The value of question‐first citizen science in urban ecology and conservation

Citizen‐science projects focused on ecology and conservation have been growing in popularity in recent years, offering many opportunities for researchers and volunteers alike. Two principal approaches to citizen‐science projects in ecology can be characterized as the data‐first approach and the question‐first approach. Here, we highlight the value of question‐first citizen‐science projects for providing insights into the ecology and management of urban wildlife, using case studies on (1) beneficial insects (pollinators, predators and parasitoids) and (2) possums and gliders in Australian cities and towns. The question‐first approach has many benefits, offering a platform to engage volunteers with the scientific process and the broader context of an ecological or conservation problem, while also connecting them with their local environment. Identifying the questions to be addressed in a citizen‐science project ahead of data collection allows for co‐design and stronger collaboration with volunteers, community groups, local experts, and landscape managers. Question‐first citizen science can also provide valuable ecological data that extend substantially beyond presence‐only records, including presence‐absence data collected via timed surveys and information on animal behavior and interspecific interactions. However, establishing and maintaining question‐first citizen‐science projects can be challenging, requiring the building and maintenance of many relationships and a multidisciplinary approach that goes well beyond the usual activities of an academic researcher. Well‐designed, question‐first citizen science has the capacity to achieve both scientific rigor and meaningful engagement with volunteer participants.


| INTRODUCTION
Citizen science (also known as public participation in scientific research, or PPSR) is a scientific practice that involves volunteers in the collection and/or analysis of scientific data (Meredith, 1996;Shirk et al., 2012). Enthusiasm for citizen-science projects that are focused on biodiversity, ecology, and conservation has been growing in recent years (Follett & Strezov, 2015;Pocock et al., 2017;Steven et al., 2019). For example, the Australian Citizen Science Project Finder (via ala. org.au) lists 262 current biodiversity projects. The scale of these projects varies from local to national, and their scope ranges from a single species or habitat type to entire ecological communities. Some projects run for years while others are structured around a single short-term event such as a community "BioBlitz", where the objective is to record as many species as possible in a defined place and time frame (Robinson et al., 2013).
Citizen science offers many benefits for researchers and volunteers alike, especially for large-scale and/or longitudinal field studies (e.g., see Dickinson et al., 2012;Griffiths et al., 2019;Loss et al., 2015). For researchers, these benefits include access to a large pool of enthusiastic volunteers who can collect data at multiple points in time across a wide geographic area (Kosmala et al., 2016). In an environment of constrained research funding, citizen science can facilitate projects that would otherwise be too ambitious and too expensive to complete (Dickinson et al., 2012;Flesch & Belt, 2017;Galv an et al., 2021). The scale and pace of data collection via popular platforms such as eBird far exceed what would be possible with a conventional field-based study (Callaghan, Poore, et al., 2021).
Researchers may also see citizen science as a key component of science communication: an effective way to engage the broader community with both the scientific endeavor and pressing environmental problems such as the biodiversity crisis (Peter et al., 2021;Vitone et al., 2016). Further, involvement in citizen-science programs can lead to sustained behavior change among participants. For example, participants in a program to survey backyard butterflies in France planted more nectar-bearing plants for butterflies and reduced the use of pesticides in their own gardens (Deguines et al., 2020). An online survey of 1160 participants of biodiversityfocused citizen-science programs across 12 countries in North America, Europe, Africa and Australia found similar behavioral changes, plus an increased likelihood that participants would discuss biodiversity and environmental problems with others in their sphere of influence (Peter et al., 2021).
For volunteers, the benefits of citizen science include the opportunity to contribute to research while gaining new knowledge, skills and an increased connection with the biological or physical world around them (Peter et al., 2019(Peter et al., , 2021Roger & Klistorner, 2016;Schuttler et al., 2018). A range of studies has also found that involvement in citizen science increases social capital through new social connections with other volunteers, the opportunity to share and learn with like-minded people, and increased trust and cooperation in the communities where programs operate (Conrad & Hilchey, 2011;Peter et al., 2021;Schläppy et al., 2017). Some participants also report benefits for their mental and physical well-being, such as feeling calmer, or a sense of joy and satisfaction gained through being part of something that is worthwhile and via an increased connection to nature (Peter et al., 2021).
Two principal approaches to citizen-science projects in ecology can be characterized as the data-first approach and the question-first approach. As the name suggests, data-first projects are established to engage citizen scientists in the collection of ecological data (such as incidental observations of species' presence) without having a clear research question in mind. However, these data are then sometimes used in a post-hoc way to address questions such as those relating to the distribution of species or estimates of their abundance (e.g., Clemens et al. 2016;. Data-first projects may be underpinned by a strong emphasis on the participation, training and engagement of volunteers as valuable goals in themselves. For example, The Australian Barcode for Life project aims to train citizen scientists in the methods of DNA barcoding to provide participants with valuable skills and increase their local environmental awareness and stewardship (details via ala.org.au). In contrast, question-first citizen-science projects are designed to address one or more specific ecological questions or hypotheses (see Smith et al., 2020). Examples of question-first projects include Project Martin Watch in North America, devised to address a range of questions related to the breeding behavior and reproductive success of purple martins in provisioned housing (Raleigh et al. 2019); and Echidna CSI in Australia, designed to address questions relating to the diet, gut health, stress levels and reproductive status of shortbeaked echidnas (Perry et al. 2022).
While both approaches to citizen science are valuable and have made significant contributions to biodiversity knowledge, there remains some debate about the relative merits of each (e.g., Dickinson et al., 2010;Elliot & Rosenberg 2019). Concerns about the data-first approach align with some expressed more generally about "big data" projects in ecology, which emphasize data collection as a valuable activity in its own right rather than as something designed to address an articulated question or hypothesis (Bayraktarov et al., 2019). Likens (2013, 2018) have even suggested that putting data collection first is effectively doing science backwards, potentially leading to wasted effort and a poor match between the information collected and important ecological questions requiring answers. These mismatches may be related to the scale of data collection, the methods used and/or inherent biases in citizenscience datasets that tend to be skewed towards temperate regions of the world and accessible locations such as those along roads or near human-population centers (Boakes et al., 2010;Hughes et al., 2021;Robinson et al., 2022).
In this perspective piece, we highlight the use of question-first citizen science to address a series of questions relating to the ecology and management of urban wildlife in Australia, via a mobile (cellular) telephone application. These questions were developed in collaboration with urban-wildlife managers and members of the local community, and in the two case studies we present, successfully engaged city-dwelling citizens with nature in their local environment. We discuss the benefits of this approach but also its challenges, such as the need for a dedicated champion to recruit and train volunteers and maintain their connection with each other and the project over time. We contend that question-first citizen science can (1) offer an important model for collaborative ecological and conservation projects; (2) provide valuable ecological data that extend substantially beyond presence-only records; and (3) facilitate a deep connection between volunteers and the natural world around them.

