Reaction Time, Capacity, Encoding.
[log in to unmask]
Robert F. Potter
[log in to unmask]
Department of Telecommunications
Bloomington, IN 47405
FAX (812) 855-3125
March 29, 1996
Paper submitted to the Theory and Methodology Division of the
Association for Education in Journalism and Mass Communication
Reaction Time, Capacity, Encoding.
This paper investigates the possibility that arousing television messages
more or are allocated more capacity than calm messages and that this increased
encoding results in better memory for arousing messages than for calm messages.
two of three experiments shows that secondary task reaction times are slower
during viewing of
arousing messages compared to calm messages and all three experiments show
better memory for
arousing messages than for calm messages. Arousing
Reaction Time, Capacity, Encoding.
Research on the effects of emotion on memory is plentiful. In media
studies, inquiry has
focused on emotional political ads (Newhagen & Reeves, 1991; A. Lang, 1991),
stories (Newhagen & Reeves, 1992; Brosius, 1993), emotional advertising (Thorson
1989), and emotional public service announcements (A. Lang & Sumner, 1990). By
this research shows that emotional messages are remembered better than neutral
ones (A. Lang,
1989; A. Lang and Friestad, 1993; Lang and Sumner, 1990; Reeves, A. Lang,
Rothschild, 1989; Reeves, Newhagen, Maibach, Basil, & Kurz, 1991).
A subset of this research compares different types of emotional messages to
(rather than comparing emotional messages to neutral messages). Much of this
demonstrates that people have better memory for negative material than for
(Bohannon, 1988; Christianson & Loftus, 1991; Lang, 1991; Newhagen & Reeves,
et al., 1991; Thorson & Friestad, 1989). Others provide support for the so
Principle" by showing increased memory for pleasant stimuli (A. Lang, Dhillon, &
Bradley, Greenwald, & Hamm, 1993; Matlin & Stang, 1978).
Dimensional theories of emotion (P. Lang, 1979) suggest that emotions can
categorized on the basis of two primary dimensions, valence and arousal
Greenwald, Cook, & P. Lang, 1989; P. Lang, 1979; P. Lang, Bradley, & Cuthbert,
1990; P. Lang,
Bradley, Cuthbert, in press). Work using this approach suggests that emotional
remembered better than neutral messages because they are more arousing
(Bradley, 1994). Lang,
Dhillon, & Dong (1995) suggest that much of the research comparing
messages is difficult to interpret since the level of arousal between the
positive and negative
messages was neither controlled or measured.
Work done by Margaret Bradley and her colleagues (Bradley, 1994) on memory
emotional slides consistently shows that arousing messages are remembered better
messages Actually each slide is rated on three dimensions of emotion, valence,
arousal and dominance.
However, the dominance variable accounts for less than 10% of the variance in
most analyses and
does not appear to be related to memory. and that improved memory for emotional
stimuli does not depend upon the valence of
the message, but rather on the arousal associated with that valence. For
example, Bradley et al.
(1992) found that both highly pleasant and unpleasant IAPS slides receive high
ratings from college-aged subjects and that the highly arousing messages were
recalled better both
in immediate recall and in delayed recall measures, which were taken one year
after the slides
were presented. Another study (Bradley et al, 1992) shows that highly arousing
recognized more accurately when subjects were given a forced-choice recognition
Secondary task reaction times (STRT's) obtained during the recognition task were
highly arousing new slides (foils) than for the new slides that were calm.
Bradley (1992) suggested that this last somewhat unexpected finding might
arousing stimuli are more difficult to encode than calm stimuli. Another
possibility is that more
capacity is being allocated to encoding arousing stimuli. If this is the case,
then part of the reason
that arousing slides are remembered better may be that they are encoded better.
Do these findings give any clues as to the role arousal plays in memory for
messages? If we apply the limited capacity model of television viewing (A.
Lang, 1992; A. Lang,
Geiger, Strickwerda, &Sumner, 1993; A. Lang, 1995), this secondary task reaction
time data is a
good place to begin.
Within most limited capacity theories, secondary task reaction times are
used to index the
amount of capacity subjects use to perform some task. Recently, Lang & Basil
suggested that secondary task reaction times, at least in the context of
television viewing, can best
be viewed as an index of the amount of cognitive capacity available at encoding
(A. Lang, &
Basil, 1996). Within this theoretical perspective, the slower secondary task
demonstrated in response to slides (Bradley et al, 1992) would indicate that the
slide viewers had
less capacity available (or not in use) at encoding when they were viewing
arousing slides than
when they were viewing calm slides.
