Dr Alex Goddard
Fellow in Psychological and Behavioural Science
London School of Economics
a.j.goddard@lse.ac.uk
Dr Alex Goddard is an LSE Fellow in Psychological and Behavioural Science. His research focuses on the quality of dialogue: how people reach consensus, maintain common ground, and repair norm violations. His research draws on deliberative theory, social psychology, and conversation analysis to study text-based interaction across contexts — online communities, complaints, organisational feedback, and human–AI exchange. He develops AI-assisted computational tools and psychometric methods for measuring interaction dynamics at scale: from norm violations on Reddit to hospital responses to patient complaints. Underlying all of this is a single question: what makes a good and bad conversation?
In addition to his research, Alex has worked professionally in psychometrics and has consulted for various organisations across different sectors. He currently teaches on the MSc in Social and Cultural Psychology and Social and Public Communications at the LSE.
Research Areas
-
Quality of dialogue
People easily recognise good dialogue in their own lives, but what actually makes dialogue good is highly contested. Alex's research addresses this by studying how people negotiate and repair problems (e.g., misunderstandings, conflict) in real-world dialogues. Through these analyses, he aims to improve how we define and measure dialogue quality.
-
Conversational repair
When people misunderstand each other in conversation, they tend to fix it in predictable ways; a process known as conversational repair. Alex's research examines whether these mechanisms extend beyond misunderstandings to norm violations — toxic comments, complaints, hostile exchanges. He studies this across organisational complaint responses, online communities, and human–AI interaction.
-
Text measurement in psychology
Psychologists increasingly use text as data, but turning unstructured language into reliable measures is harder than it looks. Alex developed RAMP (Repeated Adjustment of Measurement Protocols), an iterative trial-and-error method for building text classifiers that prioritises validity at every stage: ensuring tools measure what they claim to measure. He applies RAMP across his research domains — from classifying repair sequences on Reddit to measuring corrective responses in hospital complaint letters — and collaborates with other researchers to test and extend the method.
Publications
-
RAMP: The Repeated Adjustment of Measurement Protocols method for developing textual measures in psychology
Goddard, A., Gillespie, A. · Psychological Methods · 2025
-
Conversational repairs on Reddit: Widely initiated but often uncompleted
Goddard, A., Gillespie, A. · PLOS ONE · 2025
-
Validating the use of large language models for psychological text classification
Bunt, H., Goddard, A., Reader, T., Gillespie, A. · Frontiers in Social Psychology · 2025
-
Textual indicators of deliberative dialogue: A systematic review of methods for studying the quality of online dialogues
Goddard, A., Gillespie, A. · Social Science Computer Review · 2023
In progress
-
Measuring Corrective Culture in Hospital Responses to Patient Concerns about Safety and Quality Raised Online
Goddard, A., Reader, T.W., Gillespie, A. · Safety Science · revise and resubmit
-
"We don't ban you immediately for an offense": Reddit moderators reveal the importance of norm violations for maintaining the quality of online dialogue
Goddard, A., Bunt, H., Gillespie, A., Heitmayer, M.
-
Opening the black box: Safety voice and listening underpin team situation awareness and aviation accident severity
Pandolfo, A.M., Goddard, A., Reader, T.W., Gillespie, A.
-
Feeling the Friction: Developing and validating text classifiers for sludge in consumer complaints
Chesterfield, A., Gillespie, A., Goddard, A., Krpan, D.
// Feeling the Friction · manuscript in preparation█