Robo-advisor acceptance: Do gender and generation matter?

Robo-advice technology refers to services offered by a virtual financial advisor based on artificial intelligence. Research on the application of robo-advice technology already highlights the potential benefit in terms of financial inclusion. We analyze the process for adopting robo-advice through the technology acceptance model (TAM), focusing on a highly educated sample and exploring generational and gender differences. We find no significant gender difference in the causality links with adoption, although some structural differences still arise between male and female groups. Further, we find evidence that generational cohorts affect the path to future adoption of robo-advice technology. Indeed, the ease of use is the factor which triggers the adoption by Generation Z and Generation Y, whereas the perceived usefulness of robo-advice technology is the key factor driving Generation X+, who need to understand the ultimate purpose of a robo-advice technology tool before adopting it. Overall, the above findings may reflect that, while gender differences are wiped out in a highly educated population, generation effects still matter in the adoption of a robo-advice technology tool.

1. Introduction

A robo-advisor is a virtual financial advisor, based on artificial intelligence, that is currently in the spotlight in wealth management [1-4], propelling leading innovations in the financial industry. According to the European Securities and Markets Authority (ESMA) 2018 guidelines, “Robo-advice means the provision of investment advice or portfolio management services (in whole or in part) through an automated or semi-automated system used as a client-facing tool” [5].

According to Niszczota and Kaszás [6], the main obstacle to the adoption of robo-advice technology is the perception of humans as more effective than algorithms when the decision process requires subjective judgments. Rasiwala and Kohli [7], by contrast, evidence that robo-advice technology is gaining popularity, as it may reduce or eliminate the risk of human errors. Transparency, security and simplicity of processes are considered relevant factors for any information technology artifact and may lead to increased customer satisfaction and willingness to adopt robo-advice services specifically. The expectation of financial inclusion related to Fintech is high [8, 9], and Bianchi and Briere [10] confirm the potential of robo-advice technology in this direction, suggesting that the human–robot interaction can improve financial capability. To the best of our knowledge, few researchers have analyzed factors that might influence the adoption of robo-advice by end users [6, 11–15]. Specifically, Hohenberger et al. [12] examine how financial experience influences the intention to use a robo-advisor and how emotional reactions, such as anxiety and joy, can mediate this effect. They find that individuals’ behavioral intention increases when positive emotions are expected from using a robo-advisor but decreases otherwise; moreover, this relationship may change, according to individuals’ self-enhancement motives (e.g. possibility of accumulating wealth).

Cheng et al. [14] evidence the significant influencing role of supervisory control and validate the relationships between trust influencing factors: trust in technologies, trust in vendor and trust in a robo-advisor.

Finally, Lourenço et al. [15] focus on the impact of firm characteristics on consumer acceptance of pension investment advice generated by a robo-advisor. They find that consumers’ perceptions of trust and expertise of the firm providing the automated advice are important drivers of advice acceptance. Hohenberger et al. [12], Cheng et al. [14] and Lorenço et al. [15] measure the impact of the proposed adoption drivers through a structural equation model (SEM), in line with the technology acceptance methodology (TAM) by Davis [16]. Indeed, the TAM is possibly the most widely used framework in the field of information systems for measuring and forecasting technology adoption [16] and has proved empirically useful in several studies [17, 18].

With respect to technology adoption, several studies in many fields have documented the relevance of generational cohorts and gender (e.g. [19-24]). Generational cohorts have a different culture and approach to technology in general, and this effect can be exacerbated when technology is related to possible profits and losses. Further, women are often proved to be less self-confident (see, among others, [25]), and this may impact their attitude. In this regard, Chen et al. [24] find a significant gender gap in the use of Fintech services.

The impact of generational cohorts/age and gender for robo-advisory application is investigated in D’Acunto et al. [26], Lourenço et al. [15] and Cheng et al. [24], with partially conflicting results. D’Acunto et al. [26] evidence no demographic differences between users and non-users of robo-advice; Cheng et al. [14] include gender and age groups among control variables to specifically investigate trust transfer in robo-advice and show that they have insignificant effects in their research model. Conversely, Lourenço et al. [15] evidence that age has a negative and significant influence, such that older individuals and women are less satisfied with using automated tools than younger individuals. Interestingly, Lourenço et al. [15] also evidence a positive impact of education level on robo-advice acceptance.

We enter the current debate by investigating the moderation effect of gender and generational cohorts on the TAM model, applied to the robo-advisor technology, by focusing on a homogeneous, highly educated sample, to avoid possible biases related to investor education level. Indeed, generational and gender characteristics may have different effects in a highly educated framework.

The main question we address in this paper is whether there is a significant difference in the factors affecting the acceptance of the robo-advice technology across the selected generational groups and between genders. Specifically, we survey students and employees (faculty and staff) of the University of Perugia, Italy; all respondents are of working age. Hence, elders are not included; thus, the effect of biological aging is beyond the scope of this analysis.

The paper is organized as follows: section 2 describes the research methods and states our hypotheses, section 3 describes how data are collected and treated and section 4 summarizes the empirical results. Finally, section 5 discusses the implications of the research outcomes and offers some concluding remarks.

