Australian Institute of Criminology

Skip to content

Methodology

Participation in the DUMA program is both voluntary and confidential. Each quarter, trained interviewers work onsite during peak hours to conduct interviews with detainees. Access is facilitated by the police officer in charge of the watchhouse, or their delegate, who determines a detainee’s initial eligibility. Detainees are not interviewed if they are:

  • deemed unfit for interview due to alcohol/drug/medication;
  • mentally unfit;
  • children and juveniles (except for New South Wales);
  • alleged offenders who would require an interpreter;
  • considered to be potentially violent;
  • held in custody in excess of 48 hours; or
  • deemed ineligible for other reasons at the discretion of the custody manager.

If eligible, a detainee is approached by a police officer and asked if they are willing to participate in the DUMA study. Using a prepared introductory statement, the police officer tells the detainee that:

There is a researcher here who is doing some research. Would you be willing to speak to him/her for a few minutes? The researcher is completely independent from the police and anything you said would be treated in strict confidence.

At this point, a detainee may choose not to participate and will be returned to their cell. Their decision not to participate has no impact on their case or processing by the police.

If the detainee agrees to be interviewed, they are escorted from their cell to an interview room where they will have their first contact with a DUMA interviewer. Upon greeting the detainee, the interviewer advises that the research project is government funded and that participation is confidential and voluntary. The detainee is then asked to give informed consent to undertake a structured interview, as well as provide a urine sample (if relevant to the collection period). They are reminded of the confidential nature of the research and assured that none of their responses can be linked back to them in any way. Names and addresses are not kept and the data are not linked at a later stage to any official data such as their criminal record. The participant may elect not to answer questions and they may refuse to provide a urine sample if one is requested. As at December 2012, 5,044 detainees had refused to be interviewed, 48,251 had agreed to be interviewed and 35,822 had provided a urine sample (6,935 of those who agreed to answer a questionnaire refused to provide a urine sample and 4,483 did not provide a urine sample for other reasons such as being unable to produce a specimen).

If the detainee indicates to the police officer or the interviewer that they do not wish to participate, the reason for their refusal is recorded on a separate interview form (known as a refusal form) and the detainee is returned to their cell. Again, their decision not to participate has no impact on their criminal case or subsequent processing by the police.

If the detainee agrees to be interviewed, ‘informed consent’ procedures are undertaken. On several occasions during the survey, detainees are reminded of the confidential and voluntary nature of the survey, as well as their right to end the survey at any time.

At the completion of the interview, if relevant to the collection period, detainees are reminded about the collection of the urine sample. They are again asked whether they consent to the provision of the sample. Detainees who refuse to provide the sample are then read the following statement:

Your participation is completely voluntary, but I would like to remind you that no names will appear on the specimen and the results will not be given to the police or affect the outcome of your case. An independent laboratory will perform the analysis and the sample will be destroyed as soon as the tests have been done. There is no way that the results can be tied back to you. Would you agree to provide a sample?

If a detainee refuses to provide a sample after the second prompt, the interviewer thanks them for their time and ends the interview. The detainee is then escorted back to their cell. A detainee who agrees to provide a urine sample is given a urine collection bottle and escorted to an appropriate location within the watchhouse. The sample is then returned to the interviewer, the interview ends and the detainee is escorted back to their cell. Urine samples are then given a unique barcode, refrigerated and later sent to an authorised testing laboratory in New South Wales. The laboratory returns test results to the AIC in an electronic form. At no point during the survey process are the police or local data collectors informed of the individual test results. Similarly, the AIC receives no personally identifying information about the detainees and so interview forms or urinalysis results cannot be re-identified.

At the completion of each interview and for each detainee who refuses, interviewers complete the front page of the survey using information from the police charge book. Protocols for collecting charge information vary from jurisdiction to jurisdiction, although in no situation is the name of the detainee recorded on the interview form. Completed interview forms are locked in a secure cabinet until the end of the four week collection period, at which time they are couriered to the AIC using registered mail.

