This blog post is part of a series investigating different demographics and uses of mySociety services. You can read more about this series here. 

I saw a comment on Twitter the other month along the lines of: “is civic tech too boring? It’s dominated by reporting potholes to councils”.

As someone working in civic tech I find this terribly unfair because civic tech is about so much more than that! For instance, we also report dog poo to councils. 

But it’s certainly true that there are a lot of potholes involved. It’s the largest use of FixMyStreet, representing a quarter of all reports. People have submitted over 361,000 reports and over 54,000 photos of potholes. As a result, while the FixMyStreet database represents a fraction of all potholes, it represents one of the largest datasets of pothole reports covering the whole country. 

And while it’s easy to think of potholes as the obsession of people pointing at roads in local papers, they are a serious problem. There are a lot of them, they appear everywhere, cause problems on roads when people try to avoid them, and damage when they don’t.  For cyclists, potholes can be fatal. 

Given that, what does FixMyStreet data tell us about potholes?

How many potholes are reported through FixMyStreet?

Up to the end of 2019 there have been 423,736 potholes or road surface defects reported through FixMyStreet (either .com or a cobrand), with 90,000 reported in 2019. Working from a rough figure of 675,000 actual pothole reports a year, this is around 13% of all potholes reported in the UK. 

A feature of reports to FixMyStreet  is that, while the majority of reports are made by men, there are different ratios in different kinds of reports and categories are often gendered in terms of reporters. Deriving the gender of the reporter from their name, potholes and road surface defects are mostly reported by men, and disproportionately more than the site in general. 

As explored in a previous post, this isn’t an essential gender difference but is likely to result from men having far more cause to encounter potholes. In 2013, men in the UK were on average driving twice as many miles per year as women. 

People who report potholes are more likely to have reported multiple problems than other reports. Most pothole reports are made by people who have reported multiple reports and represent a smaller proportion of single report users than other report types.  

When are potholes reported?

Potholes tend to be reported during the day, but disproportionately compared to other requests around the evening commute. The chart below shows the distribution of reports by time of day, where green indicates the number of reports is higher than the general distribution of FixMyStreet data. 

While potholes are associated most with the start of the year, they occur in smaller numbers all year long. The number of potholes reported through FixMyStreet peaks on the 28th February. 

Where are potholes reported?

While reports in FixMyStreet are less likely to be made in less deprived areas in general, this effect is larger for potholes:

This effect is driven more by reporting being lower than usual in more deprived areas than especially high in less deprived areas:

This pattern was generally similar for the Income and Employment domains of deprivation. This does not necessarily mean there are more actual potholes in these areas, but possibly that people in areas with higher income and levels of employment are more likely to report them. 

Examining reports using the deprivation subdomain that measures difficulty accessing services (GPs, supermarkets, etc) shows a different pattern, where a disproportionate amount of pothole reports are made in areas with the least access to services. 

The access to services measures in Scotland and Wales also reflect that the least accessible places have  a large number of pothole reports compared to the general dataset:

The area with the worst access to services (typically a measure of distance to services) has a disproportionate amount of total pothole reports on FixMyStreet. This doesn’t necessarily indicate this is where most of the potholes actually are, but more remote, less traffic-ed potholes will rank lower in risk-based calculation than those on busier roads, and hence may go longer without fix, and make a report on FixMyStreet more likely.

Repeat potholes

Fixing potholes is a never-ending task, as they are an inevitable result of erosion of roads over time. That said, poor repairs will make the return of a pothole more inevitable than it might be. The issue isn’t just that  the same pothole returns: if a pothole initially formed because the road surface was poor, others are likely to form in the same area too. 

Looking at reports on FixMyStreet up to the end of 2016, for 3% of potholes a new pothole was later reported within 10m between six months and two years after it was first reported (with an average time lag of 15 months). Expanding that ratio to a 20m radius, 7% of potholes had a new pothole reported in the same time range. 

While FixMyStreet’s data on potholes is far from universal, the geographical range gives us better scope than any single local authority’s data to see how reporting of potholes relates to social factors. You can examine this data yourself, on our geographic export, which gives counts of different categories of report by LSOA. 

Photo by Markus Spiske on Unsplash

This blog post is part of a series investigating different demographics and uses of mySociety services. You can read more about this series here. 

When people make their first Freedom of Information request using WhatDoTheyKnow they are sent an email two weeks later, asking them to complete a survey.  This survey has been running from 2012 and in that time has received 6,861 replies. Because this is an optional survey and not a requirement of making a request, this is a small proportion of the number of first time requesters in that time (around 3-4%).  This response rate reflects that the survey is currently quite long and asks questions that, while more useful when the service was new, are now less helpful in understanding its ongoing impact. 

As it’s unclear how representative this sample is of requesters of WhatDoTheyKnow, the overall results shouldn’t be read as authoritative of the user base. What is more interesting is how different groups of respondents use the site in different ways. The data from the surveys has been added to the explorer minisite, a research tool that uses chi-square tests to examine if there is a statistically significant difference in the distribution of responses.

Survey demographics

Looking at the overall picture, the average age of respondents is around 45-54. 

