Research Protocol for Medical Marijuana Dispensaries. Prepared by Legal Science, LLC

Research Protocol for Medical Marijuana Dispensaries Prepared by Legal Science, LLC July 2016

MEDICAL MARIJUANA DISPENSARIES July 2016 Research Protocol I. Dates of Protocol: June 25, 2014; July 8, 2014; July 14, 2014; August 4, 2014; August 12, 2014; August 14, 2014; August 18, 2014; February 6, 2015; July 30, 2015; February 26, 2016; August 24, 2016 II. III. Scope: Compile statutes, regulations and current dispensary-related medical marijuana/cannabis laws across all 50 states and the District of Columbia. This cross-sectional dataset includes coding questions about medical marijuana dispensary topics such as: limitations on the number of dispensaries, location restrictions for dispensaries, dispensary license fees, minimum dispensary employee age, supply to outof-state patients, product supply limitations, medical marijuana taxation, medical marijuana product display, and on-site consumption of medical marijuana products. Primary Data Collection A. Dates covered in the dataset: June 2, 1999 July 1, 2016. This dataset project began as a cross-sectional study intended to capture present day law as of September 1, 2014. Each state record is a compilation of relevant statutes and regulations that were in effect as of September 1, 2014. The states record effective date is the effective date of the most recent statute or regulation within the state record. Thus, there are laws within this dataset that were enacted prior to September 1, 2014, however they are still effective so they are included in the state record and assigned the appropriate states record effective date. B. Data collection methods: The team building this dataset consisted of two legal researchers ( Researcher #1 and Researcher #2 or Researchers ) and one supervisor ( Supervisor ). The two legal researchers were each assigned half of the states at random. C. Databases used: Searches conducted using WestlawNext, Lexis Advance, HeinOnline, National Conference of State Legislatures website, and state-specific legislature websites i. Full text versions of the laws collected were pulled from each respective state legislature website. D. Search terms: Medical marijuana; medical cannabis; dispensary; dispensaries business; compassion center; retail marijuana dispensary; RMD; producer license; medical marijuana center; medical marijuana and business; alternative treatment center; medical cannabis and sales; medical cannabis and business i. Key word searchers were supplemented by examination of the table of contents of each relevant section of the state law identified. Researchers also collected effective dates for the most recent version of relevant statutes and regulations ii. Once all the relevant statutes and regulations were identified for a jurisdiction, a master sheet was created that summarized the relevant statute or regulation, included the most recent statutory history for each statute and regulation, and added the effective date for that version of the law.

3 iii. The initial 10 states were 100% redundantly researched to confirm that all relevant law was being collected by the Researchers; 20% of the remaining states were redundantly researched. E. Information about initial returns and additional inclusion or exclusion criteria: Included laws providing definitions of qualified patients and dispensary. Excluded laws pertaining specifically to caregivers and patient identification cards without mention or application to medical marijuana dispensaries. IV. Coding A. Development of coding scheme: To develop coding questions the Supervisor consulted with one subject matter expert, Rosalie Liccardo Pacula, PhD. The Researchers, Supervisor and the subject matter expert drafted coding questions and circulated them for review until all parties felt they had been sufficiently refined. Once the coding questions were finalized, they were entered into the LawAtlas Workbench. i. States vary on whether they call the substance medical marijuana or cannabis. Although the language in this dataset will use the term medical marijuana, state laws using the term cannabis will be collected and coded identically. ii. Statutes on dispensary licenses were included if they referred to dispensary license as a dispensary registration, registration certificates or dispensary certificate. The team decided that these should be coded the same as a dispensary license, as there appears to be no actual difference in the law. 1. Furthermore, not all state laws use the term dispensary when referring to a center that issues medical marijuana to caregivers and patients. For example, Rhode Island calls dispensaries compassion centers. The team closely examined such states to determine that these facilities are identical to dispensaries and differ only in name. Therefore, these facilities were coded just as they would be if they were called dispensaries. iii. While researching the law, the Researchers