​​​​ ​​​​

Blog

How to use AI-Assisted mass focus groups

September 20, 2024

This article was written with items taken from Loopanel

AI is transforming the way we conduct focus groups, making the process faster, more efficient, and more insightful. From generating AI focus group questions to analysing transcripts and drafting reports, AI focus group tools are helping researchers at every stage of the process.

 

The research challenge

Anyone who has conducted social research will have faced these challenges:

  • Focus groups are too small to be really representative of the targeted demographic
  • Residential focus groups processes generate a lot of social biases, which can predict what people generally say on an issue, but not necessarily how they feel. These focus groups can be useful to determine messages that play into social conformity, but fail to give insights into other, e.g. values-based, approaches
  • Residential focus groups are difficult to organise, expensive and time-consuming

Recently developed AI-assisted tools can provide alternative or complementary methods, well worth exploring.

How does it work?

  • A question is being asked on a specific App platform to a range of participants (provided by the app randomly, sampled by the provider, or brought by the client)
  • Participants respond in their own words. Using free-form text, they can share their honest thoughts without constraints, giving you access to a rich source of data that supports your hypothesis — or takes you down an unexpected path you hadn’t considered.
  • Participants evaluate responses shared by others. This helps understand how well responses reflect the views of specific segments and discover areas of resonance.
  • Each response is immediately analysed to understand its meaning and how similar or different it is from other responses. This predicts how participants would vote on every response they didn’t see (since voting on thousands of responses would be impossible).
  • Results are generated instantly, organised and analysed by themes, codes, sentiment, etc

Benefits:

One of the most significant benefits of AI in focus groups is its ability to streamline time-consuming tasks. AI-powered tools can assist researchers in creating questionnaires, transcribing audio recordings, and generating reports. 

Also, AI algorithms excel at identifying patterns, sentiment, and key themes within large volumes of qualitative data, so AI can help researchers quickly identify trends and connections across multiple focus groups, enabling them to draw more comprehensive conclusions.

One of the biggest challenges of focus groups is the amount of data you’re dealing with. If you’re running multiple focus groups, you may end up with 100s of pages of transcripts, notes, and observations. While we as people find this volume of data extremely overwhelming, AI can help categorise it for us to make it easier to consume, process, and review. 

Challenges:

However useful it can be, it’s important to remember that AI is not a replacement for human expertise and judgement. While AI can provide valuable assistance, researchers must still use their skills and knowledge to design effective focus group studies, interpret the results, and make informed decisions. As with any new technology, it’s crucial to understand the limitations and potential drawbacks of AI in focus groups, such as the risk of biased or inaccurate outputs. This risk is the result of AI algorithms being developed from data provided by a mainly white Western audience.

Furthermore, AI will not be able to identify determinant cues such as facial expressions, or nuances such as humour, sarcasm, etc. 

Using AI in focus groups involves collecting and analysing large amounts of personal data. This raises important concerns about privacy and data security. Researchers must be transparent about how they’re using AI and ensure they have robust measures in place to protect participants’ personal information. For example, be very careful if you’re using open chatbots like ChatGPT to analyse data—make sure you’ve opted out of letting them use your data for training purposes.

Providers:

A useful list of tools, with their respective qualities, is available HERE.

F l/r aming with Pride

June 29, 2024

TOO PROUD ?

Some years ago, I got a call from a National Council of Muslim organisations, who wanted to react on a poster that was created for the International Day Against Homophobia, Transphobia and Biphobia. This poster was depicting two girls with rainbow-colored head scarfs holding hands.  I was already bracing myself for a tough homophobic conversation but to my utter surprise, the Council’s request was not to protest about the idea of Muslim Lesbians. Actually, they were very supportive of the depiction of sexual diversity among Muslim women (you probably can sense it was a very progressive country, which I won’t name). Their opposition “merely” centered on the attitude of the women: “You see, they explained, the veil is a symbol of modesty. It is not their sexuality that triggers our members, it is the indecency of their attitude[1]”. In other words, they were just “too proud”.

We are in the heart of Pride season and global communications are flooded with the frame of Pride, which triggers just as much exhilaration as acrimony.

While sexual and gender minorities are carrying the Pride frame literally as a flag, opponents voice their (at best) uneasiness by a notion they perceive as (at best) irrelevant, antiquated, and exclusionary.

There are countless moves to try to discredit the “Pride” approach, to make it look like aggressive and indecent instead of what it is: a source of hope and self-respect.

At the other end of the opposition spectrum, we have all raised our eyebrows in disbelief at the idea of a Straight Pride, but daft as it is, the idea reflects that the frame of “Pride” is something that is worth investigating more.

First, a very quick reminder of what framing is: our usage of certain words (and/or images, symbols, etc.) triggers certain pre-existing meanings, representations, notions, values, feelings, etc. that we hold in our minds. They act like shortcuts: The image of a heart triggers the notions of care, pleasure, comfort, and surely other things depending on any particular culture.

Framing is the art of activating these shortcuts, while avoiding triggering the ones we don’t want.

Good Pride Turned Bad

(Hi, Rihanna!)