Once the interview forms are received by the AIC, they are coded by DUMA staff in accordance with a range of coding frameworks, including the Australian and New Zealand Standard Offence Classification (ANZSOC). The coded surveys are then entered into a database by an external data entry contractor. Data are returned to the AIC for cleaning and analysis. Questionnaires and samples are matched by their barcodes at the AIC. No records of names are kept and all urine samples are destroyed once the AIC receives and validates the results.

Quality control processes

Before each data collection period, interviewers undergo training in the questionnaire and operational procedures specific to their site. Interviewer error reports are an important part of the quality-control processes employed in the program. In the first stage of this process, site coordinators audit each questionnaire and any errors that are identified are then reported back to interviewers. In the second level of quality assurance, the questionnaires are audited by the AIC. Error reports are then compiled by the AIC and distributed to each site manager in time for the next quarter. Errors that occur frequently are:

  • nil responses being recorded on particular questions—where an interviewer fails to record a response to a mandatory question;
  • non-recognition of internal skip patterns—where an interviewer incorrectly follows a specified skip pattern, leaving some mandatory questions unanswered; and
  • incorrect coding.

Error rates are generally higher when an interviewer is new to the program or when an interviewer has been with the program for some time and becomes complacent. However, by conducting interviewer training at the beginning of every quarter, the AIC is able to keep the overall error rate within an acceptable range.

The AIC also monitors the level of urine compliance at both the site and interviewer level. This internal monitoring allows for the identification of emerging issues and provides an opportunity to address individual or site-based problems if and when they arise.

A teleconference is also held at regular intervals with members of the AIC’s DUMA team and site coordinators and managers. The teleconference
is a forum in which issues related to the administration of the questionnaire or addendum can be discussed in depth.

Most serious offence

The Australian Bureau of Statistics’ ANZSOC (formerly known as the Australian Standard Offence Classification) scheme is used to assign charges to eight categories—violent offences, property offences, drug offences, drink driving, traffic offences, disorder offences, breaches and other lesser offences (ABS 2011). Since 2009, the ANZSOC codes have been adopted and all historical codes have been updated to reflect these new codes. DUMA detainees are assigned to the most serious of the charges collected. The hierarchy from most serious to least serious is:

  • violent offences;
  • property offences;
  • drug offences;
  • drink driving;
  • traffic offences;
  • disorder offences;
  • breaches; and
  • other lesser offences.

Therefore, according to this classificatory scheme, if a detainee interviewed for the DUMA program has been charged with a violent offence and a property offence, the violent offence takes precedence.

Response rates

Appendix A provides information on the fieldwork dates for quarterly data collection. This includes information on the periods during which fieldwork was undertaken, the number of hours interviewers were in the police station or watchhouse, the number of detainees approached and interviewed and the number of urine samples collected at each site.

As shown, data collection at Parramatta and Kings Cross alternated each quarter. Kings Cross operated during the first and third quarters of 2011 and 2012, and Parramatta operated during the second and fourth quarters.

In 2011–12, a total of 7,716 detainees were interviewed, of whom 7,586 were defined as adults in their relevant jurisdiction and 130 were juvenile detainees from the three New South Wales sites. Detainees can choose to complete the interview without providing a urine sample. During a 2011–12 review of the DUMA program, the urine collection schedule was altered. Urine samples are now collected from detainees every second quarter, rather than every quarter. In 2012, the new collection cycle commenced. Urine was collected in the first and third quarter for the sites of Adelaide, Brisbane, East Perth, Darwin, Footscray and Southport. In New South Wales, Bankstown urine collection was conducted in the first and fourth quarter, while urine was collected once in 2012 for Kings Cross (first quarter) and Parramatta (fourth quarter). In 2011–12, of those who agreed to interview in a urine collection quarter, 75 percent also provided a urine sample. The rate of urinalysis compliance in 2012 (73%) was four percentage points lower than in 2011 (77%). The collection rate achieved in 2011–12 is consistent with that recorded in previous years.