found that state medical marijuana programs were regulated by different branches of the government depending on the state (i.e. the Department of Health). It was decided with the Supervisor that any references to the Department would represent the department that is responsible for regulating the medical marijuana program in that state. Furthermore, a question was added that asks which state agency regulates dispensaries. Because some states have agencies who play minor roles in medical marijuana regulation, the team will code for the state agency or agencies that generally regulates dispensaries under the state medical marijuana law. iv. Some states make a distinction between dispensaries that are classified as nonprofits because of their tax status under the IRS and dispensaries that run not-for-profit. To answer our question about dispensaries operating as not-for-profit entities, the Researchers noted the difference between a nonprofit tax status dispensary and a dispensary ran not-for-profit. 1. In coding this question, researchers will only code for not-for-profit entities. v. The Researchers discovered that the medical marijuana dispensary laws in both California and Connecticut seemed unique and required additional explanation for comprehension. vi. The California Compassionate Use Act was passed by voters in 1996, making it legal for patients and their primary caregivers to possess and cultivate marijuana with a recommendation or approval from a California-licensed physician. Senate Bill 420, passed by the California state legislature in 2003 established the Medical Marijuana Program. The bill was codified into law as the California Health and Safety Code Section 11362.7-11362.83 and allows patients to form collectives and cooperatives. Additionally, the Act established the voluntary identification card program and gave the Attorney General the authority to set forth and clarify details regarding medical marijuana regulations. The Medical Marijuana Identification Card program is overseen by the California Department of

4 Public Health (DPH). This program only handles medical marijuana identification cards and the registry database for verification of qualified patients and primary caregivers. Medical marijuana users are not required to have a state identification card; however, having a card may protect them from arrest. The DPH does not have jurisdiction over cooperatives, dispensaries, or collectives, and therefore advises interested parties to follow the guidelines given by the State Attorney General s Office. Due the lack of clarity regarding medical marijuana in California, in 2008 the Attorney General issued guidelines specifically addressing the laws, patients, primary caregivers, and collective and cooperatives. In the guidelines, the Attorney General acknowledged the existence of medical marijuana dispensaries, but reiterated that they are not recognized under California law. Only cooperatives and collectives are. However, he goes on to say that those collectives and cooperatives that dispense medical marijuana through a storefront are lawful as long as they are abiding by these guidelines. 1. We coded answers using the Attorney General Guidelines and the statues passed by the California legislature. Only collectives and cooperatives, not medical marijuana dispensaries, are recognized under California state law. According to the Attorney General Guidelines, however, the dispensing of medical marijuana through a storefront by collective or cooperative may be considered lawful if they abide by these guidelines. We coded these collective or cooperative laws as dispensaries to represent the medical marijuana program in California while noting the difference in this state. vii. In Connecticut, a dispensary is defined as a person who has received a license from the Commissioner of Consumer Protection and is therefore able to dispense marijuana to qualifying patients and primary caregivers. This licensed person works at a dispensary facility, which is a place of business where marijuana is dispensed or sold. No more than five dispensaries may work at a dispensary facility without approval from the Commissioner. All fees associated with a dispensary must be paid before a license is issued or renewed. 1. We used the laws applicable to dispensary facilities in order to code dispensary related questions. viii. Definitions of terms used by the Researchers: 1. Medical marijuana: Marijuana refers to the seeds, leaves, buds, and flowers of the cannabis plant and any mixture or preparation thereof, but does not include the stalks, and roots of the plant. Medical use means the acquisition; administration; delivery; possession; transfer; transportation; or use of marijuana to treat or alleviate a registered qualifying patient's statutorily specified debilitating medical condition or symptoms associated with the patient's debilitating medical condition. 