with general developmental trends. However, older children often struggled too. For example, older children (9- to 10-year-olds) gave uninformative responses in response to 33% of questions about sexual body part knowledge. Furthermore, they were no more likely than younger children to provide specific information when asked additional questions about sexual body parts.

Our results expand on Burrows and colleagues’ (2017) conclusions regarding questions about sexual body part knowledge. We replicated their findings that children’s spontaneous references to sexual body parts are often unclear, that clarity fails to increase from 5 to 10 years of age, that questions asking children to name their body parts are often uninformative, and that questions about the function of sexual body parts are more productive. However, Burrows and colleagues’ (2017) were unable to examine the relation between question type and uninformative responses or child-generated information, because they failed to distinguish between option-posing questions and wh- questions. Our results revealed the superiority of wh- questions over option-posing questions in eliciting child-generated information about sexual body parts. Moreover, we extended this finding to other types of questions about sexual touch.

Implications for Practice

Our results highlight the challenges of eliciting descriptions of sexual body parts and sexual touch from 5- to 10-year-old children. Interviewers should anticipate that children’s initial disclosures will fail to provide clear descriptions of their sexual body parts. In attempting to elicit unambiguous responses, they should include questions about the excretory functions of the sexual body parts, as these are often superior to questions about alternative names or locations.

Attorneys’ preference for option-posing questions in our sample may reflect their belief that option-posing questions overcome children’s difficulty in talking about sexual body parts and sexual touch. However, our results revealed no consistent advantage of option-posing questions over wh- questions. Our qualitative results suggest that persistent use of option-posing questions can elicit acquiescent affirmations in young children, subjecting them to impeachment. Furthermore, our findings reinforce the well-established advantages of wh- questions in eliciting elaborative, child-generated details, in contrast with the predominantly one-word responses obtained by option-posing questions (Lyon & Henderson, 2021). These advantages have been demonstrated even when children undergo the stressful procedure of testifying in open court (Andrews et al., 2016). Consequently, we suggest that using more wh- questions in court would increase child-generated information about sexual body parts and abusive touch without eliciting more uninformative responses.

Limitations and Future Research

We were unable to assess the accuracy of children’s descriptions of sexual body parts and sexual touch. Future work examining children’s descriptions of genital and anal touch in a medical setting could expand on existing findings by shifting focus from whether children disclose being touched to how children describe their experiences of touch (Dupree et al., 2016). We were also unable to determine whether and how often children’s non-verbal responses were uninformative or lacked specificity. Although potentially suggestive when used as screening tools (e.g., Bruck et al., 2016), non-verbal aids such as human figure diagrams may facilitate the recall of additional details about previously disclosed touch (Aldridge et al., 2004; Teoh et al., 2010). Future research with access to the diagrams children used and video recordings of pointing could assess whether sexual body parts and the location of touch may be clarified through non-verbal means.

Future research should assess the degree of specificity needed for questions to elicit clear descriptions of sexual touch. On the one hand, perhaps additional requests for free recall ultimately would lead to clear descriptions. Practitioners are now routinely advised to maximize their use of broad open-ended content-free requests for recall, also known as invitations (Lamb et al., 2018), but this type of question was virtually non-existent in our sample of court transcripts. Similarly, Burrows et al. (2017) found very few invitations in the forensic interviews they examined. Therefore, we cannot say whether a series of invitations may be able to elicit clear descriptions of sexual body parts or sexual touch.

There are reasons to worry, however, that given children’s reluctance and their difficulty in discerning the intent of the questioner, specific wh- questions may be necessary. As noted above, Sullivan et al. (2021) found that miscommunications often occurred in court testimony when children failed to understand that attorneys’ broad questions were asking for specific details about sexual touching. Generally speaking, invitations often elicit higher rates of uninformative responses than other question types, leading Wolfman et al. (2016) to warn that “[i]nvitations may be challenging for children because they do not specify what kind of information the child should include in their response” (p. 114). Conversely, wh- questions can elicit information that invitations often overlook, such as children’s emotional reactions to abuse (Stolzenberg et al., 2021).

We found evidence that even wh- questions are often insufficiently specific. Only about a third of wh- questions about sexual body parts elicited responses that increased the specificity of body part identifications. Our qualitative assessment identified examples in which children’s responses to quite specific wh- questions were informative, but did not clearly identify the body part that was touched. For example, the children who answered “where the pee-pee comes from” with “underneath the clothes” or “what do you do with that part” with “go to the bathroom” either failed to understand what the attorney needed to know, or were reluctant to provide the needed body part. Interviewers and attorneys may need to be still more specific, and may need to include supportive statements that emphasize the importance of the information and reassure children about the propriety of explicit description.

We observed that questions about penetration were frequent despite the rarity of cases where proving penetration was legally required. Research using a sample of trials with more severe charges could explore whether potential miscommunications and impeachment on the topic of penetration may affect trial outcomes. Furthermore, future research should also investigate potential differences between prosecutors’ and defense attorneys’ questions about sexual body parts and sexual touch, as defense attorneys may specifically avoid asking wh- questions to control the scope of the testimony and minimize child-generated information (Pozner & Dodd, 2004). Indeed, defense attorneys may decide to avoid the topic altogether, instead focusing on peripheral details of the child’s report (Szojka et al., 2017). Furthermore, field research should examine the role of child characteristics that may influence knowledge of sexual body parts and sexuality in general, such as culture and native language (Kenny & Wurtele, 2008), and socio-economic background (Gordon et al., 1990).

Conclusion

Attorneys and children predominantly referred to sexual body parts using unclear colloquialisms that do not distinguish between the vaginal and anal area. Questions about the function of sexual body parts elicited the fewest uninformative responses and were most likely to increase the specificity of body part identifications. Attorneys primarily asked option-posing questions about sexual body part knowledge, the location of touch, the method or manner of touch, skin-to-skin contact, penetration, and how touching felt for the child. However, in most question categories, wh- questions performed better than option posing questions, increasing the likelihood of child-generated responses without increasing uninformative responses. Coupled with prior research on the advantages of wh- prompts in court (Andrews et al., 2016), our results suggest that attorneys should ask more wh- questions when questioning children about sexual body parts and abusive touch.