A STUDY OF USER RESPONSIVENESS IN 
PARTICATORY SOCIAL APPLICATIONS

Caren Crowley, Rafael Bachiller, Wilfried Daniels, Wouter Joosen and Danny Hughes
iMinds-DistriNet, KU Leuven, Leuven, Belgium
caren.crowley@cs.kuleuven.be

ABSTRACT
Participatory applications rely upon crowd sourcing, whereby a community of users contribute resources such as: text reports, sensor data and multimedia files. The participatory application model is scalable, low cost and enables the development of a new class of user-centric applications. Online social networks are a natural fit for recruiting and communicating with user communities. However, the success or failure of participatory applications is dependent on developing and maintaining a community of users that is willing to supply data as required by the application. In this paper we explore the factors that influence user responsiveness to queries through an exploratory 30-day study in which 3,055 messages were sent to 70 participants.

INTRODUCTION
An increasing number of participatory applications are emerging which rely upon a user community to contribute content such as: text, audio, images and video (Christin et al., 2011; Hoseini-Tabatabaei et al., 2013). Participatory applications are scalable, low cost and enable the development of a new class of user-centric applications. These applications can be divided into two classes; automatic and manual participatory applications. In the case of automatic participatory applications, the level of human involvement required is relatively low. Data is automatically sensed by software running on the userÕs device and reported to the application servers. Communication is therefore primarily between the application and the users device. Examples of automatic participatory applications include environmental sensing of real-time traffic conditions (Mohan et al.  2008), or human-centric sensing, involving the recording of data to infer the activities and location of participants (Eisenmann et al., 2007). In the case of manual participatory applications, higher levels of human involvement are both enabled and required. This allows for the creation of content and capture of phenomena that would not be possible in an automatic fashion. Examples of manual participatory applications include weather reports (Demirbas et al., 2010) and micro-blogs involving audio and visual content (Gaonkar et al., 2008).  However, the benefits of this application model must to be balanced against the inherent risks of expanded human involvement. Humans are inherently unreliable, the risks of increasing user responsiveness are twofold: Firstly, there is a risk that query messages will be lost if the user does not notice them or chooses to ignore them. Secondly, users may take an arbitrarily long time to respond to a message, which is a significant problem for time-sensitive applications. This leads us to pose the question: what can be done to increase user responsiveness to application queries? 

PARTICIPATORY APPLICATIONS
Manual participatory applications are an emerging area and much of the prior work has focused on automatic participatory sensing architectures. As such research has focused primarily on building appropriate software architectures  (Jia et al., 2007; Rachurri et al., 2011) and the deployment of participatory applications in real world settings (Milluzzo et al., 2008; Eisenmann et al., 2007). To highlight the importance of responsiveness to query messages and understand why it has received limited attention in the literature, we examine the difference between querying (or tasking) in automatic and manual participatory application models respectively. Christin et al. (2011) provide a detailed overview of automatic participatory sensing applications. In such a model, Querying or Tasking by the application is simplified as communication occurs between the application and software installed on the participants devices. The Tasking component thus automatically distributes sensing tasks to the devices. The task specifies the data to be collected (e.g. traffic or weather conditions), the modalities used to collect it (e.g. text report or microphone sensor), the location and the timeframe of interest. Data is collected automatically by an application running on the participantsÕ devices whenever the participant happens to be in the desired location during the desired time frame and transmitted to the servers of the participatory application.

In contrast, communication in manual participatory applications is between the application and the user rather than the device. In order for the user to respond to the message within the required time frame the query needs to be noticed and user must alter their behaviour to undertake tasks on demand and then report the required data to the application. Lane et al., (2008:12) argue that in the case of applications where users must consciously choose to answer queries, socio-technical techniques must be developed to encourage user involvement. Such techniques are necessary in order to ensure the participation of a large community of users and to help mitigate the negative impact that arises when a participatory application interrupts normal user behaviour. 

Prior research has confirmed that users are willing to capture data and create content to be used by participatory applications (Christin et al., 2011, Hoseini-Tabatabaei et al., 2013). However, high rates of message loss and slow response times remain a significant impediment to the development and widespread adoption of the participatory application model (Demirbas et al., 2010). Lane (2008, p.12) argue that in the case of applications where users must actively answer query messages, socio-technical approaches must be developed to encourage user involvement. Such techniques, Lane et al. (2010) argue, are required to minimise the negative effects of unreliable human behaviour. The responsiveness of users is likely to vary due to differences in how they interact with the application and their current context and relationships with other users. By better understanding general usage patterns and the interactions between user attributes and contextual conditions, we can reduce the burden of participation on users while increasing user responsiveness. 