2. Not-for-profit: An organization that uses revenues in excess of operating costs in order to facilitate its operation instead of distributing profits. A medical marijuana dispensary classified as a not-for-profit does not need to be recognized as taxexempt by the International Revenue Services. 3. Dispensary: A business that acquires, possesses, cultivates, manufactures, delivers, transfers, transports, supplies or dispenses marijuana, and/or related supplies and educational materials, to registered qualifying patients and/or their registered primary caregivers who have designated it as one of their primary caregivers. B. Coding methods: Researcher #1 and Researcher #2 were responsible for coding the dataset in the Workbench. The Researchers were assigned a number of states to code independently. i. Several issues were identified in the coding through discussions between the Supervisor and the Researchers and we coded accordingly. ii. Specifically, we narrowed our scope to focus on the information the average citizen would want if they were concerned about dispensaries operating in their community. This meant

5 that we excluded questions that detail the process required to open or run a dispensary such as questions relating to dispensary application requirements and employee registration requirements. iii. In regard to the question, Are there fees to obtain a dispensary license? we answered in the affirmative for states that do have dispensary fees, even if the exact amounts have yet to be determined by the appropriate state department. iv. In regard to the question, Are dispensaries prohibited from displaying marijuana products? we only answered in the affirmative if the statute specifically prohibited storefront display of marijuana products. We answered in the negative if the statute only prohibited cultivation or production from being visible. v. In regard to the question, Does the state prohibit consumption of marijuana products at dispensaries? the team decided that we would code for prohibitions on consumption of marijuana products by patients. Therefore, our answers were not affected by states that do allow for consumption by employees during certain times (Oregon) or states that allow consumption for the purpose of teaching or demonstrating the use of products (Massachusetts). C. Quality control: Thirty-one percent of all records were redundantly coded by both Researchers. i. Of the 19 states with laws on medical marijuana dispensaries, 16 were redundantly coded. Because of the complex area of law and adjustments to the questions, eighty-four percent of the states with laws on medical marijuana dispensaries were redundantly coded. This means that two records were created containing the same law, questions and date. Both Researcher #1 and Researcher #2 coded these records, resulting in two records for the same jurisdiction and date. Records were selected for redundant coding by the Supervisor. ii. The Supervisor reviewed the redundant coding by downloading the data from the Workbench into Microsoft Excel and comparing the records, variable by variable, looking for divergences. When a divergence was identified, it was discussed with the researchers. The reason for the divergence was identified and resolved. A measure of divergence was calculated by the Supervisor and the duplicate record was deleted. 1. In the preliminary coding phase, 100% redundant coding occurred. Each Researcher redundantly coded the other Researcher s first five states. The preliminary rate of divergence on July 16, 2014 was 39.7% Because of the high rate of divergence, the team re-evaluated their questions and answer choices to make it more representative of what is actually present in the law, as discussed above. After these changes were made, the first five states for each Researcher were re-coded and on July 28, 2014 the new rate of divergence was 13.8% 2. Each Research redundantly coded another three states each. The average rate of divergence for this round of redundant coding was 28.1%. The Researchers met with the Supervisor to discuss any additional problems with the coding scheme. 3. The overall rate of divergence for the redundant coding as of August 11, 2014 is 29.8%. iii. The Supervisor oversaw the overall quality of the data by downloading the data from the Workbench into Microsoft Excel and reviewing it in order to find caution flags, missing answer choices, divergences in redundant records and errors in the coding. Daily coding review sheets were sent to the Researchers for their review. Issues in the coding were discussed by the Supervisor and the researchers in coding meetings. iv. A naïve coder was brought on towards the end of the process to code 20% of the records. On August 5, 2014, the Supervisor assigned five states at random to the naïve coder who had not previously worked on this dataset. Once naïve coding was complete, the team met with the naïve coder to discuss and resolve any divergences. 1. The rate of divergence for the naïve coding was 27.6%.