| The app
The CAUL Urban Wildlife App (the App) is a free application for mobile (cellular) phones, available for iOS and Android, designed to harness the power of citizen scientists across Australian cities and towns. Approximately 90% of the Australian population lives in urban areas (ABS, 2019), representing a large pool of potential participants. The App contains four modules; we developed each of these to address one or more ecological questions and to contribute to the research program of the Clean Air and Urban Landscapes Hub (CAUL) and the Threatened Species Recovery Hub (TSR) of the National Environmental Science Program (NESP), Phase 1. This program was funded by the Australian Government through the Department of Agriculture, Water and the Environment from 2015 to 2021 and comprised six research hubs. As their names suggest, the CAUL Hub conducted research on urban environments from air to landscapes, while the TSR Hub conducted research to help protect and recover threatened species and ecological communities (see https://www.dcceew.gov.au/ science-research/nesp/phase-1 for more information).
We developed the App in consultation with wildlife experts, urban ecologists, land managers and community groups to address key information gaps in urban-wildlife management around Australia. Its four modules focus on beneficial insects (pollinators, predators and parasitoids), bell frogs, flying-foxes (also known as fruit bats), and possums and gliders ( Figure 1; Table 1). We selected these groups because they are readily observable in urban landscapes and are often the target of management actions due to their public appeal (e.g., butterflies, native bees, bell frogs), the ecosystem services they provide (beneficial insects, flying-foxes), their status as threatened species (bell frogs, flying-foxes, and some possums and gliders including the western ringtail possum, greater glider and yellow-bellied glider; Table 1), and/or their potential to come into conflict with human residents (flying-foxes, possums). However, information on their fine-scale distribution, behavior and interspecific interactions in urban environments is currently lacking, especially from private land such as residential gardens (Steven et al., 2021;Van Helden et al., 2020). The App's structure allows for further modules to be added as new questions emerge regarding additional taxonomic groups.
The CAUL Urban Wildlife App provides a platform for data to be collected using the same protocols that researchers employ, a key feature that sets it apart from many other mobile apps. For example, after completing a short training and certification session, citizen scientists can conduct timed surveys for beneficial insects or frogs using standardized ecological-research protocols. Incidental observations can also be submitted for all four modules, and citizen scientists have the opportunity to upload photos, videos and audio recordings to assist with species identification. Participants may choose to engage with all four modules or focus on their preferred taxonomic group. To date, the App has engaged more than 350 citizen scientists, with over 4000 observations submitted between June 2017 and January 2023. These have helped identify important ecological interactions, provided information on how our target species use and move through cities and towns, and revealed sources of human-wildlife conflict and mortality (Soanes, Lentini, et al., 2020).
The number of insect observations made through the CAUL Urban Wildlife App is similar to that recorded through the 2021 Wild Pollinator Count (another nationwide citizen-science survey; https://wildpollinatorcount. com). However, we have so far received relatively few records of frogs and flying-foxes. Citizen-science activity on these projects has been constrained first by the Black Summer bushfires of 2019-2020, which burnt a large proportion of the flying-foxes' range in eastern Australia, and by lockdowns and travel restrictions associated with the COVID-19 pandemic. This is the first published paper to describe the App and report on its findings. We have How are possums and gliders using public and private land in cities? Where are these animals coming into conflict with people?