A NETWORK PERSPECTIVE ON USER QUERYING
We analyse participatory applications from a network perspective (Tsai 2001, Kenis and Oerlemans, 2008), in order to understand how attributes of nodes in the network affect the flow of information. A network in this sense is simply a set of actors linked by a set of ties of a particular type (Borgatti and Halgin, 2011). 

Specifically, participatory applications form a 2-mode network wherein the actor types are the application and the participants. In this network, ties are directional with information flowing between the application and user and vice-versa. The ties that link these actors are the queries sent from the application to the participant (i.e. outgoing ties) and the responses sent from the user to the application (i.e. incoming ties). The application and its users are therefore embedded in a network that is coordinated through the messages sent by the application (Tsai 2001). From a network perspective, user responsiveness is a response attribute that may be understood as the outcome of the combined influence of user attributes and query attributes (i.e. actor and tie attributes respectively). 

Query Attributes
Querying is a significant component of existing manual participatory applications. Gaonkar et al., (2008) include querying as a necessary feature of their Micro-Blog application. The application allows users to create multimedia blogs involving a mix of audio, images and text content. While users were free to upload content according their personal preferences, the trial found that users were much more likely to upload content in response to query requests. In the case of Micro-Blog, queries were location specific and the user could send queries, which were then sorted by the application and sent to other users based on current geographic location. While Gaonakar et al. (2008) do not address response times directly, they do give users an option to set queries as active for specified time-limit only. In addition, Gaonkar report that most Micro-Blogs are uploaded after 3pm with the highest density of blog uploaded between 5pm and 9pm. Feedback from trial users of Micro-Blog application confirmed the importance of timing with users expressing a desire to buffer messages to be responded to later at their convenience. 

The timing of query message tranmission is clearly an important attribute likely to influence responsiveness. For instance, research on survey response has found that surveys are more likely to be completed if they are received on a Tuesday or Wednesday afternoon relative to other days of the week (Dillman, Singh & Christian 2000).  The importance of networking has also been recognised by social applications such as GroupOn which send promotional emails to their members at 10AM, as this time was found to be optimal for their target audience of stay at home parents (Park & Chung, 2012). The rate of querying is also important (i.e. the number of messages that are sent to a user within a given time-period). While more frequent querying may improve the freshness or spatial resolution of application data, it is intuitive that users can only handle a limited number of query messages. Lane et al. (2008) argue that the probability of human cooperation is likely to fall as the daily barrage of queries causes fatigue and eventually annoyance.

In terms of the communication channel through which users are contacted, Online Social Networks (OSNs) are growing in popularity as a medium to support participatory applications. This popularity is driven by two factors. Firstly, OSNs have a massive existing user population. Facebook has over 1.2 billion monthly-active users [Facebook 2012], while Twitter has over 500 million (Twitter, 2012). Secondly, these platforms provide a natural mechanism through which applications can interact with users across many different devices (from desktops to mobile phones).  Demirbas et al. (2010) for instance, recruits users in this manner. However, the feasibility of using OSNs as a generic mechanism to recruit and communicate with participants requires more detailed examination.

User Attributes
In terms of user attributes, there is an increasing emphasis in the research literature on the social and emotional issues that influence participation in applications (Chen & Sakomoto, 2013). In this paper, we examine the influence of remotely observable social factors on participation. Specifically, we focus on tie strength between participant and the individual who recruited them to join the application and the social distance between the application user and the individuals managing the application. This is a complement to prior studies of online social network structure (Mislove et al. 2007; Backstran 2011). Granovetter (1978) developed the concept of tie strength, distinguishing between strong and weak ties. Gilbert et al. (2009) apply this concept of tie strength to online networks, demonstrating through a lab-based experiment that it is possible to use OSN profile data remotely measure tie-strength between OSN users with an accuracy of 85%. Based upon these prior results, we anticipate that stronger ties between participants and their recruiters will lead to greater levels of participation.

In addition to tie strength, the social distance between a participant and those managing the application is likely to influence participation. The more removed a participant is from those using their data the more likely they are to have trust and privacy concerns. For example, Efstratiou et al., (2012) find that users had significant concerns over how data was released to other users of the application. However, such concerns could be overcome when users were able to control the type of data released. 

The devices through which users respond to queries are also likely be important. Advances in smart phone technology mean that non-professionals are likely to possess all the necessary to quickly and conveniently create, capture and report data and other content to the application (Hoseini-Tabatabaei et al., 2013). However, while smart-phones are becoming increasingly ubiquitous, the limited user interface of these devices and prohibitions against their use in the workplace may have a negative effect on the responsiveness of their users.

