Tuesday’s claim was that listeners hear a slow yes as a reluctant one. Today the room tests it, with stopwatches, on itself.
Before the lab#
Bring a phone with a stopwatch and a lap button. The tutor can lend one. Work in threes: asker, responder and timer. You can take a quiet station at the edge of the room; tell Tamsin if you need the prepared-data alternative below.
Run sheet#
Times below run from the 14:05 start; finish at 14:55.
| Time | What happens |
|---|---|
| 0:00–0:06 | Learn the timer procedure below. Make six folded cards, one for each invitation × delay combination. Put the invitation outside and straight away, count one or count three inside. Shuffle before each asker starts. Do one unscored trial. |
| 0:06–0:24 | Round 1. Each asker reads the outside of a card; only the responder sees its delay. The responder waits, then says only “yes”. The asker rates willingness from 1 to 5 before seeing the measured time. Use all six cards, then rotate every role. |
| 0:24–0:34 | Round 2. Each asker has two more trials, one per invitation. The responder begins “hm, let me think” immediately, then waits a count of three before “yes”. Rotate after two trials. |
| 0:34–0:44 | Calculate each asker’s mean rating for the four conditions. Put the group means on the board, alongside the range of individual ratings. Keep the response and yes-onset times separate. |
| 0:44–0:50 | Compare what happened, including no change or an unexpected direction. Each person writes one finding about these trials and one limit on extending it beyond them. |
The invitations and the clock#
Use the same words in every condition:
- “Want to have lunch after this?”
- “Would you like to study in the library tomorrow?”
The timer starts at the end of the asker’s last spoken word. Without a token, stop at the first sound of “yes”. With a token, press lap at the first sound of the token and stop at the first sound of “yes”. Write both elapsed times. The first is response onset; the second is yes onset. Keep the screen away from the asker until their rating is written. No audio or video recording.
Manual starts and stops are rough measurements, especially for short gaps. Do not use them to decide whether a reply crossed exactly 700 ms. If the timer misses a boundary, mark the time missing and repeat that trial once.
For the edge-of-room data alternative, use these fictional trials and calculate the same four means: straight away, ratings 4, 5; count one, 3, 4; count three, 2, 3; count three with a token, 4, 3. Their fictional yes-onset times in seconds are 0.2, 0.3; 1.1, 1.0; 3.2, 3.0; 3.8, 3.9; token onsets in the last pair are 0.2, 0.3. State that you analysed prepared data instead of rating live replies. Do not pool it with observed results.
Things this test gets wrong on purpose#
Everyone knows the yes is coming, the invitations aren’t real, the conditions are repeated, and a spoken count is not a calibrated delay. Voice and wording can change too. This is a classroom test of perceived willingness, related to the question Roberts and colleagues asked. It does not replicate their controlled telephone stimuli or test Kendrick and Torreira’s distribution of rejections and acceptances. There are no rejections in our response set. The token condition may also lengthen the time to “yes”; use the recorded times to describe that difference rather than claiming the token alone caused a rating change.
What leaves the room#
A results table showing your ratings across the delay and token conditions. If the ratings did not change, record that too.