RESEARCH ARTICLES

 

Public order transparency: Using freedom of information laws to analyse the policing of protest

 

 

David Bruce

Independent researcher specialising in policing and criminal justice. davidbjhb@gmail.com

 

 

---

ABSTRACT

This article discusses two research projects that have used the Promotion of Access to Information Act (PAIA) to analyse protest in South Africa and the policing thereof. A total of 23 information requests were submitted on behalf of the two projects, 19 to the South African Police Service (SAPS) and four to the Independent Police Investigative Directorate. The article starts by discussing police transparency in South Africa, information on the policing of protest that the SAPS routinely publishes in its annual report, the PAIA framework, and some of the limitations of the projects. It then focuses on insights into SAPS information on levels of protest and protest-related violence in South Africa that emerged from the two projects. This includes information disclosed by the police regarding their use of force during protests, and police accountability for this. The article concludes by reflecting on the implications of and lessons from these exercises in police transparency.

---

 

 

In contemporary democracies police are increasingly exposed to public scrutiny. One reason for this is the proliferation of technologies such as closed-circuit television (CCTV) and cell phone cameras. In recent years, CCTV and cell phone videos, exposing apparently unjustified police actions, have often circulated on social and traditional media, notably in the United States (US), but increasingly in South Africa too. Accountability legislation and oversight architecture, and laws promoting public access to state information also increasingly compel police to share information. Police could embrace such transparency as a means to build trust and cooperation with civilians and civil society groups;¹ however, they do not always welcome scrutiny and may be resistant to transparency. This possibly reflects what Joshua Chanin and Salvador Espinosa call a 'preference for reticence' motivated by '[m]istrust of the media and scepticism about the motivations of the information-seeking public'.²

Related to the fact that protest is often linked to political conflict and social divisions in society, the policing of protest is a source of controversy in many countries. As a result, police may be wary about opening themselves to scrutiny around it.³ In South Africa, protest is a key issue of public concern and debate.⁴ Public Order Police (POP) are the key component of the South African Police Service (SAPS) responsible for dealing with protest, especially where there is violence or the risk thereof. Scrutiny of POP in South Africa has intensified in recent years in the wake of a series of deaths during protests and strikes, most notably at Marikana in August 2012, where police shot and killed 34 striking miners and wounded 76 others.⁵ Protests on university campuses in September and October 2016 involved widespread disruption of teaching programmes and some incidents of violence by protestors. Police action on university campuses in response to this protest also became a source of heated contention, particularly in relation to some incidents in which excessive force was allegedly used.⁶ In brief, when and how public order policing is practiced in South Africa remains contentious.

This article focuses on the intersection between police transparency and the policing of protest in South Africa. Researchers have started using freedom of information (FOI) laws to access SAPS information on crowd incidents and public order policing. It discusses two projects that have used the Promotion of Access to Information Act (PAIA) to obtain information on protest and its policing. It focuses on key insights gleaned from the data regarding protest numbers and the use of force by police during protests. The conclusion reflects on SAPS compliance with PAIA, the unsatisfactory quality of some of the information that has been disclosed, continued official reliance on this information, and the implications of the two research projects discussed in this article for how the SAPS understands transparency.

 

Police transparency in South Africa

Evaluations of an organisation's transparency often focus on how it responds to requests for information lodged under FOI laws. For instance, reports by a civil society network in 2013 and 2014 indicated that the SAPS had responded positively to less than 50% of requests that were submitted to it.⁷ But assessing degrees of transparency is not only about an organisation's willingness to disclose information but also about whether the information is reliable, accurate and up to date. A 2007 assessment of the SAPS in relation to 'indicators of democratic policing' noted that the standard of reporting in the SAPS's annual reports was relatively good, compared to that of many other government departments.⁸However, the auditor-general has consistently raised questions about the reliability of information on the SAPS's performance against set indicators provided in its annual reports.⁹The system through which the SAPS releases crime statistics has also frequently attracted criticism. This is partly because when crime statistics are released, they are already six months out of date.¹⁰

Information routinely provided on public order policing

The SAPS's annual reports consistently include information on the number of POP units, the number of members of these units, the procurement of public order equipment, public order training, and arrests during crowd incidents. In addition, the section on POP consistently provides data on two categories of 'crowd related incidents' distinguished as either 'peaceful' or 'unrest'.¹¹ The data comes from a database of incidents to which POP units have responded, known as the Incident Reporting Information System (IRIS). Table 1 provides IRIS data on these two categories of incidents from April 2011 to March 2016.