There were more than double the number of male respondents as female respondents. 17% of respondents said they had a disability. Disability is a broad category where self-identification can vary, which makes comparison to national figures difficult. However in 2011, 8.5% of the population of England and Wales were ‘limited a lot’ in their daily life as a result of a health problem of disability, while 9.3% were ‘limited a little’. This suggests that use of WhatDoTheyKnow is not broadly different  from the national picture – however this could be disguising variation within different kinds of disability. 

There is a good spread of income ranges among respondents — but the average respondent has a greater income than the UK median of around £28,000.

43% of respondents were working full time, 10% were working part time and 21% were retired. A majority (57%) were university educated. 

On ethnicity, most respondents declared ‘British’ or ‘English’. 16% were part of a BAME (Black, Asian and minority ethnic) group. Because of the small number of responses over a large range of ethnicities, a second ‘reduced’ option was created by grouping responses that just presents BAME/Not BAME/NA. This would make general trends statistically detectable, but may also disguise trends when different ethnic groups have effects in different directions. 

Over time there is a small number of trends. There is a slow rise in the number of female respondents – from 24% to 33% in 2019. There was statistically a larger proportion of BAME respondents in  2015 and 2016 (19-20%) and fewer in 2012-13 (13%). The number of respondents with disabilities does not show any significant differences between years.

Authority type

BAME respondents are more likely to write to Education, Central Government and Other than the general dataset, and less likely to write to health, local government, emergency services, military services, and media and culture.  BAME survey respondents make up 5% of requests to media and culture and 29% of education respondents. 

Female respondents are more likely than male respondents to write to education and health authorities than the general dataset, and are less likely to write to emergency services, media and culture, transport, and military and security services. Female survey respondents make up 14% of requests to military and security services, and 40% to education. 

Respondents with disabilities are more likely to write to health-related authorities and the emergency services than the general dataset, and less likely to write to transport and education. Respondents with disabilities make up 9% of requests to education authorities and 26% for health authorities.

Retired respondents are more likely to write to environment-related and local authorities, and less likely to write to central government and education. Retired respondents make up 7% of requests to education authorities to 33% for environment-related authorities . 

Reversing the lens to look at one type of authority, respondents writing to education authorities are more likely than the general dataset to be female (but still majority male) and more likely to be part of a BAME group (but still majority white). They are more likely to be below the age of 24 and less likely to be above 55 than the general dataset and (related to that) more likely to be in education and less likely to be retired. 

Message concern

46% of respondents said they were writing on behalf of ‘all people in the community’. This group was more likely to be retired, less likely to be part of BAME group, but more likely to be part of a community group (but not a political group alone). 

20% said they were writing on behalf of themselves/family as well as similar people. 

14% said they were writing on behalf of all people — this group was slightly more likely to be earning less than 12,500, have excellent internet access, and more likely to be involved in political activity (less likely to be part of a community group), to have made FOI requests before, and to make lots of FOI requests.  

13% said they were writing on behalf of themselves or family. This group has a spread on age, but is more likely to be older than 75 (and less likely to be 45-54) than the general dataset. 46% are still university educated, but this is less than the general dataset and this group is more likely to be have secondary or technical college qualifications, and slightly likely to be part of a BAME group than the general dataset (while still majority not BAME). This group is more likely to not be involved in groups or to previously have made requests.

Previous FOI use

The profile of a user who had never made a request before using the site is in many respects similar to other users. This group contains slightly more 25-34 year olds and those in full time education. They are more likely to be making requests where the information is mostly relevant for themselves/family or people similar and more likely to not be involved in community or political groups.

What’s next?

While this survey has found some interesting things about our users, it’s currently overly-long and has a much lower response rate than some of our comparable surveys. We’re looking at the best way of modernising the questions and survey platform to replace this survey, while maintaining continuity with some of the trends identified above. 

Photo by Shawn Ang on Unsplash

This blog post is part of a series investigating different demographics and uses of mySociety services. You can read more about this series here. 

Just as there is interesting information to gain from where people make reports, there are also interesting things to discover from when an issue was reported. 

There are four interesting times in the life of a FixMyStreet report:

1. When a problem happened
2. When a problem was noticed
3. When it was reported
4. When it was fixed

In the FixMyStreet dataset we have lots of information for when a problem is reported, but less about the other times. A follow-up survey gives us some idea if a problem was fixed inside a month –but this isn’t universally responded to. 

Reka Solymosi, Kate Bowers and Taku Fujiyama (2018) examined FixMyStreet data and found some signs that enough reports are made close enough to the time of time a problem is noticed that they show a statistical difference. Using reports of broken streetlights (which should be more noticeable when it’s dark), they showed that more reports were made at night compared to other kinds of reports. 

This analysis is replicated on the Explorer minisite, which shows that more problems with street lights are reported during the winter months; and also that they are disproportionately likely to be reported during darker times of day than other reports as a broken streetlight is more noticeable at night (while other kinds of problems become less obvious). The below graph shows when street light problems were reported. While a fair number of reports are made during daylight (reflecting that not all issues are reported close to when they were observed, or that some street light problems are noticeable during the day), compared to the dataset as a whole the nighttime reports for this category stand out.

