Preview
FILED: NEW YORK COUNTY CLERK 07/06/2023 11:59 AM INDEX NO. 155947/2023
NYSCEF DOC. NO. 29 RECEIVED NYSCEF: 07/06/2023
EXHIBIT 1
FILED: NEW YORK COUNTY CLERK 07/06/2023 11:59 AM INDEX NO. 155947/2023
NYSCEF DOC. NO. 29 RECEIVED NYSCEF: 07/06/2023
Expert Report of Jonah Berger, Ph.D.
July 5, 2023
FILED: NEW YORK COUNTY CLERK 07/06/2023 11:59 AM INDEX NO. 155947/2023
NYSCEF DOC. NO. 29 RECEIVED NYSCEF: 07/06/2023
I. Background and Assignment
1. On November 16, 2022, the New York City Department of Consumer and Worker
Protection ( "DCWP") published a proposed rule in the City Record to implement Local Law
115 of2021 (the "Proposed Rule"). 1 According to DCWP, "Local Law 115 of2021 charged
the Department with studying [food delivery workers who work with apps as independent
contractors] and developing an appropriate minimum pay rate to ensure adequate compensation
for these workers." 2 Also in November 2022, DCWP published a report entitled "A Minimum
Pay Rate for App-Based Restaurant Delivery Workers in NYC" that discussed, among other
things, DCWP' s "proposed rule to establish a minimum pay rate" for food delivery work, as
well as three different surveys that DCWP conducted, including "an online survey distributed
to 123,000 workers who performed deliveries in NYC in the fourth quarter of202 l " (the "NYC
Delivery Worker Survey").3
2. DCWP states that it conducted the surveys in order to gain information about NYC
delivery workers' expenses, demographics, safety conditions, work history, and experiences
with discipline and non-payment on food delivery apps as well as NYC restaurant owners and
managers' delivery experiences. 4 Information obtained from these surveys formed, in part, the
basis for DCWP' s Proposed Rule establishing a minimum pay rate for NYC delivery workers. 5
3. On December 16, 2022, Dr. Itamar Simonson submitted a report (the "Simonson
Report") before DCWP regarding the Proposed Rule. 6 In this report, Dr. Simonson evaluated
the three surveys that DCWP conducted. Dr. Simonson opined that these surveys "were biased
and were bound to produce unreliable results" because they "violated fundamental principles
of survey design."7 For instance, Dr. Simonson finds that the survey preamble explicitly told
respondents that the purpose of the survey was to raise pay for app delivery workers, which in
1
New York City Department of Consumer and Worker Protection, ''Notice of Adoption of
Final Rule" ("Notice of Adoption of Final Rule"), p. 1.
2
Notice of Adoption of Final Rule, p. 1.
3
Notice of Adoption ofFinal Rule, p. 22; New York City Department of Consumer and
Worker Protection, "A Minimum Pay Rate for App-Based Restaurant Delivery Workers in
NYC," November 2022 (the "Minimum Pay Rate Report"), p. ii.
4
Minimum Pay Rate Report, pp. 2- 3.
5
Minimwn Pay Rate Report, p. ii.
6
Expert Report of Dr. Itamar Simonson Re: Proposed Rule on Minimum Pay for Food
Delivery Workers, December 16, 2022.
7
Simonson Report, 1.
Page 1
FILED: NEW YORK COUNTY CLERK 07/06/2023 11:59 AM INDEX NO. 155947/2023
NYSCEF DOC. NO. 29 RECEIVED NYSCEF: 07/06/2023
turn led responses to exaggerate their estimates of expenses. 8 As such, Dr. Simonson
determined that "it would not be consistent with accepted survey science to rely upon [the
surveys'] results for use in justifying the Proposed Rule."9
4. DCWP subsequently issued a "Notice of Adoption of Final Rule," which included,
among other things, purported responses to Dr. Simonson's commentary. 10 DCWP also
clarified that, while it relied on data from its NYC Delivery Worker Survey in developing the
proposed minimum pay rate, it did not use data from the other two surveys for this purpose. 11
5. I have been asked by counsel for DoorDash, Inc. and Grubhub, Inc. to i) review the
Simonson Report and offer my opinion on the flaws pertaining to DCWP's surveys identified
therein; and ii) review DCWP's responses to the commentary in the Simonson Report to
determine whether the flaws identified by Dr. Simonson have been resolved.