Many people assume 'public order policing' means the policing of protest, and interpret the data on crowd incidents as data on protests in South Africa (see Alexander et al. in this issue of SACQ).¹² This misreading is reinforced by the limited supplementary data in SAPS reports on such incidents. For instance, the 2015/16 SAPS annual report refers to 'peaceful incidents' as including 'assemblies, gatherings and meetings'. 'Unrest-related incidents' are said to include 'labour disputes, including the mining sector, dissatisfaction with service delivery by local municipalities, demarcation of municipality borders, but also in the transport and education sectors'.¹⁴ This does not explain whether the crowd incidents recorded were all protests. A further persistent source of confusion is the tendency of both the public and senior police to interpret the statistics on 'unrest' crowd incidents as statistics for violent protest.¹⁵ Some have suggested that police deliberately misrepresent IRIS data, and that they do so to present an inflated picture of levels of violent protest, thereby potentially feeding into a situation where the state is seen as justified in stifling dissent.¹⁶

 

The Promotion of Access to Information Act

PAIA gives effect to the right of public access to information provided for in Section 32 of the constitution. The act obliges a 'public body' (the act also has provisions regarding private bodies) to provide access to 'records' that it holds if these are requested. This is subject to specified grounds for refusal and procedural requirements.¹⁷ The act therefore provides access to 'records' that are already held in one form or another, rather than imposing a general obligation to provide information. The act also provides that public bodies must publish a manual to inform members of the public about how to submit PAIA requests to that body.¹⁸Both the SAPS and the Independent Police Investigative Directorate (IPID) publish this information on their websites.¹⁹

PAIA distinguishes between records that may be requested and those classified as 'automatically available'.²⁰ For instance, according to the SAPS PAIA manual, 'National Instructions' are automatically available.²¹ This means that requests for them do not have to be evaluated in terms of PAIA criteria and they can simply be provided to the requestor.

 

Information requests on the policing of protest

In 2014 and 2015, the Freedom of Information Programme at the South African History Archive (SAHA) submitted 23 information requests to the SAPS and IPID on behalf of the two research projects discussed in this article. These requests resulted in the release of 95 records. The records that have been released are available online on the SAHA 'PAIA Tracker'.²²

Social Change Research Unit, University of Johannesburg

Three of the PAIA information requests were submitted on behalf of the Social Change Research Unit (SCRU) at the University of Johannesburg. The first of these was submitted in March 2014. The unit has researched protest in South Africa since 2009 and the requests were submitted in line with this interest.²³ These three requests sought information from IRIS and resulted in 43 records being released. The most notable of these requests was submitted in October 2014 and resulted in the release of 34 records containing IRIS data in nine categories, including all incidents in the crowd management 'peaceful' and 'unrest' categories, spanning 17 calendar years from 1997 to 2013. The SCRU's analysis of this data represents a significant breakthrough in understanding protest in South Africa (see Alexander et al. in this issue of SACQ).

Open Society Justice Initiative

Twenty other PAIA requests were submitted to the SAPS and IPID in terms of an international comparative 'transparency audit' under the auspices of the Open Society Justice Initiative (OSJI), a programme of the Open Society Foundation.²⁴ The audit was carried out simultaneously by locally based researchers in Brazil, India, Mexico, South Africa and the United Kingdom.²⁵ Questions were agreed on through consultation between researchers and were organised around the themes of preparedness of police forces, policing in practice, and the oversight and accountability of police forces in respect of the policing of gatherings and protest.²⁶

Of the 20 requests submitted in South Africa, 16 were to the SAPS and four to the IPID. Requests submitted to the SAPS resulted in the release of 47 records. Those to the IPID resulted in the release of five records.

Other PAIA requests on the policing of protest

In addition to these requests, the OSJI project also submitted five PAIA requests to five municipalities. These were concerned, inter alia, with 'conditions' that people who are planning protests are supposed to comply with. The response to these requests is not discussed here. PAIA requests linked to the policing and regulation of protest have also been submitted by other parties.²⁷

 

Information not available or that was refused

Records that were released, and which may be of interest to people interested in protest and the policing of protest, include the national instructions on the policing of gatherings, National Instruction 4 of 2014 (classified as 'automatically available'), training materials, SAPS organograms for the Operational Response Services division, and others.²⁸However, not all of the requests for information were successful.