Potholes are reported at the start of the year, and disproportionately in the afternoon. Dog fouling is also reported more at the start of the year, but this is more of an early morning report, with a peak as people arrive at work towards 9:00 am:

Some patterns reflect how people’s activity changes when not working. Issues in parks and open spaces are reported more at the weekend, while potholes are reported more during the week. 

While some problems are driven by physical processes that raises their occurrence at certain times of year and their report at certain times of day, other reports result from the activity of other people. Rubbish is reported in the morning, but also has peaks on Sunday (following Saturday night) and Monday, as regular commuters return. 

Similar to the idea that more 311 reports are made in spaces that are contested between different communities, Solymosi and colleagues suggest that reports can also be driven by the handover of the same space between different groups: “The narrative descriptions included with [FixMyStreet] reports reveal that these reports are made by people who are waking up to go to work, and encountering signs of activity that took place in the same location, but at a different time. They see signs of another activity in the space their routine activity pattern takes them through but is incongruent with their current use of this space, and interpret these as a signal disorder, attributing meaning which can result in heightened fear or anxiety.”

For people writing to their representatives on WriteToThem, there are similarly differences in when people write to different kinds of representatives. These might be times people are exposed to something that makes them want to write to their representative, or when they have the time to write.  Compared to all messages sent through WriteToThem, people writing to MPs are more likely to be writing before work and in the late afternoon, while Councillors are sent more messages  between 8:00 am and 4:00 pm. 

While few people write during the night, compared to other types of representative messages are written to Lords more often at night. Looking at the gender of people writing to MPs, the data shows that men are disproportionately likely to be writing at night compared to women (although again, most messages by men are still sent during the day).

Examining the time people make reports helps to create a better picture of when people encounter an issue that a mySociety service might be helpful for, as well as when people have time to do something about it. This suggests possible ways a service could be differently reactive at different times of day and helps sharpen potential research questions.

Photo by Jon Tyson on Unsplash

We’ve previously written about how best to proceed when an FOI request has been refused, but when there isn’t a response at all   that’s a slightly different problem. However, up until now, we’ve treated both in the same way. We’ve now made some changes to reflect the difference.

If you do receive a response, but feel it’s inadequate or that your request has been wrongly refused, there are two ways of contesting the outcome. The first is to ask for an internal review, where the request is reassessed inside the same authority (by a different team or person). The second is to appeal to the Information Commissioner’s Office (ICO).

To reflect this, our approach when providing prompts to WhatDoTheyKnow users has been an escalation ladder — suggesting people request an internal review, and then appeal to the ICO if still dissatisfied. However,  as we learned when talking to the ICO earlier in the year, that this isn’t always the fastest route to getting a response.

An internal review is useful when disagreeing with a decision, but when the issue is that the statutory deadline for a response has passed, and follow-up messages haven’t been answered — a situation our colleagues at Access Info aptly refer to as ‘administrative silence’ — it can be better to complain directly to the ICO.

According to the statutory code of practice, a public body may take up to 20 working days to undertake a review, and so this route is likely to result in further delay, whereas an intervention by the ICO may have a faster result.

So when a request is overdue, our email prompt will no longer suggest that users might want to seek an internal review, but instead we suggest sending a follow-up message to the authority and note that they can appeal directly to the ICO.

However, if you have an issue with the actual decision of a request (for instance, disputing an exemption applied), internal review is the correct first port of call — and in a surprising amount of cases can be very successful. While we don’t have figures covering all kinds of authorities, for requests made to central government, 22% of internal reviews resulted in some change to the original decision (and 9% were completely overturned) and for local government this figure is between 36-49%.

Both internal reviews and appeals to the ICO can be effective methods at redressing disputes around Freedom of Information requests, but it is important to consider which is the right tool for the situation.

—

Image: Ümit Bulut

This blog post is part of a series investigating different demographics and uses of mySociety services. You can read more about this series here. 

WriteToThem is a service that assists people in writing to their representatives. Given a postcode, it lists the associated elected representatives at every layer of government and provides a form to write an email to them.

This can also be seen as a bundle of services. The main use of this website is to write to MPs, but this is just under half of messages ever sent (48%), with most messages sent to representatives in devolved or local government. Different services have different profiles of use and so need to have their effect judged separately.

In 2015, the British Election Study asked whether people had contacted a “politician, government or local government official” in the prior 12 months and found that 17% had. Based on this, over 11 million adults wrote to a representative or official that year — and WriteToThem’s 187,000 emails accounted for 1.6% of this. These results also showed that 20% of men had made contact compared to 15% of women, meaning that 57% of those doing the contacting were men. Extending this into a logistic regression shows that older respondents and those with higher levels of education were more likely to contact, with no significant difference for income and ethnicity once age and education were controlled for.

Demographic profile of WriteToThem users

Looking at the profile of people writing to MPs using WriteToThem, there is an uneven use by different demographics. Over all time,  60% of messages sent have been from men and  60% of people writing had written before. Using the index of multiple deprivation, more messages are sent by better off areas, with 55% of messages being sent by the less deprived half of the country, and 7% of messages coming from the most deprived decile (you would expect 10% if this were evenly divided).