2. Research methods and hypotheses

The TAM [16] is a path model that links specific constructs (latent variables) through causal relationships to be estimated in a SEM setting. According to the original TAM model, represented in Fig 1, the variables linked to the acceptance of any technology and its adoption are perceived ease of use (PEU), perceived usefulness (PU), attitude towards use (ATU) and behavioral intention to use (BIU).

Fig 1

The technology acceptance model.

Davis [16] defines perceived usefulness as “the prospective user’s subjective probability that using a specific application system will increase his or her job performance within an organizational context” and perceived ease of use as “the degree to which the prospective user expects the target system to be free of effort.”

According to the TAM, behavioral intention to use refers to the prospective use of a given information system and thus determines technology acceptance. Attitude towards use and perceived usefulness jointly influence behavioral intention, which is indirectly affected by perceived ease of use as well. Attitude towards use is directly affected by both perceived usefulness and perceived ease of use, and perceived usefulness is directly influenced by perceived ease of use. Further, according to the TAM, perceived usefulness and perceived ease of use may be affected by external variables (EXT). In turn, both perceived usefulness and perceived ease of use mediate the effect of external variables on the user’s attitude towards use and behavioral intention to use (Fig 1).

The purpose of this paper is to apply the TAM to the adoption of a robo-advisor service, by analyzing whether the propensity to use this technology varies across gender or generational groups. In this study, perceived ease of use is defined as the degree to which a respondent believes that learning to use a robo-advisor requires a relatively low degree of effort, and perceived usefulness is defined as the degree to which a respondent believes that using a robo-advisor would enhance their investment performance. The definition of the other variables is intuitive: attitude is the individual’s positive or negative feelings about a behavior, and behavioral intention is the individual’s perceived probability that they will use the system, here the robo-advisor.

Moreover, we introduce a new construct, the personal investment approach (PIA), to represent the propensity of individuals to actively manage their portfolio, and we assume that this construct also influences the technology acceptance of robo-advice as an external variable (EXT = PIA).

Specifically, we test whether and to what extent the TAM links are moderated by generational cohorts or by gender.

Moderation analysis can be performed alternatively by using interaction terms or by splitting the sample into sub-groups of interest before estimation (so-called multi-group analysis [MGA]). We adopt the latter approach, which permits assessing both the magnitude of TAM path relationships within sub-groups and the significance of their differences. The hypotheses tested in the paper are summed in Table 1.

Table 1

Hypotheses under investigation.

H1		The perceived ease of use of the robo-advisor positively affects its perceived usefulness
	a.	The effect is moderated by the generation cohort
	b.	The effect is moderated by the gender
H2		The perceived usefulness of the robo-advisor positively affects the attitude towards its use.
	a.	The effect is moderated by the generation cohort
	b.	The effect is moderated by the gender
H3		The perceived ease of use of the robo-advisor positively affects the attitude towards its use.
	a.	The effect is moderated by the generation cohort
	b.	The effect is moderated by the gender
H4		The attitude towards use of the robo-advisor positively affects the behavioral intention of its adoption.
	a.	The effect is moderated by the generation cohort
	b.	The effect is moderated by the gender
H5		The personal investment approach positively affects the perceived ease of use of the robo-advisor.
	a.	The effect is moderated by the generation cohort
	b.	The effect is moderated by the gender
H6		The personal investment approach positively affects the perceived usefulness of the robo-advisor.
	a.	The effect is moderated by the generation cohort
	b.	The effect is moderated by the gender

3. Data collection, measures and procedure

We designed an online questionnaire using SurveyMonkey (https://www.surveymonkey.com/) and submitted the questionnaire link to students, faculty and staff of the University of Perugia in October 2019. Overall, 214 respondents completed the questionnaire and are considered a convenience sample for the following analysis. The response rate for the complete questionnaire is around 5% for University faculty and staff and 0.45% for University students (around 25,000 enrolled in 2019/2020). Section 1 of the questionnaire (12 questions) allows us to establish the sample’s descriptive characteristics, summarized in Table 2.

Table 2

Questionnaire: Section 1, descriptive statistics.

Attributes	Frequency (N = 214)	Percentage
Gender
Male	79	36.9
Female	134	62.6
Other	1	0.5
Generational group
Generation Z	102	47.7
Generation Y	42	19.6
Generation X+	70	32.7
Education
High school	90	42.1
Bachelor’s degree	81	37.9
Graduate studies	43	20.1
Occupation
Student	102	47.7
Other	112	52.3
Financial literacy
No	85	39.7
Intermediate	119	55.6
High	10	4.7
Confidence in the evaluation of financial offers
No, I trust the proposal	34	15.9
No, I am not confident but I try to obtain external information	135	63.1
Yes, I am usually able to evaluate	45	21.0
Use of internet-based information
Often	101	47.2
Seldom	63	29.4
Never	50	23.4
Investment horizon (years)
<1	88	41.1
3	49	22.9
5	45	21.0
10	20	9.3
>10	12	5.6
Trusted advisors support
Yes	111	51.9
No	101	47.2
Previous use of robo-advice
No	210	98.1
Yes	4	1.9
Interest in robo-advice
No	142	66.4
Yes	72	33.6
Cyber-risk aversion
No, I trust	16	7.5
Yes, that’s why I am not interested	79	36.9
Yes, but I would use it anyway	35	16.4
I have no idea	84	39.3

The sample is composed of highly educated individuals. Most respondents hold at least a bachelor’s degree, of whom 20% also have a post-graduate diploma. Further, those having only a high school degree are mainly students currently pursuing their bachelor’s. Overall, the sample fits well the aim and context of this study.