In terms of Section 23 of the act, where a record cannot be found or does not exist, an affidavit must be provided to this effect. In response to a request for records of disciplinary action for police misconduct connected to gatherings, the SAPS provided an affidavit from an SAPS official stating that he was not aware of any disciplinary action being taken against any officers in his unit and that he did not know of any cases being opened or lodged against any member of his unit.²⁹ Some of the issues raised by this response are examined in the discussion of IPID data on complaints and disciplinary cases relating to protest, below. The SAPS previously proved unwilling to respond to information requests regarding disciplinary action against police involved in the Marikana incident.³⁰

The SAPS did not respond to a query for information on shooting incident investigations relating to protests.³¹ In addition to the provisions of Section 23, Section 25(3) of PAIA provides that, in the event that requests are refused, reasons must be provided. However, no affidavits or reasons were provided for not responding to this query.

 

Other limitations of SAPS responses

The SAPS only released some of its crowd management training manuals after an 'internal appeal' process provided for in the act.³² Apart from this, all the records that were released, were released spontaneously. The SAPS ultimately provided one or more records in response to 16 of the 19 requests, including all three of the SCRU and 13 of the 16 OSJI requests. However, the records that were given did not necessarily contain the sought-after information. Requests for records about the number of people injured in protests, or about complaints received by the SAPS in relation to the policing of protest, were not answered in a meaningful way.³³ A request for information on the budget for public order policing was similarly unsuccessful. This may in part be explained by the formalistic and bureaucratic nature of the PAIA process. Information officers at public bodies are legally obliged to comply with the request for information within the framework of PAIA, but this does not extend to an obligation to clarify requests that are not clearly formulated. It is possible that in these cases the information request did not clearly set out the needed information.

Another limitation of the process was that some information appeared to be incomplete. For instance, a 2011 policy document on public order policing indicates that training programmes for POP include courses not only for platoon members but also for platoon commanders, first line operational managers and operational commanders.³⁴ But, in response to the request for training materials, the SAPS only provided training manuals for the course for platoon members as well as materials on crowd management from the basic training curriculum. No explanation was given why manuals were provided for some, but not all, SAPS crowd management courses.

 

IRIS information provided by the SAPS

The SAPS released a number of records from the IRIS system in response to requests from the University of Johannesburg's SCRU. Of the records provided in response to the OSJI requests, 15 related to the IRIS system. In total 57 of the 92 records released by the SAPS were therefore related to the IRIS system.

The most important set of documents released by the SAPS were the 34 records of IRIS data on 'crowd management peaceful' and 'crowd management unrest' that were released to the SCRU. The records cover 156 230 incidents over the years 1997 to 2013.³⁵ Analysing this large volume of information presented a major challenge to the unit. A team of four research assistants was assigned the task of interpreting and coding a stratified random sample of the incidents.³⁶ Two reports based on the analysis of these records have been published.³⁷

The reason these records are important is that each incident record includes detailed 'notes'. These are narrative entries by POP officials. The SCRU's 2016 analysis is based on the narratives in these 'notes' for a stratified random sample of 4 520 incidents over the 17-year period. This work shows that the only way to systematically distinguish protest incidents from other incidents on IRIS is through analysis of these notes. There is no other data field on IRIS that can be used for this purpose. For example, IRIS contains a field for number of arrests, but it is not possible to provide the number of people arrested during protests without isolating protest cases from non-protest incidents through a painstaking analysis of the notes section on a case-by-case basis - as the SCRU has done.

Through its analysis the SCRU has been able to make findings on the number of protests reflected in the IRIS data for 'peaceful' and 'unrest' crowd incidents. Overall figures emerging from the SCRU analysis for the 1997-2013 period are provided in Table 2. The SCRU report indicates that there are substantial variations from one year to another in the percentage of crowd incidents that are protests.