JI. Summary of Opinions
6. I agree with the Simonson Report that DCWP's surveys suffer from numerous flaws
that run counter to standard practices for designing and conducting surveys. Accordingly,
DCWP's surveys are biased and unreliable. Among others, these flaws include (1) relying on
an inappropriate and leading initial prompt, which should encourage demand effects; (2)
multiple problems with question construction, including the use of leading/biased questions,
focalism bias, and inappropriate use of closed-ended, rather than open-ended, questions; (3)
failing to include control, (or "phantom") questions, or any other controls to account for
guessing and demand effects; and (4) failing to validate survey responses.
7. DCWP's responses to Prof. Simonson's assertions do not resolve the flaws in the NYC
Delivery Worker Survey, and, as a result, the survey remains unreliable. First, DCWP's
response does not address the fact that, even if customary in other DCWP surveys, the prompt
that appears at the beginning of the NYC Delivery Worker Survey invalidated the survey
results in at least two ways: it biased the responses of those workers who participated ("Answer
Bias"), and it biased the survey results as a whole by influencing who took the survey in the
first place ("Participation Bias"). These biases threaten the survey's representativeness and
make it invalid. Second, DCWP's finding that respondents had difficulty providing certain
8
Simonson Repmt, ir,[ 31-33.
9
Simonson Report, ,r 4.
10
Notice of Adoption of Final Rule, pp. 22-25.
11
Notice of Adoption of Final Rule, p. 22.
Page 2
FILED: NEW YORK COUNTY CLERK 07/06/2023 11:59 AM INDEX NO. 155947/2023
NYSCEF DOC. NO. 29 RECEIVED NYSCEF: 07/06/2023
open-ended answers does not support using leading and inappropriate closed-ended answers,
but instead suggests that respondents may have guessed and biased their responses. Third,
DCWP provided no evidence for its claims that control questions and other validation
techniques were not necessary, even when such practices are standard in survey design.
Finally, DCWP failed to address that it did not attempt to verify certain responses to questions,
for example, by asking for receipts to prove claimed purchases.
III. Qualifications
8. I am a Marketing professor at the Wharton School at the University of Pennsylvania. I
received my Ph.D. in Marketing from the Graduate School of Business at Stanford University
and my B.A. in Human Judgment and Decision Making, also from Stanford University.
9. My research focuses on analyzing consumer behavior, word of mouth, and social
influence, through the use of scientific techniques such as surveys and natural language
processing. I have published over 70 articles in top-tier academic journals in marketing,
consumer psychology, and other disciplines including in the Journal of Marketing, Journal of
Marketing Research, Journal of Consumer Research, and Journal of Consumer Psychology. I
have also written three best-selling books that have been printed in over 30 different languages.
In addition, I serve as an Associate Editor at the Journal of Marketing. In my academic and
other work I have conducted or evaluated over a thousand surveys.
10. I have received numerous awards for my research and writing. In 2014, I received the
American Marketing Association's Leonard L. Berry Marketing Book Award, which is given
for a book that has had a significant impact in marketing and related sub-fields. In 2017, I
received the Journal of Marketing's William F. O'Dell Award for the article that made the
most significant, long-term contribution to marketing theory, methodology, and/or practice. In
2023, I received the Sage 10-Year Impact Award for one of the three most impactful articles
published in all Sage journals. I was recognized from 2008 to 2023 by the American Marketing
Association as one of the most productive researchers in marketing. The American
Management Association named me one of the top 30 leaders in business and Fast Company
magazine named me one of the most creative people in business.
11. Jn addition to my research, I have extensive experience teaching and consulting on these
topics. I have won numerous teaching awards from the Wharton School and have consulted
with hundreds of companies, including many Fortune 500 companies and foundations.
Page 3
FILED: NEW YORK COUNTY CLERK 07/06/2023 11:59 AM INDEX NO. 155947/2023
NYSCEF DOC. NO. 29 RECEIVED NYSCEF: 07/06/2023
12. A complete list of my publications and other academic and professional experience is
in my curriculum vitae, which is attached to this report as Appendix A.
IV. The Simonson Report Identified Numerous Flaws of DCWP's Surveys that
Render the Surveys Unreliable
13. The Simonson Report concludes that DCWP' s surveys "violated fundamental
principles of survey design" and identifies a number of flaws in the surveys.12 I agree with the
Simonson Report that the surveys suffer from numerous flaws that run counter to standard
practices for designing and conducting surveys. Accordingly, DCWP's surveys are biased and
unreliable. I briefly summarize the critical flaws identified in the Simonson Report in the
following paragraphs.