Only one respondent does not disclose information about gender; 63% are females and 37% are males. Respondent ages range from 18 to 67 years, grouped by generational cohort: Generation Z (born 1994–2012; 47%), Y (born 1980–93; 19.6%) and X+ (born before 1980; 32.7%). A generational cohort is a group of individuals identified by common social and historical life events [27]; however, the definition is usually loose and has been questioned by Rudolph and Zacher [28]. For the purposes of proper contextualization with mainstream literature, we cautiously use generational groupings consistent with those defined in [21, 29, 30] while acknowledging the theoretical limitation highlighted in [28].

After asking for basic demographic information, the survey investigates respondents’ level of financial literacy and awareness, familiarity with the internet and willingness to use innovative technologies such as robo-advice.

With respect to financial literacy, most respondents (60.3%) report having an intermediate or high level of financial education. Notably, only 16% declare a low ability to evaluate investment proposals, and 21% declare a high ability to evaluate investment choices, in line with the overconfidence of investors evidenced in behavioral finance studies (e.g. [31-33]). In addition, nearly half the respondents report using the internet often to find information necessary to making financial decisions. Although most have a long life expectancy, they plan their investments within a very short time horizon (less than 1 year for 41.1% of respondents), losing the opportunities offered by true investment planning, which usually takes advantage of a longer investment horizon.

With respect to the use of robo-advice, only 2% declare having previously used such technology, as expected. Moreover, only a minority are interested in future use of robo-advice, in line with other findings [34]. Not surprisingly, of the 72 respondents declaring interest in this technology, 58.3% belong to Generation Z. In addition, interested respondents declare that they would use robo-advice only for small amounts (45.1%), for low-fee services (29.6%) or on well-known investment platforms (22.5%).

The low interest in robo-advice is partially motivated by respondents’ cyber-risk aversion: only 7.5% fully trust algorithms and internet-based platforms, and only 16.4% show enough risk tolerance to pursue using internet-based solutions.

Note that the proportion of respondents interested in the future application of robo-advice is not conditioned by their previous reliance on a trusted professional (interested: all sample = 33.6%; interested: no advisor = 33.7%; interested: with advisor = 32.4%).

The second and core part of the questionnaire includes five questions with 7-point Likert-scale responses to measure the TAM constructs used in this study. Likert-scale levels indicate respondents’ agreement or disagreement with the proposed assertions (1 = Strongly agree; 7 = Strongly disagree). The complete questionnaire is reported in Appendix A in S1 Appendix.

Before estimating the TAM links, we assess the validity of the measurement model through suitable statistics such as composite reliability (CR) and average variance extracted (AVE), as recommended in [35, 36], among others. Indeed, latent constructs are sufficiently reliable and display a quite satisfactory AVE, slightly below the desired level for the PIA variable.

In addition, no collinearity issues are evidenced, since the variance inflation factor (VIF) is far below the conservative 3.00 threshold for all latent constructs and across sub-groups.

Following Hanseler et al. [37] and Mahmoud et al. [21], among others, we also apply the measurement invariance of the composite model (MICOM) procedure to the TAM constructs. The outcomes of a permutation test (1,000 permutations) show that the original correlations of the constructs, when segmented by gender or generational cohort, are not statistically different from 1, and that the null hypothesis of compositional invariance cannot be rejected with all p-values far above 5%. Details of this model validation assessment are summarized in Appendix B in S1 Appendix.

4. Empirical results

As a preliminary step, the TAM is estimated on the entire dataset, evidencing positive and significant links between the TAM latent constructs. A full description of this preliminary analysis is provided in Appendix C in S1 Appendix. In particular, we find that model outcomes do not change significantly in the whole sample by controlling for gender, whereas a moderation effect is evidenced for generational groups.

This finding further motivates the MGA analysis to investigate which and to what extent causality links differ between genders and between generational groups. Specifically, the parameter estimates are obtained via the Multi-Group SmartPLS routines [38], and the outcomes are illustrated in Figs 2 and 3, respectively, for the generational and gender groups. Path coefficients are reported for each of the hypotheses together with the p-value (in brackets) for the corresponding significance t-test.

Fig 2

Technology acceptance constructs and causality links segmented by generation.

Fig 3

Technology acceptance constructs and causality links segmented by gender.

Hypothesis H1 holds true across generational and gender groups: perceived ease of use (PEU) of robo-advice has a positive and significant impact on perceived usefulness (PU). Specifically, the path coefficients are 0.566, 0.587 and 0.263 respectively for Generations Z, Y and X+ and are strongly significant (p < 0.001) for the first two groups. As for gender, perceived ease of use (PEU) has a strongly significant impact on perceived usefulness (PU), with path coefficients of 0.448 and 0.424 respectively for male and female respondents.