The SCRU analysis also engages with questions about the focus of protests (but not other 'crowd incidents') during the 1997-2013 period. The analysis indicates that 46% were 'labour protests' and 22% 'community related', with the balance falling into seven other categories.³⁹The SCRU research demonstrates that the classification of an incident as 'unrest' does not necessarily mean that it involved violence. It also highlights the need for official systems for recording protests to differentiate incidents that involve injury to persons or damage to property, from those that are disorderly in other ways.⁴⁰It is not clear if IRIS can be modified to address these issues, or if an entirely new system must be developed.

 

Monitoring the use of force during protest

One of the key questions regarding the policing of protest in South Africa is about the justifiability of police use of force. One OSJI request asked for information on the 'three gatherings in relation to which most arrests were made' over the period April 2012 to March 2014.⁴¹ In response to this request, the SAPS provided eight documents containing data from the IRIS system on three incidents - allegedly the protest incidents during this period at which the largest number of arrests were made. The incidents were at Woodstock station in Cape Town in June 2013 (184 arrests), at the Waterworks informal settlement near Randfontein in Gauteng in September 2013 (176 arrests), and in Aliwal North in the Eastern Cape in January 2014 (163 arrests).

The eight documents are of interest partly because they seem to provide all the information recorded on IRIS on these three incidents. The information is broader than the nine categories of information in the records released to the SCRU. These documents indicate that there are about 36 categories of information recorded on IRIS. These include, for instance, information about the notification process (if there is prior notification),⁴² the route, organisations involved, any weapons used by participants, the 'security force' units involved, and ammunition used by them. Many of the 36 categories include a number of subcategories. An interesting aspect of these documents is the information on the use of force, including the weapons and ammunition used by the police. There is no evidence of police use of force in the Woodstock incident, but some information on the use of force is provided in the documents dealing with the Waterworks and Aliwal North incidents. Notably, in the incident at Waterworks, during which residents of the informal settlement used burning tyres to barricade a turnoff on the N12 highway, IRIS documents indicate that police used close to 1 100 rubber rounds, five teargas canisters and 10 stun grenades. However, in the category 'person injured as a result of police action', no information is recorded.

One of the press reports dealing with the Waterworks incident records that a woman was shot in the shoulder with a rubber bullet.⁴³There appears to be no information in the public domain suggesting that other people were injured during the protest. Nevertheless, considering the number of rubber bullets used, it is likely that more people were injured. In incidents where rubber bullets are used, it is possible that force may be used relatively indiscriminately.⁴⁴ It is also likely that those injured by rubber bullets would flee from police rather than wait to have their wounds documented and risk arrest. As a result, it may not be possible for police to comprehensively record injuries. Nevertheless, one would at least expect the police to acknowledge the likelihood that some people were injured, even if they indicate that the exact number is unknown.

The Aliwal North documents are, at best, vague on the use of force by police. They include no detailed information about use of force, although there are indications that rubber bullets and one stun grenade were used. They also indicate that seven civilians were injured and taken to hospital, although no explanation is provided on how they were injured.

These documents therefore add to the concerns raised by the SCRU about the quality of information recorded on IRIS.⁴⁵ In particular, they raise questions about whether IRIS is a reliable record of the use of force by police during protests. It would seem that IRIS reports may downplay injuries resulting from police use of force. Concerns that IRIS data may not be comprehensive were also raised in a 2007 report that discussed IRIS data on people killed by police during demonstrations.⁴⁶

Information held by the IPID

The IPID also provided two documents, with overlapping information, on complaints about the policing of protests. The documents provide information on 68 complaints (11 in 2014 and 57 in 2015) by members of the public against the police. These relate to 52 incidents of protest. One anomaly in the documents is that, although they are supposed to provide data on complaints relating to gatherings during 2014 and 2015, they contain no cases from Gauteng, the province with the greatest population and which accounts for more protests than any other province.⁴⁷ They also contain only four cases from KwaZulu-Natal, South Africa's second most populous province. This suggests that the information comes from a system that does not classify and record complaints in a reliable manner.

The documents also suggest that there are extremely few, if any, cases where complaints against police working at protests result in a finding against individual police members. Among the 69 cases there appears to be only one where the IPID concluded that there was wrongdoing on the part of the police. In this case the report contains no indication that the police responsible were identified, or faced disciplinary action. Many other cases are closed because the police officer allegedly responsible for wrongdoing cannot be identified. In others the case is closed by the IPID because the person laying the complaint is unable to prove that they were not part of a group who were protesting violently. In effect, complainants, some of whom claim to have been bystanders who were hit by rubber bullets while in the vicinity of protests, carry the burden of proving that they were not part of a violent protest. If they fail to do so the IPID places no obligation on the police to justify the use of force.