Response Attributes
Two attributes of response messages are related to user responsiveness: whether the query is lost (i.e. whether the participant fails to respond) and the latency of the response (i.e. how long the participant took to respond). User responsiveness is especially critical for time sensitive applications. This can be clearly seen in the case of the Twitter Weather Radar application implements a participatory weather monitoring system (Demirbas et al. 2010). The application used a dedicated Twitter account to send query ÔtweetsÕ (i.e. Twitter messages) to its ÔfollowersÕ (i.e. subscribers to the account). These queries requested current weather conditions in various locations. Followers of the application were then expected to respond by manually sending a response tweet with the requested weather conditions. The weather monitoring application achieved an average accuracy of 85% for measuring current weather conditions. However, in terms of user participation, Demirbas et al. (2010) reported only a 15% response rate to queries and slow response times, with 50% of responses taking longer than 30 minutes. While the high accuracy of the application demonstrates the potential of participatory sensing applications, slow user response times and high loss rates are a clear pitfall of manual sensing approaches.

Nazir et al., (2008) developed and launched three experimental social gaming applications on Facebook and used the data collected from these applications to examine the response times of messages sent between users. Message requests were sent in an unpredictable fashion and users were often located in different countries, which increased the unpredictability of message transmission times. The authors found that response times for foreign queries were similar to those of local queries. The authors report an average response time of 16.52hrs with the longest response taking as much as 567hrs. If such high response times were inherent in participatory application architectures, this would preclude the development of time-sensitive applications. This clearly motivates further research into user responsiveness.

Conceptual Framework 
A network perspective provides an understanding of message responsiveness as the outcome of the combined influence of actor attributes and tie attributes. In this case, the actor types are the participants and the application (i.e. a 2-mode network). The ties, which link these actors, are directed. Queries are sent from the application to the participants while responses are sent from participants to the application. Figure 1 visualises this theoretical framework and illustrates how this network operates along with the relevant Dependent Variables (DV) and Independent Variables (IV). The precise definition of the variables is provided in the Methodology section.


Figure 1: Theoretical Framework - A Network Perspective on User Responsiveness

While a large body of previous work has investigated how to build participatory sensing architectures (Hoseini-Tabatabaei et al., 2013), the relationship between query attributes, user attributes and responsiveness has yet to be systematically explored. Motivated by this issues we conducted an empirical study that involved the recruitment and automatic querying of experimental participants.

METHODOLOGY
In order to examine how contextual factors affect message loss and user responsiveness, we designed an original experiment that ran for 30 days from December 15th 2012 to January 14th 2013. In total 70 participants were recruited for the study and a total of 3,055 automatic query messages were sent to users during the experiment. 

The methodology that we employed in this study can be broken down into the following three stages: recruitment of participants, empirical study and follow-up survey.

Recruitment of Participants
The purpose of the recruitment phase was to gather sufficient experimental participants to support the empirical study. We selected viral recruitment using Online Social Networks (OSNs) as this approach is increasingly being applied to build and support participatory sensing applications. Furthermore, this kind of viral recruitment is essential for any participatory application that does not have access to a large existing user base. As with the work of Demirbas et al. (2010.) and Milluzo et al. (2008), the success of our experiment shows that OSNs are a feasible mechanism to recruit and maintain contact with application participants.

In order to recruit users, we created a short message asking individuals to participate in our experiment. This message included a HTTP link to a web-based registration system. Using the registration system, users were invited to register their name, social network username(s) and email address. The message used to recruit participants was limited to 140 characters as this is the maximum size of a single Twitter message. The message read as follows: ÒKU-Leuven is conducting an experiment on user performance in participatory applications: please help. More information and sign up at [link]Ó, where [link] is a HTTP link to the sign-up and information webpage.

This recruitment message was initially shared by four of the authors of this paper using Facebook and Twitter. Each user who received this message and registered online was then invited to share a personalized recruitment message with their friends. In this way, the recruitment message spread across the social networks according to how the users chose to redistribute it. User recruitment peaked within the first five days, but continued throughout the 30 days of the experiment.

The graphs shown in Figure 2 are sociograms wherein experimental participants are represented as nodes and social ties are represented as edges. The ties shown in light grey are prior relationships (friendships in Facebook and followers in Twitter). The links shown in dark blue are relationships that were successfully used to recruit an experimental participant. Of these participants, 1 user (1%) used only Twitter, 16 users (23%) used only Facebook and 53 users (76%) used both social networks. The maximum path length that a successful recruitment message travelled was 3 hops.

Figure 2: Recruitment Sociograms for Facebook (left) and Twitter (right)

As can be seen from Figure 3, participants were widely geographically distributed, being located in 11 countries. Participants volunteered from Belgium (57%), the UK (18%), Australia (9%), the US (5%), Spain (3%), Poland (2%), India (2%), Ireland (1%), Sri Lanka (1%), Portugal (1%) and Andorra (1%).