Similarly, H2 is fully confirmed: the path coefficients from perceived usefulness (PU) to attitude towards use (ATU) are very high across all generation groups (0.560, 0.690, 0.743) and genders (0.821 and 0.635) and are strongly significant (p < 0.001).

The relationship in H3 between perceived ease of use (PEU) and attitude towards use (ATU) holds true for Generation Z, with a strongly significant path coefficient (0.298, p = 0.001); it is lower (0.229) and weakly significant (p = 0.078) for Generation Y and vanishes (0.067) for Generation X+ (p = 0.541). Further, the link is positive and significant for the female group (0.206 with a p-value of 0.014), but non-significant for the male group, evidencing that perceived ease of use may act as a driver in the adoption of a robo-advisor for Generation Z and females.

Hypothesis H4 is valid for all considered groups: attitude towards use (ATU) is positively linked to behavioral intention (BIU), with path coefficients of 0.673, 0.757 and 0.700 respectively for the three generational cohorts, and 0.705 and 0.695 for males and females. All parameters are strongly significant, with p-values lower than 0.001.

The effect in H5 of personal investment approach (PIA) on perceived ease of use (PEU) is strongly significant, except for Generation X+. Indeed, the magnitude of the parameter decreases across generational cohorts, with path coefficients of 0.509, 0.422 and 0.187 respectively for Generations Z, Y and X+. Links are positive and significant for both genders, with slightly higher values for the male group.

Finally, the impact of personal investment approach (PIA) on perceived usefulness (PU) H6 is confirmed to be positive for all generational groups (path coefficients of 0.315, 0.319 and 0.580) and both genders (path coefficients of 0.495 and 0.429). The results are strongly significant (p < 0.001) across all groups.

The Multi-Group SmartPLS routine includes statistical tests, based on the bootstrap algorithm provided by SmartPLS, to test whether the above differences between parameter estimates of different groups are significant, and hence to assess the presence and the magnitude of moderation effects. The outcomes of these tests are summarized in Table 3 for generational and gender groups; both the relative differences of the path coefficients and the corresponding statistical test p-values are displayed. Notably, no significant difference is detected for any path coefficient between Generation Y and Generation Z. Conversely, some causality links are significantly different between Generation X+ and the two other groups. Specifically, the effect of perceived ease of use on perceived usefulness is significantly lower for Generation X+, confirming the moderation effect of generation (H1a). In addition, the magnitude of the effect of personal investment approach (PIA) on perceived ease of use (PEU) is lower for Generation X+ than for either Generation Y or Generation Z, but only the latter is significant. Conversely, the effect of personal investment approach (PIA) on perceived usefulness (PU) is higher for Generation X+ with respect to the other two groups. Once again, the difference between Generations X+ and Z is significant, whereas the difference between Generations X+ and Y is only weakly significant.

Table 3

Outcomes of the tests between generational cohorts and gender groups.

Hypothesis		Gen Y vs Gen Z		Gen X + vs Gen Z		Gen X+ vs Gen Y		Females vs Males
		Rel. diff.	p-value	Rel. Diff.	p-value	Rel. Diff.	p-value	Rel. Diff.	p-value
H1	PEU → PU	0.022	0.831	−0.303	0.014	-0.324	0.041	−0.024	0.844
H2	PU → ATU	0.130	0.375	0.184	0.153	0.053	0.718	−0.186	0.090
H3	PEU → ATU	−0.069	0.656	−0.231	0.109	-0.162	0.347	0.144	0.259
H4	ATU → BIU	0.084	0.339	0.027	0.736	-0.058	0.584	-0.010	0.875
H5	PIA → PEU	−0.088	0.563	−0.332	0.041	-0.234	0.228	−0.137	0.297
H6	PIA → PU	−0.004	0.957	0.265	0.033	0.261	0.078	−0.066	0.555

Note. ATU = attitude towards use; BIU = behavioral intention to use; PEU = perceived ease of use; PIA = personal investment approach; PU = perceived usefulness.

Summing up, the outcomes in Table 3 evidence that causal relationships in H1, H5 and H6 are moderated by generational cohort, with a significant difference between Generations X+ and Z for H1, H5 and H6 and a significant difference between Generations X+ and Y for H1. Hence, only H1a, H5a and H6a hold true.

Interestingly, the path coefficients are not significantly different between the two gender groups, as reported in Table 3, except for the causality link from perceived usefulness to attitude towards use, for which the difference is negative and weakly significant. Hence, only H2b is weakly validated. Note also that although non-significant, the difference between the causality link perceived ease of use to attitude towards use is positive for the female and male groups and negative between the younger and older generation groups.

5. Discussion and concluding remarks

We apply the TAM to assess whether the causality links for robo-advisory adoption change significantly between generational groups and genders when focusing on a highly educated sample.