14. Flaw # 1: The Surveys did not avoid "Demand Effects," which are sometimes referred
to as "Demand Artifacts." 13 Demand effects refers to a phenomenon that leads survey
respondents to form an interpretation of the survey's purpose, and based on that, respond to the
questions in a manner that they believe the researcher wants, or that reflects the respondent's
self-interest. Here, for example, DCWP told the survey respondents that the purpose of the
sw-vey was to raise pay for app delivery workers. 14 According to the Simonson Report, demand
effects affected DCWP' s surveys and " [t]his fatal flaw, by itself, makes the surveys' results
unreliable." 15
15. Flaw #2: The Surveys suffered from multiple problems with question construction,
including the use of leading/biased questions, introduction of focalism bias, and inappropriate
use of closed-ended, rather than open-ended, questions. 16 When designing a survey, it is
important for questions to be objective, rather than to lead respondents toward a desired
response. As the Simonson Report correctly notes, leading questions can have a particularly
strong biasing effect when (as in DCWP's surveys), questions pertain to a respondent' s self-
reporting of their behaviors or experiences.17 For example, the Simonson Report identified the
question "How many hours per week is your car in use for something other than app delivery?"
12
Simonson Report, ,i l.
13
Simonson Report, ,i 2a.
14
Simonson Report, ,i 32.
15
Simonson Report, ,i 30.
16
Simonson Report, ,i 26, 2d, 2e.
17
Simonson Report, ,i 18.
Page 4
FILED: NEW YORK COUNTY CLERK 07/06/2023 11:59 AM INDEX NO. 155947/2023
NYSCEF DOC. NO. 29 RECEIVED NYSCEF: 07/06/2023
as leading because "the survey's introduction encouraged respondents to indicate that their car
and associated expenses are almost entirely attributed to their food delivery work." 18
16. Further, DCWP's surveys introduce focalism because they direct respondents to
consider items that respondents would not consider had the item not been singled out. 19
Focalism is understood to be a cognitive bias in which the choice of a survey respondent is
influenced by a piece of information that stands out in the context of a survey in a manner that
does not reflect the real world. The Simonson Report correctly notes that focalism bias exists
with respect to various expense categories provided in the Delivery Worker Survey. 2 ° For
example, the Simonson Report identified the question "When you started the job, how much
did you spend on buying delivery accessories like: GPS tracker, ... etc.?" as affected by
focalism because it "suggested categories of expenses that might not have applied to
respondents or that were purchased for additional reasons."21
17. Finally, the Simonson Report identifies that DCWP's surveys inappropriately rely on
closed-ended questions when open-ended questions would be possible and preferable. The use
of closed-ended questions is problematic in this case because the list that is provided to survey
respondents may be understood to include "correct" answers. For example, the Simonson
Report identified the question "During the past 12 months, how many batteries have you bought
for your moped?," followed by options "O, 1, 2, 3, 4, 5 or more" as an inappropriate closed-
ended question because the response options "suggested the expected response range."22 I
agree with the Simonson Report in this regard because closed-ended responses in this case
suggest a normative answer, or answer most respondents are likely to give. Consequently, a
respondent who did not recall whether he purchased more than 1 battery, for example, might
choose a response greater than 1 because the response options suggest that is what the correct
answer is likely to be. The use of closed-ended questions is especially problematic when the
list does not include responses such as "I don't know" or "I don't remember," or "None of the
above." DCWP's surveys use many such inappropriate closed-ended questions, rendering the
responses unreliable. 23
18
Simonson Report, 9i[ 38.
19
Simonson Report, 9i[ 23.
20
Simonson Report, 9i[ 38.
21
Simonson Report, 9i[ 35.
22
Simonson Report, 9i[ 36.
23
While some survey questions allowed respondents to skip providing an answer, most
questions required an answer to allow the respondent to continue. See NYC Delivery Worker
Survey Questionnaire.
Page 5
FILED: NEW YORK COUNTY CLERK 07/06/2023 11:59 AM INDEX NO. 155947/2023
NYSCEF DOC. NO. 29 RECEIVED NYSCEF: 07/06/2023
18. Flaw #3: DCWP's surveys failed to include control (or "phantom'') questions, or any
other controls to account for guessing and demand effects. The use of control questions is
well-accepted by survey researchers as an effective method to minimize bias in survey
responses. As the Simonson Report points out, an appropriate way to control for respondents'
guessing about expenses they incurred would have been to include expense categories not
related to food delivery work. 24
19. Flaw #4: DCWP's surveys Jailed to validate survey responses. While in some cases
it may be difficult or impossible to validate survey responses, the Simonson Report particularly
identifies that survey respondents could have been asked for purchase receipts as proof of
claimed expenses. 25
V. The Flaws of The NYC Delivery Worker Survey Are Not Resolved Based on
DCWP's Responses and, Accordingly, the Survey Remains Unreliable.