Figure 3: Geographic Location of Experimental Participants

Empirical Study
Recruitment of participants followed a snowball sampling method, with four of the authors of this paper reaching out to contacts using their chosen online social network(s). A brief message was sent to potential participants asking them to register in order to participate in study on participatory applications conducted by the iMinds-Distrinet research group within the Computer Science department based at KU Leuven. Following self-registration, participants began to receive messages from the client application at a random time of day, but at a controlled rate of messaging. The distribution of messages to users via the online social network clients was scheduled using standard Linux CRON jobs. Custom OSN client applications were created for managing the automatic messaging of clients. In the case of Twitter, the client sent a directed tweet  (Twitter message) to participants, while in the case of Facebook; the client sent a private chat message to participants. During the 30-day experiment a total of 3,055 query messages were sent to the 70 participants in the fashion described above.

A simple query message and task was deliberately chosen over a rich participatory application, as the purpose of this exploratory study was to examine the effect of social and contextual factors on user responsiveness, independent of the inherent appeal of a given application. Messages sent to participants contained the following text: ÒOSN connectivity Ð click this link to help by checking-in: [link]Ó. Where [link] is a HTTP link to the online check-in webpage. The check-in webpage presents users with background information on the experiment and on how many Ôcheck-insÕ they have completed. The webpage also allows participants to either drop out of the experiment or share the recruitment message with their OSN contacts.

We explored the effect of rate of querying has on user responsiveness, as follows. Initially all users received 1 message per day. However, on day 15, the group was randomly split (based on odd and even user IDs), half of the users were placed into the Òrate-experimentÓ group and half into the ÒcontrolÓ group. The rate of message was then increased every few days for the Òrate-experimentÓ group such that by the end of the experiment half of the users were receiving 5 messages per day.

As users Ôchecked-inÕ with the application we were able to gather information on attributes of the response message. Each request that is sent to an OSN user embeds a unique web link that associates each response to its corresponding query, OSN and username. When a user follows the check-in link, the web server resolves the IP address of the user to a location, which is used to calculate the local time. The browser agent string, a standard feature of the HTTP protocol (Fielding et al., 1999) is then used to identify the device type through which the user is responding.  It should be noted that our approach to gathering important user attributes gathers no more data than a standard website tracker. Naturally, the gathering of additional meta-data on users would allow for richer models however this comes at the cost of more invasive monitoring. 

Follow-up Survey
Following completion of the survey all enrolled participants were invited to complete a short survey. In addition to basic demographic information regarding participants we collected information allowing us to calculate the tie strength between the participant and the person who recruited them to join the experiment. The survey was completed by 54 (78%) of participants. Responses to the demographic questions indicated that the majority of respondents were male (80%) and aged between 18 -35 years. Clearly our sample is not representative of the general public however the age profile is representative of typical social media users (Duggan and Brenner 2013).

The concept of tie strength was introduced by Granovetter (1973:17) and is defined as a Ôcombination of the amount of time, emotional intensity, the intimacy (mutual confiding), and the reciprocal services which characterise the tieÕ. Gilbert and Karahalios (2009:212) take this definition of tie strength and develop 5 indicator questions to measure tie strength amongst Facebook users. In the follow up survey we adapt three of these questions to develop a measure of tie strength1 between participants and the person who recruited them to join the survey. The tie strength questions used in the survey are listed in Table 1. Questions were answered using a 5-point Likert scale with two extreme values at either end. The responses were then used to create a scale for tie strength ranging from a minimum of 3 to a maximum of 15. 

Table 1: Tie Strength between Users and Recruiter

Measurement of Variables
The dependent variables associated with Query messages were controlled through a custom experimental software system, which also measured the independent variables associated with Response messages. Of the three User variables, path length was calculated when a user signed up to the experiment, device type was recorded whenever a user performed a check-in and tie-strength was estimated based upon responses to a follow-up survey. 

As can be seen from Figure 2, a QUERY has three important attributes, which can be controlled by the APPLICATION: the time-of-day when the user is queried, the rate of querying and the OSN used to query the user. Each USER has three important attributes: their social distance or path length from the application managers, the tie strength between the user and their recruiter and the device used to access the OSN. While an application cannot control the attributes of a user, it can use these attributes to inform which users are selected from the population to receive a particular query.

Table 2: Explanation of Variables

RESULTS 
In this section, we examine the effect of various contextual variables on user message loss and user response times.
Message Loss
We used logistic regression analysis to examine the effect of relevant predictor variables, time of day, online social network, rate of messaging and path length on message loss. We examined both the main effects and interaction effects. Table 3 shows the results of the logistic regression analysis.