Along with the absence of information on disciplinary action against POP unit members described earlier, this suggests that complaints that are lodged with the SAPS or IPID are highly unlikely to result in disciplinary charges being brought against a police officer. Apart from general difficulties in establishing the truthfulness of complaints, one obstacle would appear to be the difficulty of identifying POP members who are allegedly responsible for inappropriate force or other abuses. If there are members of the public who feel aggrieved in public order incidents it is likely that, as a general rule, they are unable to identify the individual police officer allegedly responsible. Even if the SAPS member wears the requisite name tag it is unlikely that this will be visible to the victim of POP use of force.

Another issue is highlighted by the IPID response when a six-year-old girl was hit by rubber bullets while waiting for transport to school. The IPID report indicates that the investigation had concluded that '[i]t is unfortunate that the child was at the wrong place at the wrong time. There is no evidence to suggest the police official intended to injure the child.'⁴⁸ This is clearly inadequate as an assessment of whether the use of force by police was appropriate or not. The questions raised by the IPID should include whether police were aware of the presence of children in the area and took sufficient care to ensure that they were not harmed. Along with the evidence that people injured by POP are sometimes bystanders, this raises the question whether POP take adequate steps to ensure that force is targeted at individuals who are involved in violations of the law, rather than against the general public in the vicinity of a protest.

 

Conclusion

This article explores the intersection between two of the rights provided for in the South African constitution: the right to freedom of assembly and the right to information. Rather than generating their own data, the research projects discussed here have used information held by government departments, obtained through requests for information in terms of PAIA. In response to 19 requests lodged on behalf of the two projects the SAPS disclosed a total of 90 records, and the IPID disclosed five. One conclusion is that the SAPS responds positively to many requests for information. Although some requests did not generate the information that was sought, this may have been because they were not clearly formulated. It is therefore difficult to draw firm conclusions about what protest-related information the SAPS does not have or is not willing to disclose.

The SAPS exhibits relatively high levels of transparency. But many of the key records provided indicate that SAPS data on protest and how police respond to it have limited utility. In combination, the uneven standards of data entry, the absence of categories that differentiate protests from non-protest incidents, and the ambiguity of the unrest category, may lead one to conclude that IRIS mystifies more than it explains. There is a need for information that more readily lends itself to analysis of how the legal framework regarding protest is interpreted and applied by police.⁴⁹

The IPID also responded positively to the PAIA requests it received. But the IPID data did not include any cases from Gauteng, the province with the highest annual number of protests. This suggests that the data does not reflect all cases of protest that the IPID receives.

It remains unclear why senior government officials repeatedly present the IRIS data on crowd incidents as data on protests.⁵⁰If this does not demonstrate a deliberate misrepresentation of the data, it indicates that there is confusion about what the data represents, even at senior levels within the SAPS. It also indicates that SAPS data on protests and police responses to protests must be improved.

Better quality information is necessary in order to evaluate demand for, and resourcing of, public order police. It is also important to better understand when, why and how force is used in public order policing. As shown in this article, the SAPS is limited in its ability to answer questions on these issues. It also shows that mechanisms for holding police accountable for the use of force during protests are inadequate. This is sometimes due to the fact that individual police officers accused of abusing force cannot be identified, and sometimes because the criteria that are applied in assessing officers' use of force are inadequate.

One argument in support of police transparency is that sharing information with the public may lead to 'police data being analysed in new ways', leading to insights that were not previously available.⁵¹ A further argument is that sharing information can build familiarity and trust in communities, and promote organisational legitimacy.⁵² However, according to US security expert Brian Jackson, 'making more data available to the public is a strategy for improving police-public trust'. But, 'that strategy will work only if the data is trusted'.⁵³

The exercises in transparency discussed in this article do indeed reveal the SAPS's willingness to respect laws governing transparency. They have also provided an opportunity for conducting new analyses of data held by the SAPS. However, these exercises reveal serious shortcomings in the SAPS, and expose it as an organisation that is uninformed about the nature of protest and its own responses to it. It therefore highlights one risk of transparency for the police: that their inadequacies may be exposed, resulting in their being subject to increased criticism. Hopefully the SAPS, IPID and others will continue to recognise the value of providing data to members of the public in compliance with South Africa's FOI laws. But transparency will better contribute to trust only if the quality of the information provided can be improved.