Table 96 sets out the response rates for adult detainees who agreed to an interview. The data shows that there are no substantial differences by gender and that response rates are generally consistent across sites, with the exception of East Perth, where male detainees were nine percentage points more likely than females to agree to be interviewed (65% cf 56%). However, there are differences between sites in the provision of urine samples.

Urinalysis

Across the sites of Adelaide, Bankstown, Darwin, Footscray and Kings Cross, female detainees were less likely to provide a urine sample, while at Parramatta, male detainees were less likely to provide a urine sample. In the remaining sites, male and female detainees were almost as likely to agree to be interviewed. At the Bankstown and Parramatta sites, juveniles voluntarily provided a urine sample at rates similar to adult detainees, while at Kings Cross, juvenile detainees were more likely to provide a sample than adults.

In Darwin, the rate of urinalysis compliance continues to be the lowest among all the sites (50% of all detainees interviewed in urine collection quarters). At this site, there are a greater proportion of Indigenous detainees, and culture and beliefs may affect their willingness to provide urine samples. However, while there is also a large Indigenous population at East Perth, urinalysis compliance is comparable to other sites (75% of all detainees interviewed in urine collection quarters). In Indigenous culture, there are clear divisions between men’s and women’s roles (Maher 1999). It may be the case that interviewers of the opposite sex who requested samples from Indigenous detainees inadvertently breached cultural rules or norms. The introduction of same-sex interviewers has seen an increase in compliance rates, so where possible, this practice will be adopted as a minimum standard. Language may have also been a prohibiting factor because English is not the first language for many Indigenous detainees, particularly in Darwin.

Sorcery is also prominent in traditional Indigenous life and culture (Maher 1999; McGrath & Phillips 2008). Indigenous cultural beliefs about the body and bodily fluids, and in particular how such things might be misused in sorcery, may have played a role in the lower than average urine compliance rates among Indigenous detainees.

These concerns are not unique to the DUMA program—health professionals often experience difficulties in providing care to Indigenous people. As Maher (1999) suggests, this may be due to the cultural distance between mainstream Australian culture and Indigenous culture (see also McGrath & Phillips 2008). To help overcome some of these barriers, the DUMA site manager in Darwin developed additional information for use in negotiating the informed consent of Indigenous detainees that has helped to substantially increase compliance.

These issues notwithstanding, the response rates obtained in DUMA are higher than those normally achieved in social science research in Australia. For example, the response rate for the interview (68.7%) is higher than that achieved in the Australian National Drug Strategy Household Survey (50.6%; AIHW 2011).

Table 96 Response rate by sex and adult status, 2011–12
Adelaide Bankstown Brisbanea East Perth Parramatta Southporta Darwin Footscray Kings Cross
Adult males
Approached (n) 1,575 792 1,307 2,172 346 1,043 1039 570 387
Agreed to interview (n) 841 518 1,170 1,408 215 941 677 414 259
Agreed to interview (%) 53 65 90 65 62 90 65 73 67
Agreed to interview during urine collection quarters (n) 628 402 865 1,042 159 687 495 318 209
Provide urine specimen (n) 393 302 846 774 115 653 258 177 147
Provided urine (of those who agreed to interview during urine collection quarters; %)b 63 75 98 74 72 95 52 56 70
Adult females
Approached (n) 268 119 209 521 82 170 132 137 97
Agreed to interview (n) 143 80 178 294 48 149 84 101 66
Agreed to interview (%) 53 67 85 56 59 88 64 74 68
Agreed to interview during urine collection quarters (n) 105 61 135 214 30 102 67 76 56
Provide urine specimen (n) 58 43 133 162 26 98 22 32 34
Provided urine (of those who agreed to interview during urine collection quarters; %)b 55 70 99 76 87 96 33 42 61
Juveniles
Approached (n) 0 141 0 0 104 0 0 0 15
Agreed to interview (n) 0 60 0 0 59 0 0 0 8
Agreed to interview (%) 0 43 0 0 57 0 0 0 53
Agreed to interview during urine collection quarters (n) 0 53 0 0 47 0 0 0 7
Provide urine specimen (n) 0 33 0 0 33 0 0 0 7
Provided urine (of those who agreed to interview during urine collection quarters; %)b 0 62 0 0 70 0 0 0 100