Table 3: Factors Influencing Message Loss

As shown in Table 3 the main effects for both OSN and Rate have a statistically significant effect on Message Loss. We find that, on average, Facebook messages have lower rates of loss compared to Twitter. However, as there is a statistically significant effect for the interaction term OSN*Rate we need to be careful in interpreting the main effect. We plotted the interaction between the variables on Message Loss. We find that at a rate of up to two messages per day queries delivered using Facebook have lower rates of message loss relative to messages delivered via Twitter. However, when the rate of messaging increases to 3 messages per day the trend reverses and then converges when the rate of messaging is 4 or 5 messages per day. As such, queries delivered via Facebook are less likely to be lost when the rate of messaging is 1 or 2 messages per day. However, this advantage is lost when the rate of messaging increase to 3 messages per day and beyond. 

The effect of Time of Day on message loss is not significant. However there is a significant interaction between Time of Day and Rate of Messaging. As the coefficient is positive, when the time of day increases the effect of rate of messaging on message loss also increases. When we plotted the interaction we found that Rate of Messaging has the greatest impact on message loss during the Night and Morning periods. However, during the Afternoon and Evening periods response times are relative stable when the rate of messaging is up to 4 messages per day. As such in order to minimise Message Loss it is best to send messages during the Evening and Afternoon when the rate of messaging is up to 4 messages per day. 

There was no significant main effect or interaction effect between Path Length and message loss. We performed sub-group analysis on the effect of Tie Strength on message loss using binary logistic regression analysis. We do not include tie strength as an independent variable in the main analysis as we only have tie strength data for 70% of the participants. The results are reported in Table 4. We find a significant relationship between tie strength and message loss. 

Table 4: The Effect of Tie Strength on Message Loss

Message Response Times
In order to analyse the effect of the relevant predictor variables (Time of Day, OSN, Device, Path Length) on Message Response Time we perform a four-way ANOVA. We examined both the main effects and interaction effects. Due to the difficulty in interpreting interactions with several variables we only interpret interactions involving a maximum of two variables. The results of our analysis are shown in Table 5.

Table 5: Factors Influencing Message Response Time

There is a significant main effect for Time of Day on Response Time. However we also find significant interaction between, Time of Day and Device. In order to interpret this result we plotted the interaction between the variables on Response Time. We find that cellular device have a lower response time during the Morning and Evening, however this trend reverses during the Afternoon period and converge at night. This may imply that users are more likely to access their chosen OSN and then respond to the query on their work based non-cellular device during the afternoon. 

The main effect of OSN was not found to be significant. However, we find a significant interaction between OSN and Rate of Messaging. This implies that the effect of OSN on Response Times is dependent on the rate of messaging. When the rate of messaging is 1, or 2 messages per day messages delivered via Twitter have a lower response time relative to Facebook. However, once the rate of messaging increases beyond 3 messages per day the trend reverses with messages delivered via Twitter having a lower response time. The main effect of Rate of Messaging was also found to be significant however as there is a significant interaction effect we should be careful in interpreting this effect. As such, when the rate of messaging goes above 2 messages per day queries delivered via Facebook have lower response times.  

We find a significant main effect for Path Length. Interestingly when we graph the response times we find that the messages with a lower response time are associated with users of path length of 3 or 4 rather than 1 or 2. As such users of greatest social distance from application managers have a higher response times. It may be the case that such users have self-selected to participate in the experiment due to general personal interest rather than social influence. In addition we find a significant interaction effect for Time of Day and Path Length. However, it appears that the effect of Time of Day moderates the effect of Path Length with queries sent at Night having a similarly high response time for all users regardless of Path Length, which is to be expected. 

DISCUSSION: MAXIMIZING USER RESPONSIVENESS
In this section, we analyse the implications of our results for the designers of manual participatory applications.

On the Presence of Interaction Effects: We find significant interaction effects for a number of variables representing attributes of both how the queries are sent and also attributes of the user. Through consideration of these attributes, applications may more effectively target queries to their user population.
Finding 1: Improving user responsiveness to queries requires full consideration of the attributes of the network within which the application is embedded connecting the application to the user, including both query attributes and user attributes.

The Effect of Online Social Network: We find evidence highlighting the importance of selecting the most appropriate online social network through which to deliver queries. However, we also observed a significant interaction effect between the OSN used and rate of messaging and message loss. Specifically, while Facebook exhibits significantly lower message loss at rates of 1 or 2 queries per day, when the querying rate rises past 3 queries per day, the advantage of Facebook is lost. In terms of response times, Twitter has slower response times at rates of 1 or 2 queries per day, equalizes at 3 messages per day and became faster at 4 or 5 messages per day.
Finding 2: Facebook performs better than Twitter for low-rate querying of users for non-time sensitive information, however Twitter offers superior performance when higher rates of querying are necessary or the requested information is time-sensitive.