Thus, the first hypothesis in this study asks if this finding, that
secondary task reaction
times are slower for arousing stimuli than for calm stimuli, can be demonstrated
for arousing and
calm television messages.
H1: Television viewers will have faster secondary task reaction times
messages than they will during arousing messages.
This hypothesis flies in the face of much of the literature on secondary
task reaction times
during television viewing. Several studies (Reeves, Thorson, & Schleuder, 1986;
Reeves, Schleuder, 1988) have shown that secondary task reaction times are often
highly complex television messages. It has been speculated that this effect
might be caused by an
increase in arousal felt by subjects in response to complex or fast paced
television messages. It is
then suggested that this increased arousal speeds reaction times during the
Attempting to explain the seeming paradox, that complex messages somehow
capacity than simple ones, Lang & Basil (1996) have suggested that secondary
task reaction times
should not be viewed as a function of either task difficulty or the amount of
capacity allocated to
a task but rather as an indicator of capacity available at encoding. They
further suggest that
many production features of television (like cuts, edits, and video graphics)
the allocation of capacity to encoding the message (Geiger & Reeves, 1993; Lang
et al., 1993).
However, the content of the message determines whether or not that capacity is
encode the message. If the capacity allocated to encoding as a result of
orienting to structural
features is not required to encode the content of the message, then there will
capacity available at encoding, which would result in faster reaction times to
messages where the production techniques add little in the way of new
These suggestions are particularly relevant to understanding the
significance of findings
(Bradley, 1992) that secondary task reaction times are slower during viewing of
than during viewing of calm stimuli. The Bradley et al. Experiments derive, one
must recall, from
a paradigm where the stimuli are static. Subjects in these experiments
(Bradley, 1992) are
viewing slides, so the capacity required by the structure of the medium is the
same for all slides
and only the capacity required by the content varies across slides. This
suggests that the number
of structural features in a television message might need to be controlled to
Capacity, Encoding, and Memory
If more capacity is allocated to encoding arousing material than to
encoding calm material
this would begin to explicate the mechanisms which result in arousing material
better than calm material as was discussed above. If this is the case than,
memory should be
better for messages with slower reaction times - predicted to be the arousing
messages, than it is
for messages with faster reaction times - predicted to be the calm messages.
Hence, hypothesis 2:
H2: Arousing messages will be remembered better than calm messages.
This paper tests these hypotheses with a secondary analysis of data
collected previously in
three different experiments where arousingness of television content was one of
variables and secondary task reaction times and memory were dependent variable.
During this experiment participants viewed 12 two- minute emotional
which had been chosen in a pretest on the basis of valence and arousal (Lang,
Dhillon, & Dong,
1995). Pretest subjects had rated each of the 12 messages using SAM the
Mannequin (P. Lang & Greenwald, 1985).The SAM arousal scale is a pictorial
device which translates into a 9 point scale anchored by the adjectives calm,
sleepy, and restful on
one end and excited, aroused, and energized on the other. For the analysis
reported in this paper,
the 12 messages were ordered by their pretest scores on arousal. Messages which
one point of the midpoint were dropped to ensure that the arousing messages were
the calm messages were calm. As a result, only 8 of the original 12 messages
were included in the
analysis. The four most arousing messages had an average SAM arousal score of
3.03 (where 1=
highly aroused and 9=calm) while the four calmest messages had an average SAM
Dependent Variables - Secondary Task Reaction Time
Secondary task reaction times were measured by having subjects press a
button held in the
dominant hand as fast as they could every time they heard a tone embedded in the
audio track of
the television message. Five reaction time tones were placed in each message.
The first reaction
time occurred at the start of each message. The next three reaction time tones
randomly such that one occurred within each 40 second period of the message.
The final reaction
time tone was placed in the last ten seconds of the message.
The experiment was controlled by a Zenith 286 computer equipped with a
AD/DA board. Digital signals placed on audio track 2 of the videotapes both
emitted the reaction
time signal and signaled the start of the reaction time period. Each reaction
time button sent a
digital signal to the computer. The computer measured the milliseconds from
tone onset to
button push. The accuracy is plus or minus one millisecond.
Memory was measured by giving subjects a packet of 12 randomly arranged
the top of each page was a cue which identified one of the 12 messages.
Subjects were asked to
write everything they could remember about the message on that page. Six
measures were coded
from these protocols by two coders working independently. The six measures were
words, number of sentences, audio sentences, video sentences, global ideas, and
Intercoder reliabilities ranged from .86-.99. These six measures were then
combined into a global
memory scale (alpha=.991). This global memory scale is used in the analysis
Subjects were 48 undergraduate advertising students, 31 males and 17
subjects received extra credit for their participation.