a: Detainees aged 17 years were traditionally surveyed in Queensland sites but have been removed, so the number reflects only adult detainees in Queensland

b: Percentage has been calculated for the quarters in which urine samples were requested, which in 2011 was all 4 quarters and in 2012 was 2 out of 4 quarters

Source: AIC DUMA collection 2012 [computer file]

DUMA sample

It is important to note that although the sites are referred to by the name of the area in which they are located, the catchment area may not necessarily reflect the city boundaries. Because of this, the estimated size of the catchment area varies between the nine DUMA sites. Further, state legislation governs length of detention, reason for detention and the procedures for detention. These will then vary between the DUMA sites.

In regard to the randomness of the DUMA sample, none of the sites have 24 hour coverage and interviewers enter the sites at times when the number of detainees is expected to be at a maximum. During these periods, all eligible detainees are asked to participate in the study. One criterion is that a person has not been held in custody for more than 48 hours (63 detainees were excluded on this basis). Some detainees are also deemed by local police staff to be ineligible for interview. This is usually due an assessment of probable risk to the interviewer. In 2011–12, 933 detainees were deemed by the police to have been unfit for interview, representing eight percent of the potential sample. The number also varied by site; for example, 23 percent of detainees in Adelaide were declared unfit to interview. Across the other sites, the percentages ranged from three percent in Southport and Darwin to 10 percent in Parramatta. As a consequence, the sample obtained by DUMA is not a random sample of all people detained by the police.

Two other factors affect the randomness of the sample. First, in all six jurisdictions, the police use a variety of mechanisms through which they can reduce the number of people brought into the station for processing. These include diversion programs, notices to attend court (or equivalent) and cautions. Normally, these notices or cautions would be for minor offences. Diversion programs tend to focus on drug possession cases and cases involving juvenile offenders. As a result, the DUMA study generally does not survey these people.

Second, the study is anonymous, so it is not possible for individuals to be tracked across the interview periods. Given that a substantial number of detainees report that they have been arrested in the previous 12 months, it is highly likely that a small group of detainees will appear twice or more within or across quarterly collection periods. Strictly speaking, the sample is one of ‘episodes of detention’ rather than ‘individual detainees’. Detainees are asked at the end of the interview if they can recall participating in the study on a previous occasion. In 2011–12, 930 detainees confirmed having previously participated in the DUMA study (which represents 13% of the sample) and a further 30 detainees said they could not recall.

Drug testing

Research has documented the shortcomings of relying solely on self-report data when reporting on drug use (see Makkai 1999). Some of the issues affecting self-report data include the ability of the respondent to accurately recall events (especially drug use over defined periods of time) and a respondent’s willingness to share information of a sensitive nature with interviewers. These shortcomings are likely to result in the underreporting of particular behaviours, including drug use and participation in illegal activities. In order to enhance the veracity of self-report information obtained from police detainees and as a cross-validation measure, the DUMA program conducts urinalysis on the urine samples voluntarily provided by police detainees. Urine testing is the most cost-effective means of objectively measuring the presence of illicit drugs. It is also a scientifically valid measure of drug use within the known limits of the test.

Urinalysis screening was conducted for five drug classes—amphetamines, benzodiazepine, cannabis, cocaine and opiates—and secondary screening tests were conducted for the opiate pharmacotherapy substances methadone and buprenorphine. A positive result is recorded when the drug or its metabolites are detected at or above the cut-off levels set in accordance with Australian Standards (prescribed at AS/NZS 4308). If a positive result is obtained for opiates, amphetamines and/or benzodiazepines, a further set of tests using confirmatory gas chromatography-mass spectrometry (GC/MS) is performed to ascertain which specific drugs are present in the urine.