The Effect of Device Type: The device variable identifies the device that users responded with using during check-in. It is therefore not possible to examine the effect of device type on message loss. Interestingly, we find a significant interaction effect between time of day and device used on message response time. On average, messages responded to via cellular devices have lower response times compared to non-cellular devices, however this relationship varies significantly based upon the time-of-day. During the morning (06:00-11:59) and evening (18:00-23:59) cellular devices have quicker response times. However, non-cellular devices provide quicker responses in the afternoon (12:00-17:59). Intuitively, both classes of device exhibit slow response times at night (24:00-05:59) when most users are asleep. We argue that this interaction effect is likely due to the work routine of participants. During the morning and evening when users are likely to be out of the office or travelling, a cellular device is the most convenient mechanism for accessing their OSNs. However, during the afternoon when most users are at work, they respond primarily using their desktop or laptop. While the increasing ubiquity of smart-phones is an exciting trend for participatory applications, this result illustrates that non-cellular devices should also be exploited to ensure user maximum user responsiveness throughout the day. 
Finding 3: Users tend to respond quicker using cellular devices in the morning and evening, while non-cellular devices perform better in the afternoon. User response times are slow for both classes of device during the night.

The Effect of Social Distance: It has been argued that online social networks hold great promise for recruiting users to participatory applications (Demirbas, 2010). This would be supported if message loss were influenced by the tie strength between a user and their recruiter. As this is a small-world property, it would enable viral recruitment to operate in a Ôscale-freeÕ fashion. If on the other hand, the social distance between a participant and the application manager (in this case, the authors of the paper) influences message loss or response times then the scale of participatory social applications may be inherently limited by the application managers existing social networks. To investigate the impact of social distance on participation, we created two variables: path length and tie strength. Path-length measures the number of ties (indirect links) between each user and the application managers. Tie strength quantifies the strength of social relationship between each participant and their recruiter. We do not find a significant relationship between message loss and path length. However we find a significant, negative relationship between tie strength and message loss. This is the opposite of what we would expect, as tie strength increases the likelihood of message loss also increases. We do not find a significant relationship between path length and message loss. However we do find a significant effect between path length and message response times. As the coefficient is positive it is again the opposite to what we expect, as path length increases the likelihood of message loss decreases. 
Finding 4: Message loss and User Response times are not negatively affected by social distance between users and their recruiter (tie strength) or users and the application managers (path length). Thus, large-scale viral recruitment for participatory applications may be feasible using online social networks, even in cases where the ties between recruiters and potential users are relatively weak. 
It is our hope that the four findings enumerated above will allow practitioners to maximize user responsiveness in participatory applications, while providing researchers with inspiration for their further study. In the following section we discuss the limitations of our study and directions for future work.

CONCLUSIONS AND FUTURE WORK
Participatory applications have attracted considerable attention from industry and academia as they offer the possibility of a scalable and low cost development approach for large-scale user centric applications. Participatory applications are however critically dependent upon user responsiveness. In this paper, we used an original 30-day study, involving 3,055 messages sent to 70 participants to investigate the effect of query and user attributes on message responsiveness. Through this analysis we demonstrate that (i.) a network perspective that considers the attributes of both users and query messages is an appropriate framework through which to understand user responsiveness. (ii.) Responsiveness is significantly effected by observable query and user attributes such as: device type, rate of messaging, time of day, OSN used, tie strength and path length. (iii.) We identify significant interaction effects between a number of variables considered. Considering the implications of these findings for the research community; our exploratory study provides a first insight into remotely observable attributes of query messages and users impact responsiveness to queries. We hope that these promising initial findings may motivate a new stream of research that builds a deeper understanding of the determinants of user responsiveness in participatory applications. For practitioners, the findings of our study may be used to more effectively target queries to users and thereby improving user responsiveness for participatory applications.

Building on this exploratory study, we intend to undertake two complementary streams of further research. Firstly, research is required to build a better understanding of how online social networks can be used to build communities for participatory applications. Secondly, the distinct characteristics of user responsiveness different message types motivate further research into how the characteristics of specific communication channels impact user responsiveness in social applications.

1) Building communities for participatory applications: while this paper focuses on understanding user responsiveness in participatory applications, the successful use of online social networks to virally recruit a population of 70 users for a 30-day task confirms the potential of online social networks for building participatory applications, thus validating prior work by Demirbas et al. (2010). Further work is needed (i.) to identify the factors that promote successful viral recruitment on OSNs and (ii.) to build an understanding of how tie-strength between participants can be exploited to maximise the reach of viral user recruitment.