20. With respect to Flaw #1, DCWP responded as follows:
It would not have been appropriate to conduct a survey without informing
respondents that it was being conducted by the City of New York or informing
respondents how their responses would be used. Such disclosures, which are
customary, do not invalidate the survey results. Had the Department not
provided appropriate disclosure, it is likely that participation would have been
lower and less representative. 26
21. DCWP's response does not resolve Flaw #1 because it ignores the well-established
result that revealing the sponsor and purpose of a survey can generate demand effects fatal to
the survey' s reliability. 27 Even if an initial disclosure is customary in other surveys conducted
by DCWP (something DCWP seems to claim without offering support), being "customary" is
not an indicator ofreliability or scientific validity-especially when there is academic research
to the contrary. 28 More importantly, DCWP did not simply inform respondents of who was
conducting the survey and how their answers would be used. Rather, DCWP told respondents
24
Simonson Report, ,r 39.
25
Simonson Report, ,r 42e.
26
Notice of Adoption of Final Rule, p.
27
See Simonson Report, ,r 13, citing Diamond, S. S., "Reference Guide on Survey Research,"
in Reference Manual on Scientific Evidence, Third Edition, Washington, D.C.: The National
Academies Press, Federal Judicial Center, 2011, 359--424 ("Diamond (2012))," p. 411.
28
Diamond (2012), p. 411; Diamond S.S and Jerre B. Swann, "Trademark and Deceptive
Advertising Surveys," Second Edition, American Bar Association, p. 212.
Page 6
FILED: NEW YORK COUNTY CLERK 07/06/2023 11:59 AM INDEX NO. 155947/2023
NYSCEF DOC. NO. 29 RECEIVED NYSCEF: 07/06/2023
that the survey's ultimate goal would be to raise pay for app delivery workers. 29 This prompt,
which appears at the beginning of the NYC Delivery Worker Survey, invalidated the survey
results in at least two ways: it biased the survey by leading the responses of those workers who
participated in the survey ("Answer Bias"), and it biased the survey by affecting the propensity
of a worker to participate in the survey ("Participation Bias").
22. Answer Bias: As described in the Simonson Report, a great deal of academic research
finds that self-interest can drive survey respondents to report incorrect information in their
responses if they believe they stand to gain something from doing so. 30 In this case, the prompt
provided at the beginning of the survey makes it clear how respondents stand to gain a higher
minimwn pay rate from providing higher estimates of their expenses. At the same time,
because respondents were told that their responses would be kept confidential, 31 there was a
negligible risk that they could suffer negative consequences from providing an incorrect
estimate. The combination of these two factors should encourage respondents to provide
exaggerated estimates of the size and frequency of expenses and biases the survey.
23. Importantly, note that the answer bias caused by the initial survey prompt affected
responses to the entire survey, regardless of whether a question asked for a monetary amount
or some other estimate. For example, DCWP stated that it "used the survey to measure the
frequency with which workers experience loss or theft of their e-bike, purchase replacement
batteries or e-bike accessories, and buy and trade-in phones." 32 A respondent that stood to gain
from a higher minimum pay rate would have incentives to report more frequent purchase of
batteries, accessories, or phones because there would be a direct connection between the
numbers he or she reported and the estimate of expenses that DCWP would derive from their
responses (because DCWP told respondents as much).
24. Participation Bias: Researchers who use surveys should strive to ensure their survey
samples are representative- the survey sample should accurately reflect the relevant
characteristics of the target population. 33 When a survey includes responses from only a part
of the selected sample, while systematically excluding another part of the sample, it is said that
29
NYC Delivery Worker Survey Questionnaire.
30
See Simonson Report, ,r 15.
31
NYC Delivery Worker Survey Questionnaire.
32
"Notice of Public Hearing and Opportunity to Comment on Proposed Rules," New York
City Department of Consumer and Worker Protection.
33
Diamond (2011), p. 380.