The urinalysis results indicate whether the drug was consumed shortly before detention, with the exception of cannabis and benzodiazepines. For these two drugs, a positive test indicates prior use of up to 30 days for cannabis and 14 days for benzodiazepines. Table 97 indicates the average detection times and the cut-off levels for a positive screen.

Table 97 Cut-off levels and drug detection times
Drug class Cut off AS 4308 (ug/L) Average detection timea
Amphetamines 300 2–4 days
Benzodiazepines (hydrolysed) 100 2–14 days
Cannabis 50 Up to 30 days for heavy use; 2–10 days for casual use
Cocaine 300 24-36 hours
Methadone 300 2–4 days
Opiates 300 2–3 days
Buprenorphine 5 2–7 days

a: Depends on testing method and equipment, the presence of other drugs, level of drug present and frequency of use

Source: Makkai 2000

For urinalysis results, there are five important points to note:

  • the screen detects the class of drug, not the specific metabolite;
  • false positives and false negatives can occur, although cut-off levels are designed to minimise their frequency;
  • detection times can vary depending on the individual person and specific rates of metabolism and excretion;
  • a positive result does not necessarily imply illicit use; and
  • the presence of the drug does not necessarily mean the person was intoxicated or impaired.

In 2006, further testing was carried out on buprenorphine results as a cross-checking mechanism. Results from these tests indicated a high level of reliability (over 80%). For more information see Mouzos et al. (2007).

All drug testing for the program is conducted at one laboratory—Pacific Laboratory Medical Services, Northern Sydney Area Health Service—in Sydney. The laboratory is accredited to the Australian Standard AS/NZS 4308: 2008.

Table 98 shows the percentage of detainees who tested positive for heroin, methamphetamine or cocaine use and also reported drug use in the previous 48 hours and previous 30 days. These data are consistent with other studies; there is a higher level of underreporting for recent use (past 48 hours) than for use in the past 30 days. Approximately half of those who tested positive to heroin, methamphetamine and cocaine reported that they had used the substance in the previous 48 hours. For the previous 30 days, self-reporting decreased to approximately two-thirds of those who test positive for heroin, methamphetamine and cocaine. From 2011 to 2012, the level of discrepancy between reported use and urine results has remained relatively consistent for heroin. In 2012, the level of discrepancy for cocaine decreased by 13 percentage points while for methamphetamine and heroin, the level of discrepancy stayed consistent over the years.

There are a number of reasons that a police detainee may not accurately report their recent drug use, despite their urine positively indicating recent use. The most obvious reason is that some detainees can be reluctant to report drug use around the time of arrest. As DUMA is primarily concerned with measuring drug use at the time of arrest, the importance of urine testing cannot be underestimated. If drug policy is to be underpinned by evidence, the evidence needs to be as reliable and valid as possible. If data are biased, program development and implementation could be harmful to both individuals and the broader community.

Table 98 Comparing urinalysis and self-reported drug use (%)a
Heroin Methamphetamine Cocaine
Positive urinalysis result Negative urinalysis result Positive urinalysis result Negative urinalysis result Positive urinalysis result Negative urinalysis result
Self-reported use past 48 hours 53 2 47 2 42 <1
Self-reported use past 30 days 71 4 75 12 71 3
Total (n) 443 3,825 975 3,289 59 4,209

a: Results for 2011–12

Source: AIC DUMA collection 2012 [computer file]

Explaining compliance levels

Relative to other social science studies, compliance levels for both the interview and providing a urine sample are high. Several factors may account for this. First, the measures taken to assure confidentiality include a signed statement from the director of the AIC. The statement is important in negotiating the informed consent of detainees. Second, the clearly established independence of a well-trained interview team is integral to the program. It is a requirement that no current or former police officers from that jurisdiction be hired as interviewers and all interviewers are required to undergo training before entering the site. This training is compulsory regardless of whether the interviewer has participated in prior rounds of data collection. Third, detainees are assured that their information will only be disseminated in aggregated form, that their names are not recorded and that the urine sample they provide will be destroyed once the AIC has validated the results.