6 V. Update (January 2015) A. Data collection methods: Research was conducted by one Researcher ( Researcher #1 ). The research sought to identify all laws enacted or amended between September 2014 and January 31, 2015 and any pending legislation close to passage. The work of Researcher #1 was overseen by a Supervisor. Searches were conducted by Researcher #1 for enacted bills amending the medical marijuana statutes and regulations in the Workbench, or new bills regarding medical marijuana dispensaries. The searches were conducted in Westlaw Next using the following syntax and search terms: Medical marijuana; medical cannabis; dispensary; dispensaries business; compassion center; retail marijuana dispensary; RMD; producer license; medical marijuana center; medical marijuana and business; alternative treatment center; medical cannabis and sales; and medical cannabis and business. i. During the January 2015 update, the Researcher coded two previous amendments to the D.C. medical marijuana laws. On July 29, 2014, the legislature passed an emergency amendment that altered the definition of qualifying medical condition and increasing the number of living marijuana plants a cultivation center may possess from 95 to 500. Because this was passed as an emergency amendment, the changes to the law are only permitted to remain in effect for no longer than 90 days. There is no record of exactly when this emergency amendment expired, but on November 12, 2014, the legislature passed a permanent amendment making the very same changes to the law. Therefore, the team decided to code these two amendments as separate iterations. B. Secondary sources used to collect the laws: state-specific legislature websites; OpenStates.org. C. Coding updated findings: Researcher #1 found laws had been amended in the District of Columbia, Colorado, Hawaii, Illinois and Maine. i. For the question, Dispensaries must be located at a minimum distance from which facilities?, the coding answer school zone requires the law to say explicitly school zone. A school zone includes the entire school grounds, the school and the surrounding residential neighborhood. D. Quality control: The supervisor assigned 20% redundant coding for the updated entries. The rate of divergence was 20%. All divergences were discussed in a coding review meeting. The final preview link was sent to our Content Expert, Rosalie Pacula Ph.D., for review. VI. Updated (July 2015) A. Scope: For the July 2015 update, we have been tracking medical marijuana laws since January 1, 2015 through July 1, 2015. B. Data collection methods: One Update Supervisor and one researcher checked all existing legal citations on Westlaw and legislative tracking websites for pending and proposed legislation since the last update. The team determined that eight jurisdictions had relevant amendments, including the District of Columbia, Delaware, Maryland, New Mexico, New York, Rhode Island and Nevada. The Researchers created new records with updated legal text for states with new laws or changes to existing laws that affected answer choices. Researchers cloned records and updated the legal text for states with changes that did not affect answer choices. C. Coding updated findings i. Coding methods: During the July 2015 update, the team found that 2 states: New Mexico and New York had substantive changes to the law, meaning that these changes affected an answer choice, which required re-coding. Researcher #1 coded the updates for New Mexico and New York. Researcher #2 redundantly coded these entries. Supervisor

7 reviewed the coding and held a coding review session. All divergences were discussed and resolved. The team also found that 6 jurisdictions: the District of Columbia, Delaware, Colorado, Maryland, Rhode Island and Nevada had non-substantive changes to the law, meaning these changes did not affect an answer choice. Because the changes did not affect an answer choice, these records were cloned and not re-coded or redundantly coded. Once these records were cloned, the legal text for each record was updated with the new law. ii. For the question How much is the initial dispensary license application fee? : The team decided to add the answer choice $10,000 based on changes to the law in New York and New Mexico. iii. For the question How much is the license fee after an application has been approved? : The team decided to add the answer choice $200,000 based on changes to the law in New York. D. Quality control: The Supervisor downloaded records into Excel and ran divergence calculations on the redundantly coded record. The Supervisor also compared newly coded entries to previously coded entries for consistency. The rate of divergence for the July 2015 update was 8.3%. Researcher #1 re-coded as necessary. All divergences were discussed and resolved. VII. Updated (February 2016) A. Scope: For the February 2016 update, we have been tracking medical marijuana laws since July 1, 2015 through February 