Page 7
FILED: NEW YORK COUNTY CLERK 07/06/2023 11:59 AM INDEX NO. 155947/2023
NYSCEF DOC. NO. 29 RECEIVED NYSCEF: 07/06/2023
the survey suffers from participation or "non-response" bias."34 When a survey suffers from
participation bias it is difficult to draw valid inferences about the population from the survey
sample. 35
25. Academic research shows that respondents who are aware that they may gain from a
survey's result are more likely to participate in the survey than are respondents who are not
aware of such gains.36 In this case, the initial survey prompt tells respondents how they stand
to benefit from the outcome of the survey. Consequently, delivery workers who actually had
higher or more frequent expenses would have more incentive to participate in the survey
because they stood to gain more from a minimum pay rate that reflects hlgher expenses, relative
to delivery workers with lower or less frequent expenses. This is a fatal flaw because it
systematically increases the propensity of certain workers to participate in the survey (i. e.,
those with higher expenses), whlle it decreases the propensity of other workers to do so (i. e.,
those with lower expenses). DCWP' s inability to validate the representativeness of the NYC
Delivery Worker Survey is a severe flaw.
26. While DCWP claims that it "used appropriate controls to ... address non-response
37
bias," the controls that DCWP used do not address the incentives of delivery workers to
participate in the survey,38 and thus cannot correct the biases resulting from an unrepresentative
sample. Moreover, because respondents with hlgher or more frequent expenses are more likely
to want to participate in the survey to influence a pay rate that "reflects [their] expenses and
needs,"39 the estimate of average expenses drawn from the survey's unrepresentative sample is
likely to overestimate the true average expenses of delivery workers.
27. In the Notice of Adoption of Final Rule, DCWP claims that, had it not provided
disclosures about the surveys' use, "it is likely that participation would have been lower and
34
Diamond (2011), p. 383 .
35
Diamond (2011), p. 383.
36
Amany Saleh and Krishna Bista (2017), "Examining Factors Impacting Online Survey
Response Rates in Educational Research: Perceptions of Graduate Students," Journal of
Multidisciplinary Evaluation, 13-29, pp. 63-74; Charles L. Martin (1994), "The impact of
topic interest on mail survey response behavior," Journal of the Market Research Society, 36-
4.
37
"Notice of Public Hearing and Opportunity to Comment on Proposed Rules," New York
City Department of Consumer and Worker Protection.
38
DCWP applied post-stratification weights based on the mix of apps used by the worker
(Uber Eats only, DoorDash only, Grubhub only, Relay only, and multiple) and the quartile of
hours worked. See "A Minimum Pay Rate for App-Based Restaurant Delivery Workers in
NYC," New York City Department of Consumer and Worker Protection Report, p. 5.
39
NYC Delivery Worker Survey Questionnaire.
Page 8
FILED: NEW YORK COUNTY CLERK 07/06/2023 11:59 AM INDEX NO. 155947/2023
NYSCEF DOC. NO. 29 RECEIVED NYSCEF: 07/06/2023
less representative.'"' 0 However, if DCWP was concerned about low participation in the
surveys, there are other, well-established methods, that can increase response rates while
maintaining the survey's representativeness. Examples include providing financial or
nonmonetary incentives. 41 Informing respondents that they stand to benefit from answering
survey questions in a specific way is not an appropriate method to address DCWP's
participation concerns because it biases the survey. 42
28. With respect to Flaw #2, DCWP responded as follows:
[T]he Department's decision to ask workers about whether they purchased
specific accessories, as opposed to an open-ended question about expenses,
followed from initial testing with delivery workers in which respondents had
difficulty recalling the accessories they purchased without prompting. Had the
Department adopted commenter's recommendation, it would have led to an
under-estimate of accessory expense.
[T]he Department did not use any responses in which a respondent was asked
to report a monetary amount in its calculation of the minimum pay rate.
[F]ield sw·veyors on [a separate survey] reported that delivery workers
experienced a high level of difficulty completing free response format or
numeric input questions, leading the Department to determine that close-ended
responses were the most appropriate format for this population and essential in
keeping the voluntary survey short and cognitively undemanding, which
increases survey completion and representativeness. 43
29. DCWP's response does not resolve Flaw #2 because it is likely that the DCWP's use
of closed-ended questions led to guessing and bias. For example, as explained by Prof.
Simonson and as I described in ,r 17 above, the closed-ended format of the question regarding
the number of batteries that survey respondents had purchased for their mopeds is likely to lead
to biased and exaggerated estimates of the number purchased. 44
30. Notably, DCWP's response hinges on the claim that survey respondents had difficulty
answering free response format or numeric input questions. Difficulty answering certain
40
Notice of Adoption of Final Rule," p. 24.