The AIC Human Research Ethics Committee first cleared this project in January 1999 for a three year pilot study. In December 2001, clearance was granted for the project to continue and in November 2003, ethics clearance was given for the extension of the program. Ethics clearance for the further extension of the DUMA study to Darwin and Footscray was obtained in December 2005 and in June 2007 to the new site of Alice Springs. An additional ethics clearance was given for the extension of the program in November 2010. Each separate addendum administered as part of the questionnaire is also cleared by the AIC’s Human Research Ethics Committee.

Oversight committees

Selected sites have their own local steering or advisory committee, which is engaged as needed to discuss key issues about the DUMA methodology and operation. The committee’s role is to support the local data collectors, monitor the local progress of the study, suggest ways of improving the project, undertake appropriate analyses of their own site data and ensure dissemination of information at a local level to relevant agencies. All of the committees comprise a cross-section of people, including representatives from local law enforcement and researchers.

An important aspect of DUMA is the dissemination of questionnaire and urinalysis results. This involves sending quarterly results from the urinalysis to the sites as soon as practicable after their receipt at the AIC. This provides timely intelligence to inform local policy and strategic initiatives. In addition, local sites are provided with anonymised unit record files for secondary analysis by request. This ensures that those in law enforcement who are tasked with tackling local crime issues are equipped with the most up-to-date DUMA data for their area, to enable them to address problems.

Uses of DUMA data

DUMA provides an important platform for in-depth research in the criminal justice field. A number of additional studies have been launched at the local sites to capture additional data for specific policy purposes. These have included the development of addenda on stolen goods, drug driving, prescription drug use, synthetic cannabis and amphetamines. DUMA provides a unique platform from which to collect data to assist in evidence-based policymaking and to inform strategic intelligence. DUMA also has the potential to assist in the evaluation of public health interventions in the longer term. Trends and issues highlighted in the DUMA data can be used to inform policy and program development, complementing and enhancing the approaches taken by law enforcement agencies. It also serves to provide insight into some areas where information has not previously been available.

DUMA data can be used at a variety of levels and for a variety of purposes. Data can be used to argue for policy shifts, for internal resources, to determine the effectiveness of particular interventions or police operations at the various sites, or for monitoring purposes. However, the data are also useful at a state and federal government level. Because data are collected, audited and documented under the same set of protocols, greater confidence can be placed on their comparability, validity and reliability—helping to inform policymaking in areas such as housing, treatment, mental health, policing, courts and correctional institutions. DUMA data are also increasingly being used in reports produced by other agencies. Links to published material can be found at the AIC’s website http://www.aic.gov.au.

Examples of agencies and organisations that have requested/used data include:

  • state and territory police services;
  • Australian Government Attorney-General’s Department;
  • Australian Customs Service;
  • Australian Crime Commission;
  • Crime and Misconduct Commission, Queensland;
  • South Australian Office of Crime Statistics and Research;
  • Department of Health and Ageing;
  • Drugs and Alcohol Services, South Australia;
  • Drugs and Alcohol Office of Western Australia;
  • Australian Institute of Health and Welfare;
  • Turning Point Alcohol and Drug Centre;
  • National Drug Research Unit, Curtin University of Technology;
  • Edith Cowan University;
  • Flinders University;
  • Griffith University;
  • United Nations Office on Drugs and Crime;
  • Alcohol and Other Drugs Council of Australia;
  • National Motor Vehicle Theft Reduction Council;
  • National Drugs and Alcohol Research Centre, University of NSW;
  • Australian National University; and
  • Newfoundland and Labrador Centre of Health Information, St Johns, NL.