Apparatus. The norms were collected using Macintosh Quadras 650. Participants wore Shure SM10A microphone headsets which were connected to PsyScope button boxes that acted as voice keys and provided millisecond timing accuracy (Cohen, MacWhinney, Flatt, & Provost, 1993). In addition, responses were recorded onto audiotape with Radio Shack lapel microphones. Pictures were displayed on 17 in. Macintosh monitors.
Materials. Three hundred seventy-six pictures of objects from a
wide array of sources were included. The published sources included Ferreira
and Cutting (1997) coded as F&C; Roach, Schwartz, Martin, Grewal, and
Brecher (1996) coded as Philadelphia; and Snodgrass and Vanderwart (1980)
coded as S&V. However, the majority came from pictures originally printed
in dictionaries and other sources and digitized by Huitema (1996). These
have not been published but are available upon request. Pictures were formatted
to a size of 168 x 128 pixels at 72 x 72 dpi.
Procedure. Individuals were tested in small rooms, one at a time. The sensitivity
of the voice key was adjusted before the subjects received the following
instructions read aloud by an experimenter while presented on a computer
monitor:
In this experiment you are going to name aloud some pictures.Name each picture as quickly as you can with the first appropriate word that comes to mind. After you name a picture, you need to indicate whether the computer properly registered your response. If the picture disappeared BEFORE you spoke or noticeably AFTER, press the ERROR button on the keyboard. Also press this button if you stuttered or said "Um." Sometimes you may need to repeat a word a few times before the computer registers the response and takes away the picture. This is also considered an error and if it happens a few times in a row, get up and tell the experimenter. Only press the GOOD button if the picture disappeared as soon as you began speaking.
Remember to name the pictures as fast as you can. You may take as much time as you like to press the GOOD and ERROR buttons. There will be ten practice trials to get you used to this procedure.
An experimenter was present for the 10 practice pictures that followed and
answered any questions that arose. The practice pictures were randomly selected
from the full set of pictures but were not recorded or included in analyses.
The rest of the session was unsupervised. Participants named 366 pictures
of objects presented in a different random order for each.
Coding of responses. In addition to using participantsí coding of their responses, a transcriber listened to the audiotapes of each session, entering responses to each picture and noting any signs of voice key problems. Percentage of target and non-target responses is based on first response produced in case of multiple responses.
Calculation of mean response latencies. For each picture, two mean response latencies were calculated. The first latency includes all naming responses; the second latency includes only trials on which the target name was produced. Typically, target names are the ones given most often by participants.
Recordings for 2 experimental trials were unavailable so these were excluded from calculations as well as practice trials. In addition, we excluded trials on which participants indicated that voice key problems occurred and those resulting in response times less than 200 ms and longer than 3 s, eliminating 9.7% of trials. Mean general response latencies, including both target and non-target responses, were calculated from the remaining 14782 trials. To calculate mean latencies for target responses, 30% of the remaining trials on which non-target responses were produced were excluded, leaving 10,348 trials. For a small number of pictures, an additional mean latency was calculated for a name other than the target one.
Data format. The data are available in a tab-delimited text file, with 23 columns and 377 rows including column headings. Column headings are listed below in square brackets with their contents described.
[PicFileName] - An 8 character file name for identifying pictures.
[Source] - The picture collection the picture came from.
[GenMeanRT] [GenStd. Dev.] [GenStd. Error] [GenCount] - Mean RT based on all names, the standard deviation and error of the mean plus the number of observations it is based on.
[TargetName] - Typically the dominant name given to a pictured object.
[TargRTMean] [TargRTSt.Dev.] [TargRTSt.Error] [TargRTCount] - Mean RT based only on target names, the standard deviation and error of the mean plus the number of observations it is based on.
[Target%] [NonTarget%] [Pass%] [Total] - These give the percentage of participants giving each type of response. Note that target percentage will correspond to a number greater than the TargRTCount whenever latencies for target responses were associated with voice key problems or were trimmed. Pass% indicates the percentage of non-naming responses, including silence, "I don't know" and "Pass."
[2nd most] [3rd most] - These are the second and third most common names aside from the target name.
[AltName] - An alternative to the target name for which mean RTs were calculated.
[AltMeanRT] [AltRTStd.Dev] [AltRTStd.Error] [AltRTCount] - Mean RT based only on trials eliciting alternative names, the standard deviation and error of the mean plus the number of observations it is based on.
[NormsName] - The name used to identify the picture in the norming study.
References
Acknowledgements & Author Note
Please cite this as:
Griffin, Z. M., & Huitema, J. (1999). Beckman Spoken Picture Naming
Norms. [On-line]. Available: http://langprod.cogsci.uiuc.edu/~norms/
(or however APA style citations develop)
updated 11/03