41
Diamond (2011), p. 383.
42
DCWP also noted that the participation rate of its survey was "several times the rate
obtained by leading academic researchers." See Minimum Pay Report, p. 3. Thus, a
potential decrease in overall participation in exchange for obtaining a representative,
unbiased sample would not have been necessarily a concern.
43
Notice of Adoption of Final Rule, pp. 24-25 .
44
See, e.g., Simonson Report, ,r 30.
Page 9
FILED: NEW YORK COUNTY CLERK 07/06/2023 11:59 AM INDEX NO. 155947/2023
NYSCEF DOC. NO. 29 RECEIVED NYSCEF: 07/06/2023
numeric input questions (e.g., "how many cell phones did you buy last year?") is more likely
caused by the inability of respondents to remember accurately than by the question format
(open-ended or closed-ended). Unlike a computer hard drive, that records and stores
information, human memory can be more fallible. Rather than simply being retrieved, and free
from error, a great deal of research shows that in many cases memory judgments are
constructed on the spot. 45 When asked a question like "In the past twelve months, have you
bought any delivery accessories like: GPS tracker, handle bar gloves, winter gear, rain gear,
rack, food bag, basket, bungee cords, helmet, lights, horn, reflective vest, water bottle, lock,
phone holster, etc.?" for example, respondents cannot simply reach into memory and pull out
an exact list, because such information does not exist. Consequently, they must construct a
reasonable response based on available information to the best of their ability.
31. Research shows that providing closed-ended answer options can reduce accuracy
because it biases respondents to select an option even when they do not have an accurate
recollection. 46 Thus, rather than achieving higher-quality responses, the closed-ended format
of the NYC Delivery Worker Survey incentivized guessing. This issue is more prominent in
this case because, contrary to best practices of survey design,47 the NYC Delivery Worker
Survey did not allow respondents to express that they did not know or did not remember a
particular answer. 48
32. While DCWP claims that it did not use any responses in which a respondent was asked
to report a monetary amount in its calculation of the minimum pay rate, many of the inputs that
DCWP acknowledges it did rely on are affected by the inappropriate use of closed-ended
questions. For example, the questions that DCWP relies on to calculate the "[ f]requency with
which workers purchase replacement batteries or phones, or trade-in phones," the "[s]hare of
respondents purchasing each e-bike accessory;" or the"[ a ]verage phone purchase and/or trade-
45
Elizabeth F. Loftus and John C. Palmer (1974), "Recon Reconstruction of automobile
destruction: An example of the interaction between language and memory," Journal of verbal
learning and verbal behavior, 13-5, pp. 585-589; Daniel L. Schacter (2012), "Constructive
Memory: past and future," Dialogues in Clinical Neuroscience, 14-1, pp. 8- 18;
Daniel L. Schacter et al. (1998), "The Cognitive Neuroscience of Constructive Memory,"
Annual Review of Psychology, 49, pp. 289-318.
46
Diamond (2012), p. 393; Reja et al., "Open-ended vs. Close-ended Questions in Web
Questionnaires," Developments in Applied Statistics, 2003, p. 163.
47
Diamond (2012), pp. 389-390
48
NYC Delivery Worker Survey Questionnaire.
Page 10
FILED: NEW YORK COUNTY CLERK 07/06/2023 11:59 AM INDEX NO. 155947/2023
NYSCEF DOC. NO. 29 RECEIVED NYSCEF: 07/06/2023
in price,"49 inappropriately led respondents to provide an answer even if they do not remember
and suggest to respondents the expected range of their answers.
33. Moreover, as explained in ,r 22 above, survey instructions that trigger self-interest
motives can encourage responses that promote self-interest. 50 Returning to the survey question
about purchases of delivery accessories, when trying to estimate a response to that question,
knowing that their responses will help set a minimum pay rate should encourage respondents
to provide higher estimates.
34. Finally, even if DCWP does not claim that it relied on other questions for the
calculation of the minimum pay rate, the responses to one question can be affected by responses
to previous questions in the same survey.51 The Simonson Report highlights multiple examples
of questions throughout the NYC Delivery Worker Survey that are affected by the
inappropriate use of closed-ended questions, by focalism bias, and by leading language. 52
DC WP cannot appropriately claim without proof that just because it did not rely on a particular
question, the presence of flawed questions in the survey did not invalidate later responses in
the same survey.