For sexual and gender minorities, the mere vision of a Pride flag triggers notions of self-worth, protection in numbers, liberation, recognition, fun, and much more. For many of our allies it triggers feelings of celebration, fun, friendship.

But what does “Pride” trigger in the rest of society? At least, what does it trigger in “Western” cultures? This is a big debate, worth of many better articles than this one but let me share a few initial thoughts.

As one of the seven deadly sins, pride has a mixed reputation. On the one hand it is viewed positively as a sign of healthy mental and psychological balance, but take it out of its socially controlled borders and it’ll become hubris, narcissism, arrogance, and vanity.

And this is where the “LGBTQI-Pride” gets tricky: By taking the authority to be naming what can be a source of Pride, and how Pride is allowed to manifest itself, sexual and gender minorities challenge what a given society’s majority  feels to be theirs. The distribution of pride and shame is one of the fundamental instruments to shape societies.

Because of its very nature to challenge the ownership of this instrument, Pride has a disruptive effect that goes well beyond our enemies.

Biases reinforce the frames

The notion of Pride is particularly uncomfortable for people when associated with two mental structures: the slippery slope bias and the zero-sum thinking. The slippery slope bias posits that there is an incremental tendency to everything. Therefore, once Pride has become the new normal it will “escalate” into arrogance and possibly (and that’s really the end of the world as we know it) LGBTQI supremacy.

Somewhat relatedly, the zero-sum game mentality implies that for someone to win something, someone else has to lose it. So if LGBTQI people can be proud, it means someone else has to be ashamed. As crazy as this seems to the rational mind, this might well be going on in the symbol-oriented emotional mind.

Because these biases are not limited to our enemies, we have to be aware of the “side-effects” of our use of the Pride frame, of course not to diminish its prevalence because its presence is vital for most, if not all, of us but to balance these side-effects out.

The “zero-sum game” bias for example can be weakened with a few “Proud Ally” signs, or with representations of the future that includes everyone, not just us. The slippery slope biases can be challenged with images of “normalcy”, like rainbow families.

In public representations of Pride, many different messages are visible, and this often provides a sort of “natural” framing balance. But in our own communications, we are much more selective on what we show, share and shine. It is therefore really important that we keep increasing our awareness of the act of framing, and our savviness in how to strategically frame our communications.

Because changing hearts and minds mostly happens under the surface.

Having said that, HAPPY PRIDE !!!!

As said, this issue deserves a stronger conversation. To share your insights, join the Creative Campaigners Facebook group.

[1] I am not going to dwell over whether this is an acceptable stance. Other conversations are needed for this.

 

Expert talk webinar recording: confronting Disinformation campaigns

May 26, 2023

On the occasion of the launch of Sogicampaigns new free online course on how to fight disinformation, we organised a webinar with experts from different world regions to share insights into challenges and responses in dealing with disinformation campaigns.

We are sharing the recording of this webinar here. It brought together:

Mariam Kvaratskhelia – co-director Tbilisi Pride

Mariami Kvaratskhelia (she/her) is a passionate advocate for LGBTQI rights and equality and is recognized as a prominent leader in the community. As a co-founder and director of Tbilisi Pride, Mariam has been tirelessly campaigning and advocating for the rights of LGBTQI individuals in Georgia since 2015. 

Umut Pajaro – Researcher and consultant on AI ethics, Colombia

Umut (they/them) is a Black Caribbean non-binary person from Cartagena, Colombia working as a researcher and consultant on topics related to AI ethics, and AI Governance focusing on finding solutions to the biases towards gender expressions, race, and other forms of diversity usually excluded or marginalized. They are part of the Internet Society as Chair of the Gender Standing Group. They were speaker and moderator on the Internet Governance Forum, Mozilla Festival, RightsCon, and other tech and digital rights conventions, mainly focusing on sessions related to AI. They also were Mozilla Festival Wrangler 2022 and Ambassador 2022 – 2023, and Queer in AI core organizer from 2020 to 2021. 

Robert Akoto Amofao – Advocacy Manager Pan Africa ILGA, Johannesburg, South Africa

Robert Akoto Amoafo (he/him) is human rights advocate, organisational development coach and certified trainer. He was the Country Director of Amnesty International Ghana from 2018 to 2021, Communications Advisor to the Ministry of Gender, Children and Social Protection in Ghana, and Technical Advisor on the HIV Continuum of Care Project at FHI 360. Robert was a member of the International Advisory Committee of the Power of PRIDE Project run by COC Netherlands, Pan-African International Lesbian, Gay, Bisexual Trans Intersex Association (ILGA) and ILGA Asia. 

Damjan Denkovski – Deputy Executive Director of the Centre for Feminist Foreign Policy, Berlin

Damjan (he/him) is the Deputy Executive Director of the Centre for Feminist Foreign Policy in Berlin, leading the Centre’s work on human rights and international cooperation. He has been with the Centre since 2020 and is mostly curious about how we can strengthen cross-movement alliances and solidarity to counter exclusionary actors and narratives. He comes from a background in civil society capacity strengthening, peace building, and research. 