Two experimenters (one male and one female) tested subjects over a four-day
Subjects were randomly assigned to one of three stimulus tape orders at the
start of the viewing
session. Subjects were given a reaction time button and instructed in its use.
Subjects were told
to pay close attention to the television messages as they would be tested later
to ensure that
watching television was the primary task. Subjects were then shown a three
containing 8 practice reaction time tones. Following this practice session they
had a chance to ask
questions. Then they viewed the stimulus tape and filled out the memory
The design for this analysis is a mixed 3 (Order of Presentation) X 2
(Content Arousal) X
4 (Message) repeated measures ANOVA. The only between subjects factor is Order
Presentation, which corresponds to the three orders of the stimulus tape.
Content Arousal is a
within subjects variable with two levels, arousing and calm. Message is a
within subjects factor
corresponding to the four messages within each level of arousal.
Results - Experiment 1
This hypothesis predicted that when watching arousing messages viewers
slower reaction times than when watching calm messages. The main effect for
Arousal on the
reaction time data was not significant (F(1,44)=1.65, p<.20), though the means
were in the
predicted direction with the mean RT for arousing message (M=709.49) being
slightly slower than
the mean RT for calm messages (M=706.41).
This hypothesis predicted that arousing messages would be remembered better
messages. Again this hypothesis is tested by the significant main effect for
Arousal (F(1,38) =
153.88, p<.000, epsilon-squared=.7967). The mean total memory score for
was just over twice as large (M=65.79) as the mean for calm messages (M=32.23).
This experiment was originally designed as a replication and extension of
(Lang & Bolls, 1995). Again subjects viewed 12 emotional messages (presented in
orders of presentation) chosen to vary on arousal and valence. Eleven of the 12
the same messages used in Experiment 1. All of the messages were chosen for
inclusion in the
stimulus tape on the basis of pre-test SAM scores. The subjects in this
experiment, however, also
rated how they felt while watching each of the messages using SAM. For this
messages were ranked, not on the pre-test SAM scores, but on the subjects' own
ratings of how
they felt while viewing each message. The four most arousing messages and the
messages were chosen. This resulted in six of the eight messages being the same
those in experiment 1, while two were different. Analysis of subjects' SAM
arousal scores shows
that the four arousing messages (M=4.77) were significantly more arousing than
the four calm
messages (M=7.19) (F(1,65)=167.80, p<.000, epsilon-squared=.7165).
Dependent Variables - STRT and Recognition
Secondary reaction time was measured using the same procedures, equipment,
apparatus as in experiment 1. Again, five reaction time signals were inserted
in each message.
Memory was measured using recognition (rather than cued recall techniques).
given a test consisting of six four-alternative multiple choice questions for
each message. These
questions were constructed to ask for specific information that was relevant to
Questions did not ask about the gist of story. Questions were designed to be
difficult but not to
test memory for irrelevant information.
Sixty seven undergraduate communication students participated in the
experiment. All 64
students completed the recognition measures. Thirty two of the students were
also in the reaction
time condition. The other 32 students were in a physiology condition not
Procedures were the same as for experiment 1 except that the stimulus was
each message to allow the subjects to rate how they felt during the message
using the Self-Assessment Mannequin.
The design was a mixed 2 (Order of Presentation) X 2 (Arousal) X 4
measures ANOVA. Again Order of Presentation is a between subjects variable with
corresponding to the two orders of presentation of the messages. Arousal is a
variable with two levels corresponding to how aroused the subjects reported
feeling while viewing
the messages. Message is also a within subjects variable representing the four
messages in each
Results - Experiment II
This hypothesis predicted that reaction times would be slower for arousing
for calm messages. The main effect for arousal on the reaction time data was
(F(1,29)=19.50, p<.000, epsilon-squared=.3814). Reaction times were
significantly slower during
arousing messages (M=771.87) than during calm messages (M=733.66).
This hypothesis predicted that arousing messages would be remembered better
messages. Again the main effect for Arousal was significant (F(1,64)=46.45,
p<.000, epsilon-squared = .4126). As predicted memory was better for arousing
messages (M=.76) than for calm
Subjects in this experiment viewed thirty 30-second television messages
(Lang, Potter, &
Bolls, 1996; Lang, Bolls, & Kawahara, 1996; Kawahara, Bolls, Hansell, & Lang,
thirty messages were chosen from a pool of over 300 messages on the basis of
arousal ratings and
production pacing. Arousal was determined for each message by three coders
independently using a 9 point scale where 1=very calm and 9=very aroused.