Results evidence that the strongest path towards the intention to use a robo-advisor is given by the link-chain perceived usefulness → attitude towards use → behavioral intention, and that the magnitude of the effect increases from Generation Z to Generation X+, evidencing how usefulness is the main driver for adoption in the latter case. Notably, the direct link from perceived ease of use and attitude towards use is significant only for Generation Z, which appears to be the only group driven by ease of use of the technology. The analysis also suggests a positive and significant inner link from perceived ease of use to perceived usefulness for all generational groups. The effect is stronger for Generations Z and Y and appears to be triggered by their perceived ease of use of robo-advisory.

As for the effect of the personal investment approach, we find interesting outcomes: the link between this external factor and the perceived ease of use of robo-advice is irrelevant for Generation X+ and significant for the other groups, with the highest value for Generation Z. We attribute this behavior to the fact that those young adults who actively manage their investment portfolios are most likely already acquainted with online trading platforms and find no hurdle in the use of robo-advice.

Conversely, the link from personal investment approach to perceived usefulness is strongly significant for all generations, with a large impact for Generation X+; indeed, a higher propensity in mature adults to actively manage the investment portfolio may increase the perception of the robo-advisory as a useful tool.

Looking at the whole picture (Fig 2) shows that the causality path to adoption of a robo-advisor is quite different for the three generational groups. Indeed, if we follow the thickest links in the graph, we find that Generations Z and Y are both driven towards adoption by the perceived ease of use construct. However, while the chain-link for Generation Y goes through the potential usefulness of the technology, Generation Z may be significantly driven towards adoption directly from ease of use, without considering its usefulness. Indeed, within this group, both the links from perceived ease of use to attitude towards use and to perceived usefulness are positive and strongly significant.

A different path sequence is highlighted for Generation X+. Interestingly, this generation moves towards adoption by completely ignoring the ease of use of the technology and paying attention to its usefulness only. Ease of use appears to worth less than usefulness, a finding perhaps explained by the greater experience and maturity of this generation compared with the other two.

Concerning gender, the highest path coefficient for males is related to the path from perceived usefulness to attitude towards use, whereas for females it corresponds to the path from attitude towards use to behavioral intention. Moreover, females seem to be partially driven to attitude towards use by perceived ease of use, a link which is instead irrelevant for males. However, the differences between path coefficients are non-significant for all the causality links, evidencing that gender does not moderate the adoption of robo-advisory within a highly educated sample. Though it may come as a surprising result, a clue in this direction could have been the proportion of female respondents, showing the interest of highly educated women on the topic of the survey.

Overall, the above findings may reflect that whereas gender differences are zeroed out in a highly educated population, generation effects still matter in the adoption of robo-advice. As a matter of fact, an intrinsic limitation of this study and those taking into account behavioral differences in generational cohorts, is the possibility that such variations may be also generated by other characteristics, such as age or status, of the corresponding groups [28]. Such a circumstance can be verified only by future studies. In our setting, one should analyze the behavior of generation groups towards new technology instruments supporting financial decisions in the next 10 to 30 years to disentangle whether their approach is triggered by age rather than by cohort [39].

Another limitation of our research is the small sample size, based on students, staff, and faculty of the University of Perugia (Italy), which is not perfectly balanced with respect to gender and generational groups and makes the results not widely generalizable.

Further, as usual in this kind of surveys, the behavioral intention can influence the answers to the questions measuring the antecedent variables, leading to a potential risk of endogeneity. However, our hope is that our research could still catch a positive signal for future generations, where gender is not relevant, facing financial innovation technology, starting from an equal education. Besides, our results may help stimulate the discussion on robo-advising. Indeed, to contribute the diffusion of such technology, providers should highlight the potential usefulness of the technology to mature customers. By contrast, our results suggest some concerns, which may be relevant for regulatory authorities, on the risks for Generation Z to adopt such technology just driven by its ease of use.

(DOCX)

Click here for additional data file.

(XLSX)

Click here for additional data file.

2 Jul 2021

PONE-D-21-14526

How does investor generational cohort make a difference in the acceptance of financial robo-advice?

PLOS ONE

Dear Dr. FIGA'-TALAMANCA,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

The manuscript would benefit from better engagement with relevant recent research on generational differences [1]. Further, please, follow JF Hair, JJ Risher, M Sarstedt and CM Ringle [2] guidelines to perform and report PLS-SEM results, especially in terms of the “Measurement Invariance of the Composite Models” (MICOM) procedure.

References

 

1. Mahmoud AB, Hack-Polay D, Grigoriou N, Mohr I, Fuxman L:  A generational investigation and sentiment and emotion analyses of female fashion brand users on Instagram in Sub-Saharan Africa .  Journal of Brand Management  2021.

2. Hair JF, Risher JJ, Sarstedt M, Ringle CM:  When to use and how to report the results of PLS-SEM .  European Business Review  2019,  31 (1):2-24.

Please submit your revised manuscript by Aug 16 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see:  http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols . Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at  https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols .

We look forward to receiving your revised manuscript.

Kind regards,

Ali B. Mahmoud, Ph.D.

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please include additional information regarding the survey or questionnaire used in the study and ensure that you have provided sufficient details that others could replicate the analyses. For instance, if you developed a questionnaire as part of this study and it is not under a copyright more restrictive than CC-BY, please include a copy, in both the original language and English, as Supporting Information.