There is a clear linear pattern of greater employment and income in an area being associated with a greater amount of messages sent.  Most of these gradients are slight, but in aggregate the effect is that WriteToThem reflects existing divisions in participation (although there are no good sources for the demographics of people who write to MPs specifically) .

But is this actually a problem? Should a service be judged for the proportion of existing represented groups making use of it, or what it does for the under-represented groups who do use it? WriteToThem has delivered 73,000 messages to MPs from people in the most deprived IMD decile alone, if this has led to dialogues that resolved issues that would not otherwise have happened, this is a positive regardless of whether the same is also true for more people in the least deprived areas. If WriteToThem lowers the cost of contact by making it easier, then it is unsurprising that many of the people making use of it would have made contact anyway — but also included in that are people who were previously unable to engage in the process.

When we look at the result of the survey asking whether a user of WriteToThem was writing for the first time, we can see that people from the bottom three IMD deciles were statistically more likely to be writing for the first time (this is also true when just looking at people writing to MPs, and when just looking at 2018). While generally the number of people using the site for the first time has decreased over time, this decline is demographically uneven and mostly occurs in less deprived areas.

For the complete time-span of the service, 47% percent of survey respondents in IMD 1 (most deprived) were writing for the first time compared to 38% of IMD 10 (least deprived). Looking at just 2018, this was 48% compared to 35%. While the service as a whole is used more by people in less deprived areas, of those using it in less deprived areas it is successfully facilitating a higher proportion of first time contacts.

The local picture

To return to the idea of bundles, WriteToThem is also quietly solving a much harder problem than contacting MPs. While people generally recognise their MP when prompted with a name, local councillors remain far more anonymous. From 2007 to 2018 WriteToThem has helped constituents send 450,000 emails to their local councillors (42,000 in 2018). This service has an effectively even gender ratio (with a female majority in 2018), with more reports coming from more deprived areas (54% by more deprived half).

If we imagine one of these bundled services being a site named “WriteToYourCouncillor”, it is in many respects a model service, with a user base displaying an even gender ratio, and more likely to be used in deprived areas. That in reality it is one function of a more well-used service in terms of numbers somewhat obscures this.

But while it is good to recognise where services are successfully reaching people we want to reach, it is also important to think about volume and overall impact.  One issue with a service used more by men or in better off areas might be if it shapes how resources are deployed or provides a false shape of the views of constituents (and emails received are certainly used by MPs to build a picture). Even a service that adequately represents under-represented groups may be ineffective if it exists in a wider ecosystem that does not.

At the moment, the systematic effect of any bias in WriteToThem outputs is marginal as WriteToThem accounts for a small fraction of parliamentary mail.  While the amount of physical mail entering the Houses of Parliament each year has decreased steadily, in 2018 it was still 24 times larger than the number of emails sent to MPs via WriteToThem. The average MP received 94 emails via WriteToThem in 2018; most MPs would receive more than this through other means in a week.

Returning to the British Election Study finding that 57% of contacting in 2015 was done by men, the equivalent figure for WriteToThem as a whole in 2018 was 55%. Being generous and bearing in mind the previous finding that the method used to assign gender from name undercounts women, this could be seen as a marginal improvement on the real world. However, it would be a marginal improvement in a pool that only represents 1.6% of the total amount of number of messages.

Defining success

Based on the above, we can think about three different kinds of ‘success’  of a civic tech service in serving under-represented groups:

Relative – The service improves under-representation relative to the current standard. e.g. a service where 60% of usage was by men is an improvement over an offline status quo of 70%.

Absolute – The service adequately (or over-) services under-represented communities to what would be expected based on their numbers in the general population.

Systematic – The service successfully services under-represented communities and is successful enough that this redresses issues of representation in competitor services/methods.

Working with these, we could say WriteToThem is a success on a relative level, servicing people in more deprived areas more than they would have been otherwise (larger proportion of first time writers), but not to the proportion of the population these groups represent.

The “WriteToYourCouncillor” part of the bundle is  a success on an absolute level, providing a relatively even amount of representation, with a slight weight towards groups who typically make contact less often.

But neither really makes a dent systematically. They may be redressing inequalities of access for individual users (which is good), but cannot significantly adjust inequalities in volume of messages and the corresponding perceptions of problems.

Making a dent in this problem is outside the scope of WriteToThem — and probably should be. While you can imagine a future where WriteToThem continues to lower the barrier to contacting representatives,  this is likely to create new users from currently-represented groups for each under-represented person successfully reached. Targeted interventions and partnerships with other organisations can avert this problem in terms of helping individuals make contact about their issues but turning the problem around, this is a platform that is unlikely to provide a balanced view of opinions and priorities of constituents.

If it is a problem that representatives have systematically skewed visions of the problems and views of their constituents, is an email platform that requires citizens rather than representatives to do work the best way to address that? A civic tech solution to this problem might look more like Consul (or similar general participation platform) than WriteToThem – but even explicitly designed online platforms still risk being skewed towards the online and present members of the community. Exploring better forms of local participation is something currently being explored through our Public Square project.

This blog post is part of a series investigating different demographics and uses of mySociety services. You can read more about this series here. 