35. With respect to Flaw #3, DCWP responded as follows:
With respect to "phantom questions," respondents to the NYC Delivery
Worker Survey reported purchasing some items at very low rates (e.g. , anti-
theft camera at 13 percent), putting a low upper bound on the frequency with
which respondents may have reported purchasing an item for delivery work
that they did not in fact purchase for that purpose. The use of"phantom
questions" is not customary in government surveys and the Department's
choice not to include them in the NYC Delivery Worker Survey does not
invalidate its results.
36. DCWP seems to argue that phantom questions are not necessary because the survey
provides a low upper bound for the frequency of certain purchases claimed by the survey
respondents. However, DCWP provided no evidence that the results from the survey are
indeed low. For example, it is difficult to validate if a 13 percent purchase rate for anti-theft
cameras is indeed a low upper bound unless one could compare this number to an external
49
Notice of Adoption of Final Rule, pp. 23-24.
50
See, e.g., Dale T. Miller (1999), "The Nonn of Self-Interest," American Psychologist, 54-
12, pp. 1053-1060; Terence A. Shimp et al. (1991), "A Critical Appraisal of Demand Artifacts
in Consumer Research," Journal of Consumer Research, 18-3, pp. 273- 283.
51
Diamond (2011), p 395.
52
Simonson Report, ,r,r 34, 36-40.
Page 11
FILED: NEW YORK COUNTY CLERK 07/06/2023 11:59 AM INDEX NO. 155947/2023
NYSCEF DOC. NO. 29 RECEIVED NYSCEF: 07/06/2023
benchmark. Even ifDCWP claims (without offering support) that phantom questions are not
customary in government surveys, it is considered best practice for researchers to try to
demonstrate the validity of their results by comparing them to external metrics, or by relying
on other strategies to correct any potential biases in the results. 53 Indeed, the City of New York
has relied on such strategies in other survey studies. 54 DCWP, however, offers no evidence of
the external validity of the NYC Delivery Worker Survey.
37. Finally, in the previous paragraphs, I explained why DCWP's responses to criticisms
included in the Simonson Report do not adequately resolve Flaws #1, 2, and 3 of the Delivery
Worker Survey and that, therefore, the Survey remains unreliable. I note that DCWP provided
no response to Flaw #4, which is the Simonson Report's assertion that there was no attempt to
verify certain responses to questions, for example, by asking for receipts to prove claimed
purchases.
38. In addition, I understand that data on hours worked and income generated was provided
to DCWP by DoorDash, Grubhub, and other delivery companies. 55 While DCWP seems to
claim that it did not rely on any survey answers relating to these variables, 56 the data provided
by DoorDash and Grubhub could have been used to test the validity of at least some of the
answers of the NYC Delivery Worker Survey. I have seen no evidence that DCWP attempted
such validation.
53
Elizabeth A. McCune and Sarah R. Johnson, "Chapter 11: How Did We Do? Survey
Benchmarking and Normative Data" in Employee Surveys and Sensing, Ed. William H.
Macey and Alexis A. Fink, Oxford University Press, 2020; Justin A. DeSimone, et al., Best
practice recommendations for data screening (2015), Management Department Faculty
Publications, University of Nebraska, Paper 124.
54
See, e.g., "Why New York Hires 200 People to Pretend They're Homeless," The New York
Times, January 19, 2018.
55
Minimum Pay Rate Report, p. 2.
56
Notice of Adoption of Final Rule, pp. 23-24.