3. Please consider changing the title so as to meet our title format requirement (https://journals.plos.org/plosone/s/submission-guidelines). In particular, the title should be "Specific, descriptive, concise, and comprehensible to readers outside the field" and in this case it is not informative and specific about your study's scope and methodology.

4. Please amend your list of authors on the manuscript to ensure that each author is linked to an affiliation. Authors’ affiliations should reflect the institution where the work was done (if authors moved subsequently, you can also list the new affiliation stating “current affiliation:….” as necessary).

5. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: I Don't Know

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Dear author! Thank you for submitting your manuscript. It is an interesting topic with some interesting findings. On the other hand, the manuscript has some issues. Below these issues are enclosed.

1) The first key issue is that the manuscript does not reasonably cover the literature on the context of Robo-advisor. One example is:

Cheng, X., Guo, F., Chen, J., Li, K., Zhang, Y., & Gao, P. (2019). Exploring the trust influencing mechanism of robo-advisor service: a mixed method approach. Sustainability, 11(18), 4917.

2)The second key issue is section number 4 (Discussion and concluding remarks). The manuscript should discuss the results (not only elaborate on the result). Several questions should be highlighted and recent related studies need to be highlighted. Moreover, the theoretical contribution should be more precise.

3) It is not clear why it is important to compare the “Digital Natives, Millennials and previous generations.”

The two studies, the manuscript, utilizes (Fulk et al., 2018; D'Acunto, 2019) did not find a significant difference in terms of age, Hence, it is critical to explain why it is important to do this comparison.

4) The manuscript argues that “Generation Y, or Millennials, which includes people born between 1980 and the early 2000s, are digitally advanced and potentially open to technology innovation in the financial sector, such as robo-advisors; however, one may argue that complete acceptance of such disruptive innovation would be achieved only by the younger Digital Natives”. It is difficult to know if this represents a subjective opinion or an assumption. Taken into consideration that the manuscript reported that “Notably, Digital Natives and Millennials show a very similar pattern, and no significant difference is detected between path coefficients.”

It is worth pointing out that several studies (e.g. Gomber et al., 2018) have reported that a high percentage of Millennials use FinTech business solutions. The manuscript should support the above argument with recent numbers, surveys, or/and studies.

Gomber, P., Kauffman, R. J., Parker, C., & Weber, B. W. (2018). On the fintech revolution: Interpreting the forces of innovation, disruption, and transformation in financial services. Journal of Management Information Systems, 35(1), 220-265.

5)It is not clear why it is important to study the opinion of the educated respondents. This also arises an enquiry regarding the possibility to generalize the results of the study.

6)The construct ‘Personal Investment Approach’ the manuscript used, represents a nice contribution. However, the manuscript needs to report the reliability and validity measures for this construct.

7)The Abbreviation ESMA needs to be clarified. There is an opportunity that some readers might not be familiar with the abbreviation of European Securities and Markets Authority

Reviewer #2: This paper investigates the moderating role of age and the differences across age groups on the technology acceptance model (TAM) in the context of financial robo-advice.

The authors find that age, recoded into 'generational cohorts', moderates some of the relationships of the TAM, and that there are differences across some of the age groups. In general, the paper is well-written. I have a few remarks and concerns regarding the paper.

RESEARCH QUESTIONS AND CONCEPTUALIZATION

1. The motivation and relevance of the research questions (the moderating effect of age) should be strengthened. Currently, the authors mention they "believe" age may be important.

While the (moderating) effects of age have been investigated in several technology-related domains before, why is age important in the specific context of robo-advice, in particular financial robo-advice? Lourenço, Dellaert, and Donkers (2020), for instance, have included age as a control in their recent study of robo-advice acceptance.

2. And then the authors discuss why there could be differences between the two youngest groups but not between those and the oldest groups.

3. The authors should also discuss, even if briefly, the differences between age-, cohort-, and period-effects and the challenges to disentangle the three. In so doing, they should clarify that they recode age into generational cohorts out of convenience, but they do not intend estimate the effects of age vs. cohort (or age vs. period nor cohort vs. period).

4. The hypotheses, which could perhaps be re-located to section 1, should be formulated in reference to the moderating effect of age and the differences across cohorts. The current hypotheses refer to the original TAM model, which have been tested countless times before.

5. In the introduction, the authors could add on p. 3, line 34 the work by Baker and Dellaert (2017).

6. While the authors also emphasize, and analyze in detail, the effect of gender, the introduction and conceptualization are silent about it. As such, mentioning on page 6, lines 98-100, that the purpose of the study is two-fold comes as a surprise, to say the least.

METHODOLOGY AND RESULTS

1. The authors do not explain how they have recoded age into generational cohorts, and what past sources have they followed to do that. That is, what birth years (current age) were considered to classify individuals into the four different cohorts?

2. The authors should also be consistent in their use of generational labels. For instance, sometimes they refer to young individuals as Millennials and sometimes as Generation Z, or Digital Natives and Generation Y. Also, are Millennials Gen. Z or Gen. Y? More importantly, 'Digital Natives' may be misleading, as Millennials are often considered Digital Natives too.