F I G U R E 2
The network of plant-insect interactions between the 12 pollinator observatories at Westgate Park (green nodes) and native and introduced insect taxa (blue and purple nodes, respectively) observed during direct-observation surveys by researchers and citizen scientists. The width of each ribbon indicates the strength of the interaction. Chord diagram created in the online implementation of Circos (Krzywinski et al., 2009); reproduced from Mata, Vogel, and Bolitho (2020). chosen to highlight two of the four modules in the App as case studies, focusing on those with the most records submitted to date (beneficial insects, and possums and gliders).

| CASE STUDY: BENEFICIAL INSECTS
The beneficial-insect module of the CAUL Urban Wildlife App has been used in a wide range of settings, including for research, in practical classes for tertiary students (Mata, Vogel, & Bolitho, 2020), and at community events (Beer et al., 2019;Mata, Vogel, & Bolitho, 2020). However, perhaps the most successful application has been for the monitoring of "pollinator observatories." A pollinator observatory is a single flowering plant species in a plot of at least 5 m 2 (Mata, Vogel, & Bolitho, 2020  repeatedly to record interactions between the plants and their beneficial-insect partners over time and across seasons. In 2017, a network of 12 pollinator observatories was established at Westgate Park, a 41-ha public greenspace located in the City of Melbourne, in collaboration with the local community group, Westgate Biodiversity: Bili Nursery & Landcare (WBBNL). Researchers surveyed the pollinator observatories for plant-insect interactions each month over a two-year period using timed observational surveys and sweep netting, and facilitated a citizenscience survey each quarter. These comprised a half-day workshop that trained and certified citizen scientists in insect identification, followed by a timed observational survey at each pollinator observatory in flower, using the App or a paper-based questionnaire (Mata, Vogel, & Bolitho, 2020). A total of 189 citizen scientists participated in one or more workshops across the two-year study period, contributing 20% of the total plant-insect observations recorded during the project (Mata, Vogel, & Bolitho, 2020).
Research questions addressed with this module included: Which beneficial insects interact with the pollinator observatories? What proportion of insect pollinators observed during direct-observation surveys were native or introduced? (Figure 2; reproduced from Mata, Vogel, & Bolitho, 2020); Which plant species support the most and fewest taxa of pollinators, predators and parasitoids? Which beneficial insects are specialists or generalists with respect to the plants they visit? (Figure 3); How can we predict when we are most likely to observe particular insect species interacting with pollinator observatories? What management actions can we take to support taxonomically and functionally diverse insect communities in urban parks? (Mata, Vogel, & Bolitho, 2020). This module of the App was championed by two researchers and a member of the WBBNL Committee of Management.
The question-first approach has been key to engaging citizen scientists with this project. For example, members of the WBBNL worked with researchers to define the research questions the pollinator-observatory project would address, which in turn informed the project's design. Further, members of the community group are now implementing management recommendations based on the findings from the project. These include actions to increase the cover and spatial distribution of plant species that provide resources for a taxonomically and functionally diverse range of common and rare insect species, and to engage park visitors with the pollinator observatories through signage, posters and a long-term "Insect pollinators of Westgate Park" citizen-science monitoring program (Mata, Vogel, & Bolitho, 2020). For us as researchers, the engagement and enthusiasm of citizen scientists has greatly advanced our understanding of management actions that will better support insect biodiversity in urban landscapes (Mata, Ramalho, et al., 2020).