Indices of deprivation are useful for mapping social phenomena onto geographic data. For a series of domains (in England: income, employment, health, education, skills and training, crime, barriers to housing and services, and living environment) all Lower Super Output Areas (LSOAs) are ranked from most deprived to least deprived. From these the Index of Multiple Deprivation is created  — which helps to illustrate which areas of the country suffer from multiple different negative factors.

The indices of deprivation are compiled separately for England, Scotland, Wales and Northern Ireland. While they cannot be combined, they do often illustrate similar measures and so are useful for cross comparison. As most FixMyStreet reports are made in England, more subtle patterns in how deprivation and reports are linked can be detected from this larger set of data.

The Explorer minsite uses the Index of Multiple Deprivation (IMD) and respective domains to understand how reports for different categories of FixMyStreet report are distributed and explore how deprivation affects reporting. This page shows the categories that are more likely than the general dataset to be reported in the lowest IMD decile (most deprived) and this page shows the categories that are more likely to be reported in the highest IMD decile (least deprived).

Missing reports

As examined in previous research, the most important finding when examining deprivation is the suggestion that there are reports that should be being made that aren’t. The Explorer minisite shows that reports of dog fouling have a peak in the middle deciles, but this does not reflect the real world incidence of dog fouling, which found that the most dog fouling was found in the bottom two deciles.

Even when actual incidence of problems is higher in more deprived areas, the reporting rate can be lower — any picture based on self-reporting is likely to have a large set of missing data. In the case of dog fouling, this means information about hotspots is not communicated to enforcement. In other cases it might mean road defects unfixed, or fly-tipping uncollected.

Exploring domains

While previous explorations of deprivation and FixMyStreet have used the index of multiple deprivation alone, the Explorer minisite lets you see how the distribution differs on each of the domains of deprivation. For instance, looking at reports of rubbish, we can see that while generally there are more in the bottom 50% of IMD deciles, there is a stronger relationship against the crime domain.

Rubbish vs Multiple Deprivation

Rubbish vs Crime IMD Domain

Examining the data for dog fouling shows that the peak in the mid-deciles is even clearer when mapped against income deprivation than for multiple deprivation. The income domain continues to show that compared to the general dataset there are fewer reports in the higher deciles than might be expected.

Abandoned vehicle reports have a scattered relationship with a few different factors, but the association with crime is much less noticeable than the association with lower housing costs. Problems with drainage generally are more reported in less deprived areas, but when focusing on access to service deprivation, they are concentrated in the most deprived areas.

Breaking down by the different domains that make up the index of multiple deprivation lets us better understand what factors are driving either problems or the reporting of problems. This in turn helps to frame questions to ask about what is driving these different uses of FixMyStreet.

Photo by v2osk on Unsplash

This blog post is part of a series investigating different demographics and uses of mySociety services. You can read more about this series here. 

Greater use by men than women is common across mySociety services. Looking just at people who had gendered names (78% using UK data from OpenGenderTracking), 38% of FixMyStreet users were women. However, because women are less represented among super contributors (users who make many reports), only 29% of reports were submitted by women. There has been a consistent year-on-year increase in the proportion of reports made by women (34% in 2018), which at the current rate will reach parity in 2025.

But what are the impacts of this? Where crowdsourced websites have a gender disparity and different genders participate differently, this leads to a difference in outcomes. For OpenStreetMap, Monica Stephens (2013) found that in discussions around proposed new categories of locations, strong distinctions are made between “swinger club, a nightclub and a brothel”, while a 2011 feature of “childcare” was debated and rejected on the grounds it was too similar to the existing “kindergarten”. If contributors are on the whole “very aware of the complexities of sexual entertainment categories, but oblivious to the age specific limits of childcare providers”, this makes the map less useful to the large potential group of users with differing priorities.

This is not an unfixable problem (and in this specific case, quickly was –  childcare was added to OpenStreetMap as a category in 2013) but reflects that crowdsourced websites and datasets reflect the interests of the people who volunteer their time towards them. In an article for CityLab about efforts to increase the number of female cartographers working on OpenStreetMap, Sarah Holder writes that:

Doctors have been tagged more than 80,000 times, while healthcare facilities that specialize in abortion have been tagged only 10; gynecology, near 1,500; midwife, 233, fertility clinics, none. Only one building has been tagged as a domestic violence facility, and 15 as a gender-based violence facility. That’s not because these facilities don’t exist—it’s because the men mapping them don’t know they do, or don’t care enough to notice.

However, as an Open Street Map contributor noted below the original version of this article, shelters for those escaping domestic violence present a particular challenge: openly mapping their locations make them easy for everyone — including the perpetrators — to locate. As such, refuges themselves may not want to be listed. While some services predominately used by women are under-mapped, others are ill-suited to an open, map-based form of discovery. For a more detailed exploration around issues of providing information for victims of domestic violence, see the Tech vs Abuse research findings.

Zoe Gardner, Peter Mooney, Liz Dowthwaite and Giles Foody (2017) found that as well as differences in the scale of activity, men and women also behaved differently in the kinds of ways they added to OpenStreetMap, with men more likely to modify existing features and women more likely to add new data in a few categories. Specific categories of label had different rates of contribution, with women more likely to add labels in the ‘building category’ (67% for women vs 35% for men), while men were more likely to make modifications to the highway category (39% for men vs 23% for women).

