Post-Indexed Communication Lag Parameterization
RCTReview
Determining social network connections in online chat platforms requires bounding response times to ensure subsequent messages reflect genuine peer feedback rather than coincidental posting.
Picture this
Think of a busy group text thread where someone asks a question. A reply sent 5 messages later is clearly answering that question, whereas a reply sent 500 messages later was posted long after the original topic was forgotten.
What the evidence says
Peer effect estimates for submission and quality stabilize when the communication lag threshold reaches $\tau \ge 80$ intermediate sentences, converging directly to the unbounded benchmark ($\tau = \infty$, $\beta = 0.533, p < 0.01$ for submission; $\beta = 0.537, p < 0.01$ for extensive quality).
- Who was studied
- N = 1,016 active Milestone 0 completers across 44 interaction clusters in virtual treatment arms.
- How
- Parametric evaluation of two-stage IV peer effect coefficients across maximum communication lags ranging from $\tau = 1$ to $\tau = 100$ intermediate channel postings.
What to do
Set communication lag parameters based on intermediate message count ($\tau \ge 80$) rather than elapsed clock time when constructing adjacency matrices from online chat logs.
From the source
"For a value of $\tau = 80$... the estimated peer coefficient reaches in every case a value very close to the coefficient estimated for $\tau = \infty$."
Peer Networks and Entrepreneurship- a Pan-African RCT.pdf