| CASE STUDY: POSSUMS AND GLIDERS
The possum and glider module of the CAUL Urban Wildlife App has been used by urban and peri-urban residents to record and share sightings of possums and gliders across Australia. Citizen scientists have contributed records from both private gardens and public parks and streets, including identifying hotspots of urban mortality. While the module includes all species of Australian possums and gliders, its use has been championed predominantly in the southwest of the continent with a focus on the critically endangered western ringtail possum (Pseudocheirus occidentalis; Figure 4; Steven et al., 2021).
The western ringtail possum was identified as a priority threatened species in 2016 by the Australian Government (DoE, 2016), catalyzing a broad-scale engagement program to raise awareness of the species' plight. The species appears capable of persisting at relatively high densities in urban environments (Van Helden et al., 2020); however, its abundance also causes conflict with some humans who resent possums damaging their garden plants and/or nesting in their roof spaces (Steven, 2020). The possums are also at risk from predation by domestic cats and dogs (Van Helden et al., 2020). This makes the species a useful target for a citizenscience approachengaging urban residents to learn more about the possum's distribution and habitat use in urban areas, particularly on private properties, and learning more about human responses to the species' presence.
We designed the possum and glider module in consultation with a suite of experts, conservation managers and community advocates. The questions addressed related to behavior and habitat use (e.g., Is the possum or glider in a tree, on a fence line or a power line? Is it in a public park or on private property?), browsing (What species of food plants are the possums and gliders eating?), and the potential for human-wildlife conflict (Is the observer happy, unhappy or neutral about seeing the animal?). This module was championed by one of the researchers in our group, who organized training workshops for volunteers in multiple states of Australia.
We have received 530 records of possums and gliders since the module was launched in 2019, with citizen scientists shedding light on how western ringtail possums use their gardens, fence lines and roofs. Much of the conflict between humans and possums in urban areas stems from the premise that they commonly occupy roofs; however, our sample found this to be true for only 5.7% of observations to date. Sadly, 14% of sightings were of animals struck and killed by vehicles on urban streets. Of 258 answers to the question "If you are observing the species on your property, how do you feel about this?", 248 (96%) indicated they were happy to see the possum or glider. The remainder were either indifferent or unhappy because the animal observed was dead. It is worth noting that this is not an unbiased sample of human attitudes to possums and gliders in the urban environment, as the participating citizen scientists were likely to have a stronger than average affiliation for the environment and for wildlife in particular.
The high engagement with this module was facilitated by community-outreach seminars and workshops in cities across the country and strong partnership with local organizations such as the South West Catchments Council. More than 300 volunteers attended training workshops in south-west Western Australia, also taking home habitat plants for the western ringtail possum to plant on their own properties. Across these workshops and additional engagement activities with some local wineries, we facilitated the planting of approximately 2000 habitat plants for this critically endangered species. The question-first approach to citizen science allowed volunteers to see how their observations could inform practical management of possums and gliders in urban environments, which may have encouraged stronger engagement. The participation of citizen scientists also provided researchers with insights into the role of private land in urban wildlife conservation, a novel habitat for many species that is otherwise very difficult for professional ecologists to survey (Steven et al., 2021). In a feedback questionnaire completed at the end of each training workshop, volunteers indicated that they would apply the things they learnt by planting new habitat plants for the western ringtail possum, keeping their pets inside at night, educating others and using the App to undertake further surveys. An example response: "I will ensure my garden is possum friendly and I will spread the word to others. I will be more careful when driving through known possum habitats at times when they are active." 5 | DISCUSSION