3. The authors should report the fit measures typically used when estimating structural equation models, and show in particular that a model with age (and age cohorts) outperforms in terms of fit a model without age. They could use a likelihood ratio test.

4. Do the results change if instead of a multigroup analysis the authors estimate a model with the entire sample but include the moderating variables (3 dummy variables, one for each of the three cohorts that compare to the reference cohort, Gen. Z)? While comparing such model, the authors should not only incude the moderating variables but also the cohort dummies as main effects.

5. What is the number of respondents who fall into the two oldest cohorts, Gen. X and Baby Boomers? In Table 1, we can only see that, jointly, there are 70 respondents in the 'other' age group.

6. What is the response rate, i.e. the ratio of respondents to the number of individuals who received the link?

CONCLUSIONS

The authors should discuss, even if briefly, the limitations of their study, namely (i) the fact that it is only based on self-reported measures (it is not a behavioral study where respondents get to evaluate an actual financial robo-advice algorithm or interface), and (ii) their sample is an Italian sample of highly educated individuals.

Typographical errors and other writing errors to correct:

p.3, line 27: 'it's' should be 'it is';

p. 4, line 60: there seems to be something missing, perhaps 'also on' right after 'but';

p. 4, line 72: should be 'four' instead of 'three'

p. 5, line 88: should be 'prospective' instead of 'actual';

p. 10, line 155: 'it's' should be 'it is';

p. 13, lines 220-231: use 'to' instead of arrows;

In the Figures, the first variable should be 'PIA' instead of 'EXT';

References used in this report:

Baker, Tom, and Benedict Dellaert. "Regulating robo advice across the financial services industry." Iowa L. Rev. 103 (2017): 713.

Lourenço, C. J. S., Dellaert, B. G. C., & Donkers, B. (2020). Whose algorithm says so: The relationships between type of firm, perceptions of trust and expertise, and the acceptance of financial robo-advice. Journal of Interactive Marketing, 49, 107–124.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

22 Oct 2021

I have attached three reports with detailed answers to the Academic Editor and the two reviewers, respectively. Please, see attached files.

Submitted filename: AnswersReviewer2 30092021.pdf

Click here for additional data file.

15 Dec 2021

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Ali B. Mahmoud, Ph.D.

Academic Editor

PLOS ONE

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #2: (No Response)

Reviewer #3: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #2: Partly

Reviewer #3: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #2: I Don't Know

Reviewer #3: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #2: Yes

Reviewer #3: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #2: No

Reviewer #3: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #2: This paper’s topic (acceptance of financial robo-advice) is a timely one. And though demographic measurements are typically used for profiling they may hold some predictive or even explanatory power of their own, relevant for practitioners and policy makers. In that sense, I understand why one could be interested in the moderating effect of age — and gender.

However, the paper has several limitations. I hope the following list helps the authors strengthen their study and the manuscript.

RESEARCH QUESTIONS AND CONCEPTUALIZATION

1. The motivation, managerial and academic relevance, and potential contribution of the study’s main research question — the moderating role of age and gender on the effect of the TAM antecedents on the intentions to use a financial robo-advisor — are again not sufficiently thought through or discussed.

2. Specifically, why are age and gender important to study, managerially and academically, in the context of financial robo-advice adoption, and given what we already know about technology acceptance in general?

3. In the Introduction, there is only one paragraph where age and gender are mentioned, and it is stated there that “several studies have documented the relevance of age and gender”, although the authors do state that in a few studies the results are “conflicting”. Is this list exhaustive? What is the direction of the effects of age and gender if one would enlarge the literature review on those variables? Searching for “technology acceptance by older adults” or “by older people” and related terms shows a long list of studies and review papers.

4. More importantly, how exactly would we expect age and gender to have a moderating effect in the context at hand?

5. One could very well imagine that age makes individuals outdated in terms of technology awareness and have a hard time using a new technology, among other effects, as could certainly be predicted from the TAM model and has been shown before. This prediction, or a similar one, though not new, is never discussed.

6. But one could further imagine that there is theory and evidence to support a moderating effect of age in the context of financial robo-advice, possibly different from the one regarding technology acceptance in general. For instance, if older individuals become not only more interested in financial products due to being close to retirement but also maybe more comfortable in admitting to an algorithm rather than a human advisor their limited financial literacy.

7. For gender, one could very well imagine that the higher risk aversion that has been documented among females in many studies could be at play here, and help generating hypotheses — or leaving the matter for empirical scrutiny in case there would be conflicting arguments when it comes to financial robo-advice in particular (maybe the higher risk aversion among females only shows in interaction with education, which would mean that the effect would not show in a highly educated sample).

8. But no such arguments, or similar ones, are discussed in the paper.

9. In the “Research methods and hypotheses” section (by the way, it would be better to avoid discussing methodologies and general (SEM) modelling details in a conceptualization section unless the contribution is a methodological one), there is again no discussion about the two moderating variables that make part of the authors’ research questions.

10. While the authors discuss details of the TAM model and the estimation of a structural equation model or SEM — and in so doing seem to mix a conceptual model with an econometric one — in the aforementioned section the hypotheses are all stated in terms of the TAM model in general, not in reference to age nor gender and their possibly moderating effects.