Moderator:

Alistair Alexander – Reclaimed.Systems, Berlin

Alistair Alexander (he/him) leads projects that explore technology and its impact on the information and physical environment, From 2016 to 2020 Alistair led The Glass Room project (www.theglassroom.org) – with exhibitions worldwide, reaching over 150,000 people worldwide.

At https://reclaimed.systems, recent projects have included: Resonant Signals sonification workshops with ZLB Libraries, Facing Disinformation online training programme, Green4Europe Hackathon for tech sustainability projects in Georgia and Ukraine, digital sustainability for the Gallery Climate Coalition

 

 

IDENTIFYING HARMFUL PERVASIVE NARRATIVES

August 22, 2022

When we communicate, our stories at the surface look like they are perfect. But we are seldom aware of the underlying narratives that these stories also propagate and which can be harmful.

One campaign video in the US featuring a Trans woman telling a perfectly uplifting story was not producing the desired effect in the audience. More research found out that the fact that the woman was featured alone was reinforcing the stereotype that Trans people are isolated and lonely. The set-up of the video contradicted the message, and visuals always win over words. The video was shot again showing the woman surrounded by friends and at last the message hit the spot.

Identifying the underlying negative frames is not easy. The Radcomms network has issues an interesting brief on this. Here are some excerpts :

As storytellers, we may reinforce tropes that perpetuate harmful pervasive ideas even when we don’t intend to. As you craft your story, or work with someone else to share theirs, avoid contributing to the proliferation of harmful, damaging stereotypes and stories. Stories that oversimplify people’s lives are almost always harmful because they lean into these established narratives. They may include:

Deservingness: These are stories that describe an individual’s moral merit. They might focus on factors like hard work or military service to show that they “deserve” success; support; and forms of public assistance like tuition aid, housing, or food assistance. The individual may be presented as an outlier who may easily be described through harmful stereotypes, but is one of the “good ones.”

Hero stories: These stories are about a single individual who, through extraordinary commitment, generosity, and skill, is able to “save” or “fix” people who are suffering the consequences of poverty. Often, this person’s success is presented without acknowledgment of others who participated in collective action.

White saviorism: In such stories, white people provide the help that they believe BIPOC need. These kinds of stories are doubly harmful because they exacerbate privilege and deprive people of agency. They also reinforce narratives that people rather than systems require fixing, and deny the power and importance of collective action.

Fixed-pie or zero-sum: These stories are written from a perspective that there is a fixed pie of resources, and that one person or group’s gain is a loss for someone else. Language that reinforces this narrative might include phrases like “getting ahead” or “left behind”.

Success stories, including “against-all-odds”and bootstrap stories : Success stories are tempting to tell for a range of reasons. Organizations often use them to demonstrate their effectiveness (in which case, they become savior stories) or to gain support from donors. These stories can be harmful because they can create an impression that if anyone can succeed against impossible barriers, everyone should.

Photo credit: Guilia Forsythe – Creative Commons

BUILDING MEANINGFUL CALLS TO ACTION

August 22, 2022

The Radcomms network has issued a useful brief on how to build good calls to action.

 

Powerful stories move people to action. Here’s how to create calls to action that work.

Great calls to action are:

1 Specific: Are you asking people to take an action that is observable and that can support your metrics of evaluation? For example, instead of suggesting that people give to your cause, provide a link and suggested amounts, or a way to give a small amount every month. Instead of suggesting that people “educate” themselves, provide a reading list; links to discussion forums; and links to where they can purchase, borrow, or rent resources.

2 Meaningful toward the issue: Will the thing you are asking people to do actually move the needle on the problem you’ve set out to address? Will they feel that, too? When people feel their actions make a difference, they’re more likely to stay engaged and keep taking action. For example, a call to action like “Stand up to racism” might inspire someone, but it could also leave them feeling like they are acting alone, and their action might seem to them like a drop in the bucket in the face of a large and complex problem. It may be more meaningful for people to share their story of how a policy harms them with an elected official, either through a meeting or a letter.

3 Achievable: Is the thing you’re asking someone to do actually possible? Sometimes, we use the goals of our efforts in place of calls to action. This can leave people feeling overwhelmed or uncertain about what action they can take, and lead them to do nothing at all. Instead of offering a call to action like “End structural racism,” which might leave even the most committed and well-meaning activist at a loss, identify the specific conditions you’re trying to change. How might the community you’re inspiring to act create pressure on those who can change the conditions?

4 Easy: As much as possible, make it easy for people to act. Link to sign-ups for events or rallies, create donation pages that make it easy to give, and provide clear instructions about what you want them to do and why it’s important.

5 Participatory: Create space and opportunities for people to bring their own voices and personalities to accepting your call to action.

6 Something it feels like everyone is doing: Our behaviors as individuals are heavily influenced by our perceptions of what people we see as similar to us are doing. So, your call to action might include language like, “People who care about ending economic injustice are [taking XYZ action].”

7 Activate emotions that keep people engaged: Such emotions might include pride, hope, awe, parental love, and sometimes anger.

123Next ›Last »