1, 2016. B. Data collection methods: Two Researchers checked all existing legal citations on Westlaw and legislative tracking websites for pending and proposed legislation since the last update. The team determined that twelve jurisdictions had relevant amendments, including California, Colorado, Delaware, the District of Columbia, Hawaii, Maryland, Nevada, New Hampshire, New York, Oregon, Vermont, and Washington. The Researchers created new records with updated the legal text for states with new laws or changes to existing laws that affected answer choices. Researchers cloned records and updated the legal text for states with changes that did not affect answer choices. C. Coding updated findings i. Coding methods: During the February update, the team found that 7 states had substantive changes to the law that required re-coding. The states included: California, Delaware, the District of Columbia, Hawaii, New Hampshire, Maryland, and Vermont. Because several of the amendments caused substantive changes in coding answers, the Supervisor assigned 20% redundant for the updated entries. ii. For the question, How far must dispensaries be located from these facilities? : The team decided to add the answer choice At least 1320 ft. based on changes to the law in New Hampshire. iii. For the question, How much is the license fee after an application has been approved? : The team decided to add the answer choice $75,000 based on changes to the law in Hawaii. iv. For the question, How much is the fee to renew a dispensary license? : The team decided to add the answer choice $50,000 based on changes to the law in Hawaii. D. Quality control: The Supervisor reviewed the original coding for each updated entry by downloading the records into Excel. Following redundant coding, the Supervisor downloaded records into Excel and ran divergence calculations on the redundantly coded record. The Supervisor also compared newly coded entries to previously coded entries for consistency. The rate of divergence for the February 2016 update was 7.8%. The Researchers re-coded as necessary. All divergences were discussed and resolved.

8 VIII. Updated (July 2016) A. Scope: For the July 2016 update, we have been tracking medical marijuana laws since February 2, 2016 through July 1, 2016. B. Data collection methods: Two Researchers checked all existing legal citations on Westlaw and legislative tracking websites for pending and proposed legislation since the last update. The team determined that thirteen jurisdictions had new laws or relevant amendments, including California, Colorado, the District of Columbia, Hawaii, Illinois, Minnesota, New Hampshire, New Mexico, Oregon, Pennsylvania, Rhode Island, Vermont, and Washington. The Researchers created new records with updated the legal text for states with new laws or changes to existing laws that affected answer choices. Researchers cloned records and updated the legal text for states with changes that did not affect answer choices. C. Coding updated findings i. Coding methods: During the July update, the team found that several states had substantive changes to the law that required re-coding and Pennsylvania added a new medical marijuana program. Therefore, the Supervisor assigned 20% redundant coding for the substantively updated entries, including Pennsylvania. 1. For the questions on fee amounts, the Supervisor added new answer choices as needed to reflect any new values from recent amendments. 2. For the state of Washington, the team consulted with the content expert to determine how to best represent the dispensary laws since the state legalized the recreational sale of marijuana. Since commercial establishments can now apply to dispense medical marijuana, we decided to capture these laws as if it were a medical marijuana dispensary. 3. After consulting with the content expert, the team determined that providing the effective date of the first medical marijuana law in each jurisdiction would be a helpful tool for researchers. Therefore, the team added the question What is the effective date of the first medical marijuana law? to each dataset and re-coded accordingly. D. Quality control: The Supervisor reviewed the original coding for each updated entry by downloading the records into Excel. Following redundant coding, the Supervisor downloaded records into Excel and ran divergence calculations on the redundantly coded record. The Supervisor also compared newly coded entries to previously coded entries for consistency. The rate of divergence for the July 2016 update on August 11 was 8.5%. All divergences were discussed and resolved. The Supervisor assigned two more states for redundant coding. The rate of divergence on August 12 was 4.25%. All divergences were discussed and resolved. i. The Supervisor also added notes to Louisiana and Florida which have passed laws allowing the establishment of limited low-thc programs, but do not have comprehensive medical marijuana programs.