2) Effect of communication channels on responsiveness: Using public messages broadcasted to users of the Twitter network, Demirbas et al. (2010) observed a loss rate of 85% and slow response times, with 50% of responses taking longer than 30 minutes. In contrast, our methodology used private query messages sent to individual users of both Facebook and Twitter, achieving a much lower rate of message loss at 50%, but with higher response times, with only 34% of responses received within 30 minutes and a median response time of 1 hour and 13 minutes. Further investigation is required to discover the impact of communication channel on user responsiveness. Specifically this analysis should consider the target scope of queries (individual or broadcast) and the privacy level of queries (private, group or public). 
To support our further work, we plan to design and implement a larger scale experiment in the second quarter of 2014. In this experiment we aim to recruit a larger user community, which enables a statistically significant study of factors affecting both user recruitment and participation.

ABOUT THE AUTHORS
Caren Crowley is a postdoctoral researcher with iMinds-DistriNet, KU Leuven. She was awarded a PhD in innovation systems and industrial clusters from the School of Business & Economics at the National University of Ireland, Galway (NUI Galway). Her research interests relate to innovation and social network analysis.

Rafael Bachiller is a PhD student with iMinds-DistriNet, KU Leuven, Belgium. He holds a masters degree in Automatics, Robotics and Telematics Master from the University of Seville, Spain. His research interests focus on cyber-physical systems.

Wilfried Daniels is a PhD student with iMinds-DistriNet, KU Leuven. He holds a masters degree in Distributed Systems from KU Leuven. His research interests focus on middleware for sensor networks.

Wouter Joosen is a full Professor with iMinds-DistriNet, KU Leuven. He holds a PhD in Computer Science from KU Leuven. His research interests focus on distributed software, middleware and secure software engineering.

Danny Hughes is an Assistant Professor with iMinds-DistriNet, KU Leuven. He holds a PhD in Computer Science from Lancaster University, UK. His research interests focus on distributed systems, sensor networks, component based middleware and peer-to-peer systems.

