The research behind Spot

Spot is based on over 40 years of research on best practices for interviewing people about emotional events.

By combining best practices from the cognitive interview with natural language processing, Spot helps employees capture the most important details about their experiences, on their own terms.

New research

We surveyed over 1,000 people about witnessing harassment & discrimination at work.

Read the witness report
Colorful wave transporting screens and other decorative elements

Cognitive interview

Conversations with Spot are highly structured, using best practices from the cognitive interview method to gather thorough, accurate detail.

Extensive research has shown that the cognitive interview, developed originally for law enforcement, is an effective way to gather information about important emotional events.1 In line with this technique, Spot always asks open-ended, non-leading questions, prompting employees to remember and document all the essential elements of their experiences.

Colorful wave transporting screens and other decorative elements
Colorful wave transporting screens and other decorative elements

Natural language processing (NLP)

Spot uses natural language processing to identify important details mentioned in an employee’s account and then probe more deeply. This type of probing reflects the cognitive interview practice of asking open-ended questions based on details mentioned by the person being interviewed.

Since humans tend to introduce bias, a bot combined with NLP may be able to conduct a cognitive interview more rigorously than a human.

1E.g., see Memon, A., Meissner, C. A., & Fraser, J. (2010). The Cognitive Interview: A meta-analytic review and study space analysis of the past 25 years. Psychology, public policy, and law, 16(4), 340.

BARRIERS

Overcoming barriers to reporting

At least 70% of employees who experience harassment or discrimination at work never report it.2 Research shows that talking with a human is related to many of the main barriers to reporting.

People who create Spot reports have the option to submit their reports for research. We remove any identifying details from these reports and then perform qualitative research to identify common themes. We also conduct in-person and online experiments with participants.

Overcoming barriers to reporting

In our analysis of hundreds of Spot reports contributed for research, employees commonly cited these main reasons for not reporting:

  • Fear of retaliation or other negative consequences
  • Fear of not being believed
  • Worry about being judged when they report
  • Feel it’s embarrassing to report
  • Know there is a conflict of interest

To learn more

Witnesses face similar barriers to reporting

After reviewing reports submitted to Spot for research, we found that 60% of incidents had at least one witness.

Because of this initial finding, we partnered with relevant NGOs and nonprofit organizations to conduct the first large international study that sheds light on the witness experience. Results from over 1,000 participants show that:3

  • Most witnesses (64%) told friends and/or family about the incident
  • Just under half (46%) shared what happened with a colleague. Of those who told colleagues, 54% told a member on the same team.
  • And yet, most witnesses (76%) say that they never reported the incident to someone in HR.

Most participants stated that they were afraid of the consequences, regretted not speaking up, and would have spoken up if they had known how.

AI BOT

Comparing Spot to other reporting methods

Capturing more detail, with greater accuracy

Research participants shared 29% more accurate detail when describing a sensitive incident to Spot, compared to speaking with a human.

We conducted in-person and online experiments4 comparing Spot to other reporting methods, including talking to a human, writing a free-form note, filling out an online form, and using a bot without AI. Conversations with Spot resulted in more detailed reports, with fewer errors. Reports written by a human who listened to someone’s account of a sensitive experience included 51% more errors than reports created with Spot.

Identifying reporting method preferences

In original research, 83% of those who used an AI bot stated they’d prefer talking to an AI bot, rather than a human, about harassment and discrimination. In contrast, participants who used other online options, including email-style reporting or questionnaires, did not prefer them over talking to a human.

Overcoming barriers to reporting

ETHICS

Spot’s research team is committed to producing original, peer-reviewed research that improves the world’s understanding of harassment and discrimination at work.

Our research studies are always reviewed by a university research ethics board, are pre-registered, and adhere to open data practices. These practices help us reduce bias and ensure ethical integrity.

3Pre-registration and materials here: https://osf.io/7xsau
4Pre-registration here: https://osf.io/rezf6