This article has been written as part of a project on police transparency funded by the Open Society Justice Initiative. PAIA requests in South Africa were submitted by the Freedom of Information Programme at the SAHA. Thanks to Sandy Coliver and others involved in the OSJI project, Toerien van Wyk, Catherine Kennedy, Imraan Abdullah, Thomas Crankshaw and others at SAHA, as well as the SAPS and IPID officials involved, for their assistance.

 

Notes

1 Joshua Chanin and Salvador Espinosa, Examining the determinants of police department transparency, Criminal Justice Policy Review, 27:5, 2016, 498-519.         [ Links ]

2 Ibid.

3 Access Info Europe, The transparency of the policing of protests, 2015, http://www.access-info.org/frontpage/15021 (accessed 27 September 2016).

4 There have, for instance, been numerous statements by President Jacob Zuma on the issue. See, for example, South African Government, President Jacob Zuma: State of the Nation Address 2014, 13 February 2014, http://www.gov.za/node/632440 (accessed 27 September 2016); News24Wire, Violent SA protests are apartheid's fault, says Zuma, 9 December 2015, http://businesstech.co.za/news/government/106687/violent-sa-protests-are-apartheids-fault-says-zuma/ (accessed 27 September 2016); Danielle Petterson, Service delivery protests - Zuma says government is delivering, Infrastructure News, http://www.infrastructurene.ws/2016/05/25/service-delivery-protests-zuma-says-government-is-delivering/ (accessed 27 September 2016).         [ Links ]

5 See Monique Marks and David Bruce, Groundhog Day? Public order policing twenty years into democracy, South African Journal of Criminal Justice, 27:3, 2014, 346-376, 362-368.         [ Links ]

6 See, for instance, Neo Goba, Wits academics and support staff join call to remove police from campus, Times Live, 18 October 2016, http://www.timeslive.co.za/local/2016/10/18/Wits-academics-and-support-staff-join-call-to-remove-police-from-campus (accessed 11 November 2016);         [ Links ] GroundUp, Wits academics call for minimal security to allow teachingto continue, 11 October 2016, http://www.groundup.org.za/article/wits-academics-call-security-allow-teaching-continue/ (accessed 11 November 2016); Shenaaz Jamal and Aron Hyman, Student shot '13 times in the back', Times Live, 21 October 2016, http://www.timeslive.co.za/thetimes/2016/10/21/Student-shot-13-times-in-the-back (accessed 11 November 2016).         [ Links ]

7 PAIA Civil Society Network (PAIA CSN), Shadow report: 2013, 3, http://foip.saha.org.za/uploads/images/PCSN_ShadowRep2013_final_20131029.pdf (accessed 27 September 2016); PAIA CSN, Shadow report: 2014, 6-7, http://foip.saha.org.za/uploads/images/PCSN_ShadowRep2014_final_20150202.pdf (accessed 27 September 2016).

8 David Bruce, Gareth Newham and Themba Masuku, In service of the people's democracy: an assessment of the South African Police Service, Johannesburg: Centre for the Study of Violence and Reconciliation and Open Society Foundation for South Africa, 2007, 53, http://www.csvr.org.za/wits/papers/papsaps.htm (accessed 27 September 2016).         [ Links ]

9 South African Police Service (SAPS), Annual report 2014-15, 337-338, http://www.saps.gov.za/about/stratframework/annual_report/2014_2015/SAPS_AR_2014-15_for_viewing.pdf (accessed 27 September 2016); SAPS, Annual report 2015-16, 296-298, http://www.saps.gov.za/about/stratframework/annual_report/2015_2016/saps_annual_ report_2015_2016.pdf (accessed 10 November 2016); David Bruce, Measuring outputs, neglecting outcomes: the Auditor General's role in SAPS performance assessments, South African Crime Quarterly, 38, 2011, 3-13, http://www.issafrica.org/pgcontent.php?UID=31216 (accessed 27 September 2016).         [ Links ]

10 Gareth Newham, The politics of crime statistics, Africa Check, 22 September 2013, https://africacheck.org/2013/09/22/the-politics-of-crime-statistics-2/ (accessed 27 September 2016).