11. The authors still suggest that age effects are equivalent to cohort effects. However, typical “age effects” have to do with biological aging, while “cohort effects” stem from common, often cultural, influences that impact in the same way different individuals born around the same time (these influences are further different from “period effects” such as financial crises or a war). The literature on age-period-cohort analysis in sociology, economics, or epidemiology is quite mature and clear about those different effects.

METHODOLOGY AND RESULTS

1. The potentially misleading suggestion referred to above, is clear from the methodology employed in the paper, whereby the authors simply discretize the continuous variable age into 3 categories that are afterwards called “generational cohorts.”

2. The discretization exercise may itself be unfit for a cohort analysis, as the width of the authors’ cohorts is not the same across the three age categories used: 18 years for gen Z, 13 years for gen Y, and, most likely, several decades for gen X+.

3. The authors’ age analysis is in the end an attempt to uncover discontinuities or non-linearities when it comes to the moderating effect of age — again, not only is the moderating effect of age not sufficiently motivated, any specific discontinuities or non-linearities are also not discussed or hypothesized.

4. The composite reliability or Cronbach Alpha analysis does not seem to be satisfactory for the construct PIA - Personal Investment Approach, which appears to be a scale constructed for this study. Would removing any item improve the Cronbach Alpha? The average variance extracted for PEU also appears problematic.

5. As for the variance inflation factors (VIFs) used to assess multicollinearity (Table 4), it should be obvious that they are equal to one when there is only one explanatory variable.

6. Tables 2, 4, and 5 (Table 3 appears to be missing), are redundant and unnecessarily long and could be summarized in one or two sentences.

7. The AIC diagnostics in Table 7 are not so clear cut, as the simpler model with continuous age (and thus the one that uses all the age information without discretizing it) appears to be better for some of the intermediate equations.

8. How do the models compare if other information criteria are used, such as the BIC, etc.?

9. The adjusted R-square Table 10 offers little fit information as it refers to one model only. That is, it does not compare the competing model specifications.

10. The results in Tables 8 and 9 (and in Table 6) are of no use to test whether the authors’ suggestion that there might be moderating effects, is plausible. Those tests are in Table 11. Thus, the length of the paper — and its readability — can be greatly improved.

11. Why are the estimates of PEU —> PU and PIA —> PEU smaller among older survey participants but the estimates of PIA —> PU are larger? Is it because older participants are more interested in investment as retirement becomes closer? (see also the conceptualization point 6 above)

12. The previous point is crucial to position the paper and help the authors re-conceptualize their approach and formulate hypotheses. Whether these results are enough to sustain an entire study / paper is another question.

13. As in any survey not designed to account for endogeneity, all the antecedent variables of BIU may be endogenous. If not the least because BIU might be what causes — within the survey — respondents to answer the way they do to the questions measuring the antecedents (PEU, PU, etc.). Again, this is difficult to account for, but should be discussed as a possible limitation of the study.

14. The use of a highly educated sample, as the authors call it, also limits the generalizability of the results.

Reviewer #3: Thank you for submitting the revised manuscript and addressing previous review comments comprehensively. I have a few more suggestions comments for improving the discussions in the paper.

1. While the gender comparison did not produce notable differences, the paper does describe some minor gender gaps that may still require attention. I suggest that the authors note finding minor gender differences and stating their nature in the abstract, as just saying "we do not find significant differences between male and female" may lead readers to think that no differences were found.

2. Overall, it is interesting - refreshing yet surprising - that no significant gender gaps were found, as much of research on new technology adoption in the recent years show that men are more accepting of new technologies than women in various domains. Is there something that's different about the nature of robo-advisors that may contribute to the lack of differences? Could it be a result of the sample composition, e.g., did males and females in the sample share comparable characteristics, or were there characteristics of the sample that may have contributed to the muted gender effects?

3. The very last paragraph stating limitations of the study needs to be elaborated much further. The sample composition pose additional challenges and limitations than what the authors have mentioned, as there are additional characteristics that make the sample far from representative. A deeper discussion of the sample limitations, and how these may have contributed to the findings need to be discussed. Additionally, it would be helpful to see examples and suggestions around the "behavioral studies" proposed to improve the objectivity and validity of the research.

4. While the paper is written in standard language and does not have major grammatical errors, use of non-academic language and subjectively toned sentences are noticeable throughout the discussions which seem unfit for a research manuscript.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #2: No

Reviewer #3: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

28 Feb 2022

See attached files.

Submitted filename: AnswersReviewer3.pdf

Click here for additional data file.

18 Apr 2022

19 May 2022

See attached file

Submitted filename: AnswerReviewerFinal.pdf

Click here for additional data file.

23 May 2022

Robo-advisor acceptance: do gender and generation matter?

PONE-D-21-14526R3

Dear Dr. FIGA'-TALAMANCA,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Ali B. Mahmoud, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

3 Jun 2022

PONE-D-21-14526R3

Robo-advisor acceptance: do gender and generation matter?

Dear Dr. Figà-Talamanca:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Ali B. Mahmoud

Academic Editor

PLOS ONE