Dynamics in using different question types in Estonian

Dynamics in using different question types in Estonian police interviews of children Kristjan Kask, Ph. D University Nord, Estonia

Question types n Open memory prompts tend to elicit longer and more accurate responses than do closed prompts (Hershkowitz, Lamb, Sternberg, & Esplin, 1997; Lamb & Fauchier, 2001; Lamb, Hershkowitz, Sternberg, Esplin, et al. , 1996; Orbach & Lamb, 1999). n Focused and closed questioning has been found to be dominant in forensic interviews of children in Israel, Sweden, Finland, the United States, and the UK (Cederborg, Orbach, Sternberg, & Lamb, 2000; Craig, Scheibe, Kircher, Raskin, & Dodd, 1999; Davies, Westcott, & Horan, 2000; Lamb, Hershkowitz, Sternberg, Boat, & Everson, 1996; Sternberg et al. , 1996; Sternberg, Lamb, Davies, & Westcott, 2001; Stockdale, 1996; Walker & Hunt, 1998; Santtila, Korkela, & Häkkänen, 2004; Korkman, Santtila, & Sandnabba, 2006).

Children’s free recall n n Young children’s free recall is accurate if they are interviewed appropriately (Flin, Boon, Knox, & Bull, 1992; Marin et al. , 1979). When children are allowed to recall information freely or through the use of general questions, even very young children can produce material which is highly accurate (Hutcheson, Baxter, Telfer, & Warden, 1995).

Use of questions n n In forensic contexts, free recall prompts such as general or cued invitations produce three to five times more information than do focused prompts (Lamb, Hershkowitz, Sternberg, Esplin, et al. , 1996; Sternberg, Lamb, Davies, & Westcott, 2001). Option-posing, ´yes/no´, and suggestive interviewer utterances are more likely to elicit inaccurate information because they implicitly encourage children to recall the suggested information or to guess (Dent & Stephenson, 1979; Poole & White, 1993; Ceci & Bruck, 1995; Poole & Lindsay, 1998).

Present study n The purpose of the present study was to examine the questioning styles employed in Estonian police interviews of child sexual and physical abuse cases, focusing q q (i) on the proportions of questions and number of words in children’s answers to the questions, and (ii) the dynamics in using questions during the interviews (see also Griffiths Question Map, Griffiths & Milne, 2006).

Method Participants n Videotaped interviews with 66 children (23 boys and 43 girls), mean age was 9 years 6 months (range 5 to 13 years). n The interviews had been conducted during 2004/5 and 2007/8, mean length 27 minutes (range 8 to 82 minutes). n Sixty three (95%) of the interviewers were female police officers and three (5%) were male.

Method n n No information was available concerning the officers’ experience, the outcome of the interviews, or what proportion led to trial or conviction. In general, Estonian police officers have had training about interviewing suspects and witnesses during their general studies, and specific trainings in interviewing children (such as the introduction to the Cognitive Interview or NICHD protocol) but these occur very infrequently and are not necessarily part of involving new investigators’ into the teams.

Coding n n Videotapes of the interviews were transcribed. The categories by Lamb and colleagues (Lamb, Hershkowitz, Sternberg, Esplin et al. , 1996; Lamb, Hershkowitz, Sternberg, Boat, & Everson, 1996) and by Yuille and Cutshall (1986) were used to characterize the interviewer utterances and to measure the amount of new information provided.

Coding n n n Explanations Invitations: q General q Cued Directive utterances Option-posing utterances Suggestive utterances Verbal affirmations / facilitators

Coding n n The number of words in the children’s responses was counted in all interviews in response on interviewers’ questions. Each interview was divided into four equal quarters regardless of how many questions were asked to investigate the proportionate usage of different question types across time within interviews.

Results n n The interviewers on average asked more than one question per utterance M = 1. 37 (0. 60) range from 1 to 9. The average number of questions asked per interview was M = 246 (164), range from 44 to 841) and the average number of words reported by children was M = 823 (671), range from 63 to 2641.

Total number of words in children’s responses n The total number of words in children’s responses increased with age F(1, 2) = 5. 39, p <. 01, η² =. 15 M = 442. 4 (553. 8) 4 to 7 -year-olds M = 969. 71 (609. 9) 8 to 11 -year-olds M = 996. 1 (721. 1) 12 to 14 -year olds

Question types n Option-posing (32%) and direct questions (32%) were used mostly by the interviewers followed by verbal affirmations (21. 6%), explanations (7. 6%), suggestive questions (4. 1%), cued invitations (1. 6%), and general invitations (1. 1%).

The number of words in response to different categories of interviewer utterance (F (1, 6) = 87. 46, p <. 001, η² =. 049) Category Mean SD Explanation 4. 37 5. 47 General invitation 10. 16 10. 61 Cued invitation 8. 16 8. 30 Verbal affirmation 4. 42 5. 98 Direct questions 5. 91 6. 05 Option-posing questions 4. 00 5. 05 Suggestive questions 3. 32 4. 72 Mean 5. 07 6. 03

Average number of words in children’s responses n The average number of words in children’s responses increased with age F(1, 2) = 219. 50, p <. 001, η² =. 041. M = 2. 69 (3. 28) 4 to 7 -year-olds M = 5. 22 (6. 13) 8 to 11 -year-olds M = 6. 41 (6. 78) 12 to 14 -year olds

The proportion of questions in different quarters of interviews (number of questions) First quarter Second quarter Third quarter Forth quarter Chi-sq. Explanation 76 54 35 42 p <. 001 General invitation 167 59 17 3 p <. 001 Cued invitation 71 63 63 26 p <. 001 Verbal affirmation 253 295 297 298 ns Direct questions 861 793 737 648 p <. 001 Option-posing questions 518 651 752 867 p <. 001 Suggestive questions 28 56 72 89 p <. 001 Category

Conclusion n The continuing failure to encourage a free narrative (or to try and maintain it once the child had begun) with many of the children is a concern, since this allows children to present their ´story` in their own words (see Westcott & Kynan, 2006). One possible explanation for the overall large number of option-posing questions may be that officers are not aware of the negative effect of such questions. Police officers may feel pressured by legal demands associated with investigative interviews on the one hand the psychological well-being of the child during the questioning on the other hand.

Conclusion n n Better police training in the adoption of bestpractice guidelines in interviewing children is required. Methods such as the NICHD protocol or cognitive interview are strongly recommended to structure the interviews.

Thank you!