11 SAPS, Annual report 2015-16, 152-153.

12 Examples of IRIS data being interpreted in this way include Bruce, Newham and Masuku, In service of the people's democracy, 30; Jenna Etheridge, Paul Herman and Jenni Evans, New crime stats 2016: 'the stuff of nightmares', News24, 2 September 2016, http://www.news24. com/SouthAfrica/News/crime-stats-2016-the-stuff-of-nightmares-20160902 (accessed 27 September 2016).         [ Links ]

13 SAPS, Crime situation in South Africa, 1 April 2015 - 31 March 2016, http://www.saps.gov.za/services/final-crime-stats-release-02september2016.pdf (accessed 27 September 2016).

14 SAPS, Annual report 2015-16, 153.

15 Peter Alexander, Carin Runciman and Boitumelo Maruping, South African Police Service data on crowd incidents: a preliminary analysis, Social Change Research Unit, University of Johannesburg, 2015, 57.         [ Links ]

16 Rebecca Davis, Is SAPS cooking the books on protest numbers?, Daily Maverick, 31 May 2015, http://www.dailymaverick.co.za/article/2015-05-31-is-saps-cooking-the-books-on-protest-numbers/ (accessed 25 November 2016).         [ Links ]

17 Promotion of Access to Information Act 2000 (Act 2 of 2000), Pretoria: Government Printer, 2000, sections 9(a) and 9(b)(i), Part 2, Chapter 4.

18 Ibid., Section 14.

19 SAPS, Manual of the South African Police Service in accordance with Section 14 of the Promotion of Access to Information Act 2000 (Act 2 of 2000), 2016, http://www.saps.gov.za/resource_centre/paia/paia.php (accessed 11 October 2016); Independent Police Investigative Directorate (IPID), Independent Police Investigative Directorate Access to Information Manual in terms of Section 14(1) of the Promotion of Access to Information Act, 2000 (Act 2 of 2000), http://www.icd.gov.za/sites/default/files/about-us/access_to_information. pdf (accessed 27 September 2016).

20 Promotion of Access to Information Act, Section 15.

21 Government Gazette, Categories of records automatically available, 20 May 2016, http://www.saps.gov.za/resource_centre/paia/annexure_a1_saps_sec_15.pdf (accessed 27 September 2016).

22 South African History Archive (SAHA), PAIA tracker system, http://www.saha.org.za/projects/paia_tracker.htm (accessed 27 September 2016).

23 Carin Runciman et al., Counting police-recorded protests: based on South African Police Service data, Social Change Research Unit, University of Johannesburg, 2016, 13.         [ Links ]

24 The author of this article was contracted to work on the Open Society Justice Initiative (OSJI) project as the South African researcher in cooperation with the South African History Archive (SAHA).

25 OSJI and Transparency Audit Network, Police transparency: evaluating access to information in relation to the policing of public gatherings in Brazil, India, Mexico, South Africa and the United Kingdom, 2015, http://www.gregmichener.com/ten/sites/default/files/articles/transparency-policing-public-protests-osji-fgv.pdf (accessed 25 November 2016).

26 Ibid., 4.

27 In 2015, following the release of the Marikana Commission report, the Right2Know campaign tried unsuccessfully to obtain information about the disciplinary action taken against SAPS members involved in the killings of the Marikana strikers. See Right2Know Campaign and Marikana Support Campaign, Marikana killer cops still walking the streets: R2K and MSC call for the dismissal of implicated police!, 7 July 2015, http://www.r2k.org.za/2015/07/07/marikana-killer-cops/ (accessed 27 September 2016); Right2Know Campaign and Marikana Support Campaign, Statement: police fail to answer our questions on Marikana killers, 9 September 2015, http://www.r2k.org.za/2015/09/09/police-fail-to-answer-our-questions-on-marikana-killers/ (accessed 27 September 2016). In 2016 a number of requests were submitted, both to the SAPS and to municipalities, by the Legal Resource Centre, an NGO. Information obtained is published on Legal Resources Centre, Protest info, www.protestinfo.org.za.