| Benefits of the question-first approach
As demonstrated in the case studies above, question-first citizen science offers a useful model for ecological and conservation projects. Identifying the questions to be addressed in a project ahead of data collection allows for a co-designed approach and stronger collaboration with volunteers, community groups, local experts and landscape managers. Volunteers also gain significant benefits, including the opportunity to engage with the scientific process and the broader context of an ecological or conservation problem.
The question-first approach to citizen science exemplified in the CAUL Urban Wildlife App facilitates the collection of data on interspecific interactions (through the modules on beneficial insects, flying-foxes, and possums and gliders; Figures 1-3) and behavioral data (e.g., frog breeding behavior and the presence of different life stages from eggs to adults, and the feeding behavior of flying-foxes, possums and gliders). Collection of data that go beyond presence records (also known as secondary data) has recently been highlighted as an important frontier for biodiversity research using citizen science (Callaghan, Poore, et al., 2021). The richness of information gathered through this approach can improve understanding of processes affecting the distribution and fitness of wildlife, as well as the threats they face in urban environments and beyond, with practical benefits for conservation management at both a local and a national scale.
Further, the question-first approach to citizen science used in the CAUL Urban Wildlife App enables the collection of presence-absence data via standardized, timed surveys in the beneficial-insect and frog modules. Presence-absence data allow for more sophisticated analyses and stronger statistical inference than can be gained from presence-only records (Elith et al., 2020;Hastie & Fithian, 2013). For example, presence-only data cannot be used to estimate the prevalence of a species, only the relative likelihood of its occurrence (Hastie & Fithian, 2013). Bias in the selection of data-collection points is common in data-first citizen-science projects. Combined with uncertainty arising from imperfect detection, this can sometimes lead to inaccurate population estimates, as was seen in a study by  that estimated global bird population sizes from eBird data (critiqued by Robinson et al., 2022).

| Challenges and constraints of the question-first approach
Establishing and maintaining a question-first citizenscience project comes with several challenges, two of which we highlight here. First, question-first projects are most successful when supported by one or more local champions who can drive the recruitment and training of volunteers, organize specific events where data will be collected, and facilitate ongoing engagement to keep volunteers connected with the project and each other over time (Capdevila et al., 2020;Phillips et al., 2019). This work involves building and maintaining many relationships, takes substantial time and energy, and goes well beyond the usual activities of an academic researcher. Scientists interested in establishing a question-first citizen-science project should be prepared to broaden their multidisciplinary skills in order to develop and lead the engagement activities required to generate buy-in to the project (Frigerio et al., 2018;Hecker et al., 2018;Steven et al., 2019). Collaboration with community groups and other local organizations that share a common goal or stand to benefit from the research can provide valuable assistance in this regard, as seen in the case studies above. Projects also need to consider the most appropriate systems for "reporting back" to volunteering citizen scientists on the results achieved (Steven et al., 2019), allowing them to see the importance of their involvement in the broader project and helping to ensure their satisfaction of achievement in contribution (Luna et al., 2018;Soanes, Cranney, et al., 2020).
Second, the quantity of data collected by a questionfirst project, especially one requiring the formal training of volunteer participants, is often smaller than that obtained through data-first projects. Related to this, data collection can also be vulnerable to unforeseen events. For example, the popular data-first project eBird has collected more than one billion records of birds worldwide (eBird, 2021), while the FrogID project established by the Australian Museum has seen more than 6500 app users submit recordings of frog calls for expert validation (Rowley et al., 2019). In contrast, two of the four modules in our app have so far collected only a modest amount of data, partly due to unforeseen events including constraints on travel and outdoor activities associated with large-scale bushfires followed by the COVID-19 pandemic. We had planned a media blitz and in-person workshops to promote the flyingfox module throughout eastern New South Wales in the summer of 2019-2020, all of which had to be canceled because of bushfires. And Western Australia, a key location for our project on bell frogs, was closed to people from other Australian states for more than two years as part of local COVID-management policies.

| CONCLUSION
Despite these constraints, we have found that questionfirst citizen science can engage volunteers who are motivated by the opportunity to contribute to current, well-defined research projects in their local neighborhood (Loss et al., 2015;Steven et al., 2019). In the pollinator-observatory project described above, almost 200 volunteers attended training workshops and conducted surveys for insect pollinators at Westgate Park (Mata, Vogel, & Bolitho, 2020), while more than 300 volunteers attended training workshops for the possum and glider module. There are great opportunities to involve this cohort of prospective citizen scientists in projects to help support urban biodiversity in cities and towns worldwide. We contend that well-designed, question-first citizen science has the capacity to achieve both scientific rigor and meaningful engagement with volunteer participants.

DATA AVAILABILITY STATEMENT
Data are available via The University of Melbourne's data repository: https://doi.org/10.26188/22216729.

SUPPORTING INFORMATION
Additional supporting information can be found online in the Supporting Information section at the end of this article.