FixMyStreet

For FixMyStreet Reka Solymosi, Kate Bowers and Taku Fujiyama (2018) found a similar difference in behaviour in terms of the categories of reports submitted by men and women and found a rough “driving vs walking” divide:

On first glance it appears that men are more likely to report in categories related to driving (potholes and road problems), whereas women report more in categories related to walking (parks, dead animals, dog fouling, litter).

This was replicated with non-anonymous data internally.  The methodology used in this paper is applied through the Explorer minisite to a wider dataset, and the gender difference in categories can be seen here. This uses an analysis that derives likely gender from first name, which is not 100% accurate and cannot derive a gender for all users. However, for broad differences, the data is sufficient – a comparison to a group of reports where reporters disclosed gender found that the derived ‘male’ group contains around 4% misallocated women, while the derived ‘female’ group contains about 1.5% misallocated men. The unknown group splits roughly 50/50, but leans towards containing more women (53%).

As women are still minority users of the site in general, categories are noteworthy if they have a greater proportion of women than the site as a whole — even if this is below parity.  For instance, women make up 40% of reports of overgrown trees, which means more are reported by men — but this is higher than use of the site as a whole by women. Women make fewer reports (and account for more first time reports than repeat reports), but these reports are focused on different categories to categories that are more reported by men (such as potholes, 74% of which are reported by men).

Encountering problems

When men and women are moving through the world differently, they are encountering different kinds of problems. In 2013, men in the UK were on average driving twice as many miles per year as women. Given this, it’s not unreasonable for men to be encountering and reporting many more potholes.

Surveys in Scotland and England suggested higher rates of littering by men and lower acceptance of littering by women — which is reflected in a slightly higher than expected number of reports of litter from women. Women make more walking trips (269 to 240) over a cumulative longer distance (10 miles more per year). Given this it would not be unreasonable for women to be encountering slightly more littering, pavement defects, dog fouling and other walking problems.

This difference is especially true for women aged 30-39 as “women in their thirties make four times as many escort education trips [school runs] than men of the same age, and walking is the most common mode used to make these trips”. Looking at reports of littering in England – reports by women are on average 154 meters (95% confidence between: 138, 171) closer to a school.  This isn’t saying that all reports of littering are made by women doing the school run, but possibly enough that it shows up as a difference in the data.

Reporting problems

In 2018 women made up 36% of reports related to rubbish — but this is masking different gender balanced on different kinds of waste. While there are very few reports of ‘discarded syringes’, three-quarters of these are made by women. Reports related to ‘leafing’ and ‘litter/litter bins’ are near parity (49%, 46%).

In 2016 there was an experiment on the  homepage of FixMyStreet.com to see if changing the  prompted categories from a focus on road problems to a prompt on parks and open spaces and changing the imagery on the homepage (happy families rather than the default “B&Q” colours and spanners) led to an increase of reports by women. There was no difference found, suggesting that the problem was more complicated than women being put off by the design. This did however change the distribution of various categories (fewer pothole reports and more reports of issues with street lights) with no shift in the gender ratio.

For reports made by co-branded websites (instances of FixMyStreet running as part of a council website), reports by women are better represented, making up 42% of reports.  This is a reminder that more than the technology is important, the perceived “officialness” and discovery routes are also important. Certain kinds of users may be more willing to use a third-party tool than the official website.

What does this mean?

If civic tech makes certain kinds of government contacting easier to do, but those forms of contact are more likely to be problems experienced by men, this may have the effect of shifting the provision of services. In the longer run, uneven reporting may entrench perceptions of public interest and respective budgeting for different areas of service.

That men and women experience their environment in different ways and so experience different problems makes this problem both important and difficult to resolve. Understanding FixMyStreet as a bundle of services gives a framework to examine the problem. Viewed this way some services (Report Potholes) are performing about as you’d expect, while others such as Report Litter are lagging. This suggests a different set of experiments to investigate the problem than a generic ‘women use FixMyStreet less’ problem suggests.

It also suggests that reaching greater gender balance in services may involve seeking out different kinds of problems. The issue is less getting more pothole reports from women but that there are neighbouring services that fulfil the same ‘ease contact with government’ role that women would be far more likely to use.

Photo by Olesya Grichina on Unsplash

This blog post is part of a series investigating different demographics and uses of mySociety services. You can read more about this series here. 

The FixMyStreet section of the Explorer mini-site helps explore the relationship between demographic features and FixMyStreet reports.

In one use case, it maps the location a report was made to a ‘neighbourhood’ sized area, and then in turn to sets of statistics measured against those areas — most importantly, the indices of multiple deprivation.  These areas are Lower Super Output Areas (LSOAs) in England and Wales, Data Zones (DZ) in Scotland and Lower Output Areas (LOA) in Northern Ireland (although NI is not covered separately in the Explorer site due a relative lack of data). These can be seen as equivalent to census tracts in the US and each LSOA has a population of around 1,500 people, while Data Zones have around 500-1000 people.