28 SAHA Freedom of Information Programme (FOIP), Records relating to standard operating procedures on the control of public gatherings, http://foip.saha.org.za/request_tracker/entry/sah-2015-sap-0012 (accessed 25 November 2016).

29 SAHA FOIP, Request for records related to disciplinary action taken for police conduct and/or abuse connected to gatherings/ crowd management/public order policing operations, http://foip.saha.org.za/request_tracker/entry/sah-2015-sap-0033 (accessed 25 November 2016); Affidavit of BRJ Heyneke, signed in Rosslyn, 28 October 2015.

30 Right2Know Campaign and Marikana Support Campaign, Marikana killer cops still walking the streets; Right2Know Campaign and Marikana Support Campaign, Statement.

31 SAHA FOIP, Copies, reports on and/or records related to investigations under Standing Order 251 and related to the Public Order Policing Unit, http://foip.saha.org.za/request_tracker/entry/sah-2015-sap-0029 (accessed 25 November 2016)

32 Promotion of Access to Information Act, Part 4, Chapter 1.

33 In this case the SAPS also provided an affidavit stating that it did not have the information. This might have been related to the fact that the request was for 'number of injuries'. It is possible that the request may have been more successful if records on the 'number of people injured' had been requested.

34 Ministry of Police, Policy and guidelines: policing of public protests, gatherings and major events, 2011, 17, http://www.policesecretariat.gov.za/downloads/policies/policing_public_ protests_2013.pdf (accessed 27 September 2016).

35 Alexander, Runciman and Maruping, South African Police Service data on crowd incidents, 11.

36 Runciman et al., Counting police-recorded protests, 31.

37 Alexander, Runciman and Maruping, South African Police Service data on crowd incidents; Runciman et al., Counting police-recorded protests, 2016.

38 Carin Runciman et al., Counting police-recorded protests: based on South African Police Service data, Social Change Research Unit, University of Johannesburg, 2016.

39 Alexander, Runciman and Maruping, South African Police Service data on crowd incidents, 44-45. Other categories are crime related, education, party political, transport issues, customary, xenophobia, other.

40 Ibid., 26.

41 SAH-2015-SAP-0031.

42 Provided for by the Regulation of Gatherings Act 1993 (Act 205 of 1993), Pretoria: Government Printer, Section 3.

43 Steven Tau, Angry residents must relocate, The Citizen, 12 September 2013, http://citizen.co.za/42980/waterworks-informal-settlement-not-conducive-to-human-settlement-according-to-department-of-cooperative-governance-and-traditional-affairs/ (accessed 27 September 2016).         [ Links ]

44 See, for instance, the use of rubber bullets described in Richard Poplack, #FeesMustBrawl: the battle of Braamfontein, Daily Maverick, 11 October 2016, http://www.dailymaverick.co.za/article/2016-10-10-trainspotter-feesmustbrawlthe-battle-of-braamfontein/# (accessed 11 October 2016).

45 Alexander, Runciman and Maruping, South African Police Service data on crowd incidents, 20, 43

46 Bruce, Newham and Masuku, In service of the people's democracy, 31.

47 Lizette Lancaster, At the heart of discontent: measuring public violence in South Africa, Institute for Security Studies, Paper, 292, 2016, 13; Runciman et al., Counting police-recorded protests, 40.

48 SAHA, Independent Police Investigative Directorate, Case 44 in untitled 51-page document with table listing outcome of 99 complaints, 2015, 26-27, http://www.saha.org.za/collections/AL2878/a200502.htm (accessed 28 November 2016).

49 See the Regulation of Gatherings Act.

50 Alexander, Runciman and Maruping, South African Police Service data on crowd incidents, 57-59.

51 Police Foundation, Five things you need to know about open data in policing, 2015, https://www.policefoundation.org/wp-content/uploads/2015/04/PF_FiveThings_OpenDateInPolicing_Handout_Rev6.18.15.pdf (accessed 27 September 2016).

52 Chanin and Espinosa, Examining the determinants of police department transparency, 501.

53 Brian A Jackson, Strengthening trust between police and the public in an era of increasing transparency, Rand Corporation, 2015, http://www.rand.org/content/dam/rand/pubs/testimonies/CT400/CT440/RAND_CT440.pdf (accessed 27 September 2016).