Page 12
FILED: NEW YORK COUNTY CLERK 07/06/2023 11:59 AM INDEX NO. 155947/2023
NYSCEF DOC. NO. 29 RECEIVED NYSCEF: 07/06/2023
App_endixA
Jonah Berger
The Wharton School• University of Pennsylvania
jberger@wharton.upenn.edu
Academic Positions
The Wharton School, University of Pennsylvania
Associate Professor of Marketing (with tenure) May 2013-
James G. Campbell, Jr. Assistant Professor of Marketing July 2010-May 2013
Assistant Professor of Marketing July 2007 - June 2010
Cornell NYC Tech, Cornell University
Visiting Professor of Marketing July 2014-- June 2015
Fuqua School of Business, Duke University
Visiting Associate Professor of Marketing July 2013-Dec 2013
Education
Ph.D., Marketing, Stanford University, Graduate School of Business, 2007
B.A., Human Judgment and Decision Making (with Distinction), Stanford University, 2002
Honors and Awards
Sage 10-Year Impact Award (1 of the 3 most cited articles in all Sage journals in 2012), 2023
MSI Scholar, 2020
AMA-Sheth Foundation Doctoral Consortium Fellow, 2019, 2020, 2021, 2022
Wharton Teaching Excellence Award 2019, 2020, 2021
William F. O'Dell Award, Journal of Marketing Research, 2017
Top 5 Most Productive Researchers in Marketing, AMA DocSig 2017-
Outstanding Reviewer Award, Journal of Consumer Research 2015-2016
Best 2012 Article Finalist, Journal of Consumer Research, 2015
Top 30 Leaders in Business, American Management Association, 2015
Emerald Citations of Excellence, article published in 2012, 2015
Berry-AMA Book Prize for Best Book in Marketing, 2014
Top 5 Most Productive Researchers in Marketing 2009-13, AMA Doc Sig 2013
Most Creative People in Business, Fast Company, 2013
Paul Green Award, Journal ofMarketing Research, Finalist 2013
Early Career Award, Association for Consumer Research, 2013
Early Career Award, Society for Consumer Psychology, 2012
Dean's Research Grant, The Wharton School, 2012
Outstanding Reviewer Award, Journal of Consumer Research 2010-2011
Outstanding Reviewer Award, Journal of Consumer Psychology, 2010-2011
"Iron Prof' Award for "awesome faculty research," The Wharton School, 2011
MBA Teaching Commitment and Curricular Innovation Award, The Wharton School, 2011
Dean's Research Grant, The Wharton School, 2011
Young Scholars Program, Marketing Science Institute, 2011
Alex Panos Research Grant, The Wharton School, 2011 ,.
Journal of Consumer Research Best 2007 Article Award Finalist, 2010
James G. Campbell, Jr. Memorial Term Professorship, 2010
Dean's Research Grant, The Wharton School, 2010
Page 1
FILED: NEW YORK COUNTY CLERK 07/06/2023 11:59 AM INDEX NO. 155947/2023
NYSCEF DOC. NO. 29 RECEIVED NYSCEF: 07/06/2023
Appendix A
AMA-Sheth Foundation Doctoral Consortium Fellow, 2006
Society for Consumer Psychology, Best Student Paper Award (Honorable Mention), 2006
Management Science Institute/JCP Research Competition (Honorable Mention), 2004
Jaedeke Scholar, Stanford Graduate School of Business, 2003
Publications
1. Packard, Grant and Jonah Berger (2023) "The Emergence and Evolution of Consumer Language
Research," forthcoming, Jownal of Consumer Research.
2. Packard, Grant, Jonah Berger, and Reihane Boghrati (2023) "How Verb Tense Shapes Persuasion,"
Journal of Consumer Research.
3. Giovanni Luca Cascio-Rizzo, Jonah Berger, Rumen Pozharliev, and Matteo De Angelis (2023),
"How Sensory Language Shapes Consumer Responses to Influencer-Sponsored Content,"
forthcoming, Journal of Consumer Research .
4. Berger, Jonah, Wendy Moe, and David Schweidel (2023), "What Holds Attention? Linguistic Drivers
of Engagement," Journal of Marketing.
5. Oba, Demi and Jonah Berger (2023), "How Communication Mediums Shape the Message,"
Conditionally Accepted.
6. Boghrati, Reihane, Jonah Berger, and Grant Packard (2023), "Style, Content, and the Success of
Ideas," Journal of Consumer Psychology.
7. Boghrati, Reihane and Jonah Berger, "Quantifying Cultural Change: Gender Bias in Music,"
forthcoming, Journal of Experimental Psychology: General.
8. Berger, Jonah, Joshua Conrad Jackson, and Ceren Kolsarici (2023), "Catalyzing Social Change: Does
Concentration Encourage Action?" forthcoming.
9. Weingarten, Evan, and Jonah Berger (2023), "Discussing Proximal Pasts and Far Futures," Journal of
Consumer Psychology.
10. Packard, Grant and Jonah Berger (2023), "Wisdom from Words: The Psychology of Consumer
Language," Consumer Psychology Review, 6(1), 3-16.
Lead Article
11. Rogers, Benjamin A., Herrison Chicas, John Michael Kelly, Emily Kubin, Michael S. Christian,
Frank J. Kachanoff, Jonah Berger, Curtis Puryear, Dan P. McAdams, Kurt Gray (2023), "Seeing Your
Life Story As a Hero 's Journey Increases Meaning in Life," Journal of Personality and Social
Psychology.
12. Berger, Jonah, Matt Rocklage, and Grant Packard (2022) "Expression Modalities: How Speaking
Versus Writing Shapes Word of Mouth," Journal of Consumer Resea