BIBLIOGRAPHY
1. Lars Backstrom. The Anatomy of Facebook, available online at: https://www.facebook.com/notes/facebook-data-team/anatomy-of-facebook (retrieved 1-14-14).
2. Stephen Borgatti and Daniel Halgin, 2011. ÒOn Network Theory,Ó Journal of Organization Science, volume 22, issue 5, pp.1157-1167
3. Delphine Christin, Andreas Reinhardt, Salil Kanhere, and Matthias Hollick, 2011. ÒA Survey on Privacy in Mobile Participatory Sensing Applications,Ó The Journal of Systems Software, volume 84, number 11, pp. 1928-1946.
4. R. Chen and Y. Sakomoto, 2013. ÒPerspective Matters: Sharing Crisis Information in Social MediaÓ In Proceedings of the 46th Hawaii International Conference on Systems Sciences Ð HIICSÕ13, Hawaii, USA, pp. 2033-2041
5. M. Demirbas,  M. Bayir,  C. Akcora, and H. Ferhatosmanoglu, 2010. ÒCrowd-sourced Sensing and Collaboration using Twitter,Ó Proceedings of World of Wireless Mobile and Multimedia Networks - WowMoMÕ10, Montreal, Canada, pp.1-9
6. L. Deng and L. Cox. 2009. ÒLiveCompare: Grocery Bargain Hunting Through Participatory SensingÓ Proceedings of the 10th Workshop on Mobile Computing Systems and Applications HotMobileÕ09,Santa Cruz, California
7. Don Dillman, Jolene Smyth and Leah Christiansen. 2000. ÒInternet Mail and Mixed-Mode Survey: The Tailored Design MethodÓ. Canada: John Wiley & Smith
8. S. Eisenmann, E. Miluzzo, N. Lane, R. Peterson and A. Campbell. 2007. ÒThe BikeNet Mobile Sensing System for Cyclist Experience MappingÓ, Proceedings of ACM Conference on Embedded Networked Sensor Systems - SenSysÕ07, November 6Ð9, 2007, Sydney, Australia
9. C, Efstraitou C., Leontidas I., Picone M., Rachuri K. K., Mascolo C., & Crowcroft J. 2012. ÒSense and Sensibility in a Pervasive World,Ó Lecture Notes in Computer Science, number 7319 pp.406-424
10. Facebook user numbers, available online at: http://newsroom.fb.com/Key-Facts (retrieved 1-14-14).
11. S. Gaonkar, J. Li, R. Choudhury, L. Cox, and A. Schmidt, 2008. ÒMicro-Blog: Sharing and Querying Content Through Mobile Phones and Social Participation,Ó Proceedings of the International Conference on Mobile Systems, Applications, and Services - MobiSysÕ08, June 17-20, Breckenridge, Colorado
12.  E. Gilbert and K. Karahalios, .2009. ÒPredicting Tie Strength with Social Media,Ó Conference on Human Factors in Computing Systems - CHIÕ09, Boston, USA
13. R. Fielding, J. Gettys, H. Frystyk, L. Masinter, P. Leach & T. Berners-Lee. 1999. ÒHypertext Transfer Protocol -- HTTP/1.1,Ó available online at: http://www.ietf.org/rfc/rfc2616.txt (retrieved 1-14-14).
14. Granovetter, M. 1978. ÒThe Strength of Weak Ties,Ó American Journal of Sociology, volume 78, issue 6, pp.1360-1380
15. Seyed Hoseini-Tabatabaei,  Alexander Gluhak, and Rahmin Tafazolli 2013. ÒA survey on smartphone-based systems for opportunistic user context recognition,Ó ACM Computer Surveys, volume 45, issue 3
16. W. Jia, P. Tso and B. Dai, 2007. ÒAnySense: A Video Communication Architecture for Urban Sensing ApplicationsÓ. Participatory Research Workshop, Proceedings of ACM Conference on Embedded Networked Sensor Systems Ð SenSysÕ07
17. P. Kenis and L. Oerlemans, 2008. ÒThe social network perspective: understanding the structure of cooperationÓ. In S. Cropper, M. Ebbers, C. Huxham and P. Smith-Ring (Eds.) Oxford Handbook of Inter-Organizational Relationships pp. 289-312. Oxford: Oxford University Press
18. A. Mislove, M. Marcon, K. Gummadi, P. Druschel and P. Bhattacharjee, 2007. ÒMeasurement and analysis of online social networks,Õ Proceedings of  Internet Measurement Conference Ð IMCÕ07, San Diego, USA 
19. N. Lane, S. Eisenmann, M. Musolessi, E. Miluzzo and A. Campbell, 2008. ÒUrban Sensing Systems: Opportunistic or ParticipatoryÓ Proceedings of Workshop on Mobile Computing Systems and Applications - HotMobileÕ08, Napa Valley, California
20. E. Milluzzo, N. Lane, K. Fodor, R. Peterson, L. Hong, M. Musolesi, S. Eisenmann, X. Zheng and A. Campbell, 2008. ÒSensing Meets Mobile Social Networks: The Design Implementation and Evaluation of the CenceMe Application,Ó Proceedings of ACM Conference on Embedded Networked Sensor Systems - SenSysÕ08, Raleigh, North Carolina
21. P. Mohan, V. Padmanabhan and R.  Ramjee, 2008. ÒNericell: Rich Monitoring of Road and Traffic Conditions using Mobile Smartphones,Ó Proceedings of ACM Conference on Embedded Networked Sensor Systems ÐSenSysÕ08, Raleigh, North Carolina
22. A. Nazir, S. Razir and C-N Chuah, 2008. ÒUnveiling Facebook: A Measurement Study of Social Network Based Applications,Ó Proceedings of Internet Measurement Computing -  IMCÕ08, New York, USA
23. J. Park and C-W Chung, 2012. ÒWhen Daily Deal Services Meet Twitter: Understanding Twitter as a Daily Deal Marketing Platform,Ó Proceedings of 3rd Annual ACM Web Science Conference Ð WebSciÕ12, Illinois, USA
24. K. Rachurri, C. Mascolo, M. Musolesi and P. Rentfrow, 2011. ÒSociableSense: Exploring the Trade-offs of Adaptive Sampling and Computation Offloading for Social Sensing,Ó The Annual Conference on Mobile Computing and Networking, Las Vegas, Nevada
25. A. Ruzzelli, R. Jurdak and G. OÕHare, 2007. ÒManaging Mobile-based Participatory Sensing CommunitiesÓ Proceedings of Participatory Research Workshop, ACM Conference on Embedded Networked Sensor Systems - SenSysÕ07, Sydney, Australia
26. Twitter user numbers, available online at: http://mashable.com/2012/12/18/twitter-200-million-active-users/ (retrieved 1-14-14).
27. J. Weijia, X. Guoliang, P. Tso and B. Dai, 2007. ÒAnySense: A Video Communication Architecture for Urban Sensing Applications,Ó Participatory Research Workshop, ACM Conference on Embedded Networked Sensor Systems - SenSysÕ07, Sydney Australia



1  Two of the questions were dropped as they dealt with willingness to lend money and helpfulness in getting a getting a job, which were deemed too subjective for the purposes of this experiment.