WE’RE NOT SAYING!
A Survey of Subway Car Announcements
METHODOLOGY
We’re Not Saying! is intended to serve as a follow-up to the May 1998 Straphangers Campaign survey entitled Say What? The methodology followed in We’re Not Saying! is similar to the methodology of Say What?, with minor exceptions as noted below.
The survey was conducted by 96 volunteers during the period from December 4, 1998 to July 14, 1999. Volunteers were instructed to complete a survey entry every time they used the subways throughout that period. The survey form was designed to measure how each subway trip conformed to the guidelines laid out for announcements by MTA New York City Transit. Surveyors noted the date and line of each trip, and noted whether the name of the upcoming station and any applicable transfer information was understandably announced prior to entering:
In addition, surveyors recorded whether there was a change in service or a delay between stations of two minutes or more. When such a change in service or delay occurred, surveyors noted whether an in-car announcement was made, whether any announcement made was audible and ungarbled, and the exact language of the announcement.
- the first stop of the trip;
- the second stop of the trip;
- the final stop of the trip; and
- the first transfer point of the trip.
Station name and transfer information announcements were grouped and totaled as one category. Both types of announcements together comprise the “basic” in-car announcement, determined only by characteristics of the line and upcoming station. A basic announcement was deemed adequate if the upcoming station name and any applicable transfer information were announced in an audible and ungarbled manner.
In our survey, we recorded 250 basic announcement opportunity observations on each of 18 lines. The J/Z and M lines as well as the Times Square and Franklin Avenue Shuttles, each with fewer than 250 survey entries, were dropped from consideration in the report.
Our findings then reflect exactly 4500 basic announcement opportunity observations; of these, our surveyors noted that for the entire system, 47% of announcements (plus or minus 2%) were adequate. The by-line confidence interval in each case is plus or minus 7%. All results are significant at the .05 level.
By-line basic announcement levels measured in this report are comparable to those reported in our 1997-98 survey. Of the seventeen lines we measured in 1997-98 for basic in-car announcements, nine—the #4, #5, #6, A, B, C, E, F and L lines—showed a statistically significant deterioration in performance between the survey periods.
One line—the N—exhibited a significant improvement. Differences in announcement performance on the #1/9, #2, #3, #7, D, G and R lines were not statistically significant. Our 1998-99 survey includes a measurement of basic announcements on one line—the Q—which was not included in the previous survey.
In order to appropriately compare system-wide basic announcement performance levels between 1997-98 and 1998-99, Q line data must be omitted from 1998-99 results. Thus for purposes of comparison, out of 4200 basic announcement opportunities on 17 equally-represented lines, our survey shows that in the 1998-99 period, 52% of announcements were inadequate (plus or minus 2%). In our 1997-98 report, we noted that 46% of basic announcements were inadequate. This decline in systemwide performance is significant at the .05 level.
Surveyors were also instructed to record whether there was a change in service or a delay between stations of two minutes or more. When such a change occurred, surveyors noted whether an in-car announcement was made, whether any announcement made was audible and ungarbled, and the exact language of the explanation given. Straphangers Campaign staff members then examined the transcription of each announcement to determine whether the announcement was “useful ” or “not useful ” to riders.
EXAMPLES OF "USEFUL" ANNOUNCEMENTS EXAMPLES OF "NOT USEFUL" ANNOUNCEMENTS fire/accident, police activity in service (no further explanation) ahead of/behind schedule, congestion schedule adjustment sick passenger fix gap in service emergency brake pulled held by supervision/dispatcher track/signal work red signal
Our surveyors observed 135 delays and service changes during the survey period, of which 36 were followed by an ungarbled, audible and useful explanation for the disruption in service (27%, plus or minus 8%). These findings are consistent with those published in our 1998 report, in which we recorded 70 ungarbled, audible and useful announcements in response to 213 measured delays and service changes (33%). The difference in results between the two survey periods is not statistically significant at the .05 level.Delay and service change data was not examined on a line-by-line basis. The number of such changes would be too low to permit statistically significant comparisons among lines.
Straphangers Campaign Home Page | Announcement Survey Summary | Announcement Survey Charts and Tables