While this statistical unit feels neighbourhood-sized and so is used to examine data for effects that may result from being in the same neighbourhood, the approach has the significant problem that what people on the ground perceive as their “neighbourhood” is unlikely to exactly overlap onto the statistical unit. On the edge of a LSOA, even a 50m radius around a home will cross into another statistical area.

Making the problem worse is that the idea of a neighbourhood is very variable. People can disagree with each other about the boundaries of their area. Claudia Coluton, Jill Korbin, Tsui Chan, Marilyn Su (2001) found that when citizens were asked to draw the boundaries of their neighbourhood these very rarely aligned with US census tracts. As the gif in this tweet shows a set of citizen-drawn boundaries for Stoke Newington in East London, and while there is a clear core, there is substantial disagreement between residents about the size of this area.

Laura Macdonald, Ade Kearns and Anne Ellaway (2013)  found that residents in West Central Scotland had a different perception of how well placed they were for ‘local’ amenities compared to the geographic distance. This reflects that what was viewed as local from the outside might not be viewed the same way by locals: there is a context gap that just cannot be bridged at this scale of analysis.

Understanding of neighbourhood effects is often positioned in terms of guardianship of a home area, and this means that certain kinds of reports might be more apparent in areas where these boundaries are less clear — leading to conflict. Joscha Legewie and Merlin Shaeffer (2016) used New York 311 calls to demonstrate that complaints about blocked driveways, noise from neighbours and drinking in public were more frequent on the boundaries of areas with differing demographics. This can also be seen in the idea that complaints about dog fouling are used for score-settling between neighbours in Chicago. Complaints can be about conflicts as well as actual problems reported.

In a related problem, Alasdair Rae and Elvis Nyanzu show in some areas the most deprived 10% of areas and the least deprived 10% are not far from each other. This means that relationships between reports and the features of deprivation might be harder to detect. The less homogenous the area, the greater the chance that features affecting how likely a person is to report will result in reports in a LSOA that is substantially different from their ‘home’ area.

This blog post is exploring a potential problem with the explorer minisite methodology. A big part of what the explorer site is doing is trying to show how much different kinds of reports are “explained” by different local features — but because of various forms of fuzziness the differences it detects may be less sharp than actually exists. In general, however, not detecting things that are there is a better problem to have than the opposite.

This blog post is part of a series investigating different demographics and uses of mySociety services. You can read more about this series here. 

A common feature in websites and services where users generate data is that a small amount of users are responsible for a large percent of the activity. For instance, 77% of Wikipedia is written by 1% of editors (with most of that being done by an even smaller fraction) and for OpenStreetMap 0.01% of users contribute a majority of the information.

This also applies to plenty of offline activities — for instance, half of the 25,000 noise complaints about Heathrow Airport were made by 10 people. People who dedicate significant time to an activity can quickly outpace a much larger group who only use the service once.

For FixMyStreet (where people report issues like littering and potholes to local authorities), the top 0.1% of users made 16% of the reports and 10% of users account for 62% of reports. Starting from the most prolific users, increasing the number of users by a factor of 10 roughly doubles the number of reports:

• 418 users (0.1%) account for 224,775 reports (16%)
• 4,181 users (1%) account for 470,384 reports (33%)
• 41,814 users (10%) account for 881,481 reports (62%)

This reflects that at any scale in the data, around half the activity is happening in the top 10%. Overall, two-thirds of users made only one report — but the reports made by this large set of users only makes up 20% of the total number of reports.

This means that different questions can lead you to very different conclusions about the service. If you’re interested in the people who are using FixMyStreet, that two-thirds is where most of the action is. If you’re interested in the outcomes of the service, this is mostly due to a much smaller group of people.

Reka Solymosi (2018) investigated the behaviour of the top 1% of reporters and found that they tended to report a wide range of categories: only “16 of the 415 contributors reported only one type of issue. The other 399 reported issues in more than one category” with an average of six categories. These also tended to cover a wide area and “there were only six people who reported in only one neighborhood [LSOA], fewer than the number of people who reported in only one category. The other 409 contributors all reported in at least two neighborhoods”. Solymosi finds four clusters of these super-contributors:

• Traditional guardians – these report in a small number of neighbourhoods covered but represent the largest number of users.
• Large-neighbourhood guardians – Report in a larger number of connected neighbourhoods.
• Super-neighbourhood guardians – People who report in a high number of connected neighbourhoods; this is the largest group.
• Neighbourhood agnostic guardians – reports are made in disconnected areas.

Collectively, this can have a wide impact — 18% of LSOAs in England have at least one report from a user who has made more than 100 reports (which is only around 900 people).

Looking at the general picture through the Explorer minisite, it’s not just that serial reporters report widely; certain kinds of reports are more likely to be made by users who are reporting more issues:

Incivilities, rubbish, road safety and bus stop damage are all categories more likely to be reported by users who have made 50+ reports. While users who make lots of reports tend to make reports across a few categories, they are often specialised in their output.

59% reports of flyposting, 57% of graffiti, 52% of litter problems are made by users who have reported more than 50 times.

It’s important to remember that these aren’t hard divides. Single report users are less likely to report potholes than serial reporters, but it is also true that one in five people who only report one issue report a pothole.

For the bundle model of understanding FixMyStreet, thinking about this group of super contributors is important, because they represent a minority of users, yet generate most of the value and impact of the site.

But this comes with a cost. People living in the same area as super contributors benefit from their efforts – but where these super contributors have different concerns or priorities from the area as a while this might shift the outcomes of the service.

As Muki Haklay argues:

The specific background and interests of high contributors will, by necessity, impact on the type of data that is recorded. This is especially important in VGI [volunteered geographic information] projects where the details of what to record are left to the participants.

Where resources are allocated on the basis of data generated by a service, the behaviour of this small group can have an outsized effect. Future blog posts in this series will explore what this looks like in practice.

This blog post is part of a series investigating different demographics and uses of mySociety services. You can read more about this series here.

A key question when looking at the role of the internet in civic life is whether it changes the demographics of who participates; or whether it simply changes the methods by which already engaged citizens participate. The two sides in this argument can be described as mobilisation and reinforcement.

The mobilisation argument says that the internet reduces the cost of communication and action, which means that more people can be involved and access becomes more broad.

The reinforcement argument says that the reduced costs of connectivity will mostly reinforce existing participation divides, making it cheaper for people already engaged to participate, but not necessarily reaching disengaged people.

This is a fundamental question for civic tech: how are these online tools used? Are they mobilising everyone or just providing more efficient processes for people who are already engaged?

This is explored in mySociety’s 2015 report Who benefits from Civic Technology?, and is a recurring question in much of our research since, such as our work on FixMyStreet, and digital technologies in sub-Saharan Africa.

Two themes we are currently investigating in this area are proxies and bundles.

Proxies are where services are used by intermediaries, on behalf of — and bringing benefits to — others: for instance, where charities engage in more effective lobbying as a result of free access to TheyWorkForYou, or where case workers find it easier to identify and write to a client’s local councillor using WriteToThem.

Bundles are about exploring how different groups of users use a service in different ways, to such an extent that one service can in fact be understood as a bundle of services serving different kinds of users.

This is the first in a series of blog posts investigating  bundles.

A common finding across mySociety services is that most people only use “transactional” services (like WhatDoTheyKnow, FixMyStreet or WriteToThem) once, to do one thing. Repeat users make up a minority of users (even if they account for the majority of actual usage).

From a technology point of view or an organisational point of view, it makes sense to understand that there is a website called FixMyStreet.com run by mySociety. But from the point of view of the majority of users, it makes sense to think of a website like FixMyStreet as dozens of different services, most of which they will never use. For one user,  FixMyStreet is a tool for reporting potholes, for another it is for reporting littering. Similarly, WriteToThem is most often used as a tool to write to MPs — but the profile of people who use it to write to their local councillors is very different.

Some services in a bundle are used by a different demographic to other uses of the same website. Understanding how to encourage FixMyStreet use in underrepresented groups requires an understanding of how there are already differences in usage across all the “services” in the FixMyStreet bundle.

To get more information about these different uses of a website, we’ve built a mini-site that helps to explore basic demographic information about each use type. Starting with FixMyStreet, personal information (names) have been anonymised and converted to gender (approximately), while coordinates are grouped into Lower Super Output Areas (LSOA) — geographic areas commonly used for statistical purposes. This means that we can look at a general, anonymised set of data representing people making FixMyStreet reports, and match this grouped data against various measures of deprivation.

Understanding more about these different patterns of users suggests possible ways a service can be used and helps sharpen new research questions.

When examining uses of one element of a bundle, the key question is whether the pattern observed reflects just the individual, or the overall pattern of the bundle. To answer this, a chi-square test is used to tell if the distribution of a sub-use of the site is different to a statistically significant extent to all other uses of the site (this method was inspired by an analysis of gender of reporters in Reka Solymosi, Kate Bowers and Taku Fujiyama’s 2018 paper on FixMyStreet). The groupings of categories in FixMyStreet use Elvis Nyanzu’s meta categories.  The mini-site highlights in red and green areas where a distribution differs from how patterns on the site as a whole respond.

We’ll be writing a number of blog posts over the next few months covering things we’ve learned from the mini-site. The first two are already up (and linked below):

Blog posts:

• 

  Can you believe we’ve reached Peak Pothole Day already?

  • Alex Parsons

  February 2020

• 

  Who uses WhatDoTheyKnow?

  • Alex Parsons

  February 2020

• 

  When people report issues on FixMyStreet

  • Alex Parsons

  January 2020

• 

  Assessing success in Civic Tech: Measures of deprivation and WriteToThem

  • Alex Parsons

  October 2019

• 

  How do different forms of deprivation affect FixMyStreet reports?

  • Alex Parsons

  October 2019

• 

  How men and women use FixMyStreet differently

  • Alex Parsons

  October 2019

• 

  Understanding what makes an area different

  • Alex Parsons

  August 2019

• 

  Data Explorer

  • Alex Parsons

  August 2019

• 

  Service Bundles

  • Alex Parsons

  August 2019

• 

  What's a neighbourhood?

  • Alex Parsons

  August 2019