Messages chosen for
this study had to be rated 1 or 2 by all three coders to be included as calm
messages or 8 or 9 by
all three coders to be included as arousing messages. Subjects in the
experiment then rated their
emotional responses to the messages using the SAM arousal scale. Subjects rated
feeling significantly more aroused during the arousing messages (M=5.6) than
during the calm
messages (M=7.22) (F(1,49)=196.66, p<.000, epsilon-squared = .7965). Four
orders of the messages were prepared.
Dependent Variables - STRT and Cued Recall
Again secondary task reaction times were measured in the same way that they
experiments 1 and 2. Two secondary task reaction time signals were embedded in
one placed randomly in each 15-second period of the 30-second messages, with the
that no signal could be within 2 seconds of a cut or edit. Memory was measured
using the same
protocol as in experiment 1. Subjects were given a packet of thirty sheets
arranged in one of four
random orders. Each sheet had a cue to one message. Subjects were asked to
everything they could remember about the message. To date, these messages have
coded for number of words written.
Subjects were 51 undergraduate communication majors participating for extra
The procedures were similar to those used in experiment 1 and 2. Subjects
instructed in the use of the reaction time button and shown a practice video.
During the viewing
of the stimulus tape the tape was paused after each message. Subjects had 30
messages to rate how they felt while viewing using the Self-Assessment
were given a warning that the message would be starting 5 seconds prior to
Following message viewing, subjects completed a questionnaire for another
experiment and then
were given the cued recall test.
The design of this experiment is a mixed 3 (Order of Presentation) X 2
X 15 (messages) repeated measures ANOVA. Order of Presentation is a between
variable. Content arousal is a within subjects variable with two levels
Arousing and Calm.
Messages has 15 levels representing the 15 messages in each level of content
Results - Experiment III
This hypothesis predicted a main effect for Arousal on the reaction time
data. This effect
was significant (F(1,48)=15.36, p<.000, epsilon-squared = .2266). Reaction
times were slower
during arousing messages (M=845.71) than they were during calm messages
This hypothesis predicted that memory would be better for arousing messages
calm messages. The main effect for Arousal on the memory data was significant
p<.000). Subjects wrote more words about arousing messages (M=12.62) than they
calm messages (M=10.95).
Results from these studies suggest that secondary task reaction times may
in fact be
slower while people are watching arousing television messages than when they are
television messages. Further, this result appears to be greatest when the
determination of what
messages are calm and what are arousing are based on subjects' own ratings of
how aroused they
felt. In experiment 1, the only case where the effect of arousal was not
significant (though the
means were in the correct direction, the determination of which messages were
which messages were not was based on pre-test subjects' judgements of the
In addition to resulting in slower secondary task reaction times, arousing
also remembered much better than calm messages in all three conditions. Both
cued recall measures were used in these three studies and both types of memory
supported this hypothesis.
These results support either of two theoretical possibilities: (1) that
require more capacity to encode, or (2) that arousing messages are allocated
more capacity to
encode. This increase in the capacity given over to the task of encoding the
message seems to be
at least part of the reason why arousing television messages are remembered
better than calm
ones. This raises an interesting theoretical question. Is this increase in
capacity automatic -- that
is, outside of the control of the viewer, or is it controlled -- that is, under
the control of the
Many people would argue that the subject matter of the arousing messages
blood, and gore) is more interesting than the subject matter of the calm
nature, conversation) and thus that people voluntarily allocate more capacity to
arousing messages because of the higher interest level. If this is the case,
than the slower
secondary task reaction times might be caused by a voluntary allocation of
capacity to the more
interesting arousing messages.
On the other hand, it is also possible that the arousing contents result in
an imperative call
for capacity and that we automatically allocate additional capacity to arousing
messages. There is
a logical argument here, too, that suggests that arousing messages tend to be
biologically primal topics like danger, food, procreation, and survival, and
that there would be an
advantage to the species if these things called attention to themselves. It is
also possible that both
an automatic and a controlled allocation of capacity combine to result in
increased encoding and
therefore memory for arousing messages.
It seems likely, however, that at least some of the increased encoding
capacity is due to
automatic allocation. Two points support this contention. First, in the design
experiments a concerted effort was made to control viewers' voluntary attention
were told to pay close attention to the television messages as they would be
tested. The practice
test they were given after viewing the reaction time practice tape was extremely
difficult than the tests given for the stimulus messages), in order to stress
the need to watch
closely. Second, in the work on slides by P. Lang and Bradley (1992) reported
times during viewing of arousing and calm slides was measured both to new slides
subjects had not viewed before) and to old slides (slides the subjects had
Reaction times were slower for arousing slides only when subjects were viewing
the slides for the
first time. When viewing slides for the second time (which means they have
encoded), reaction times to arousing slides were faster! If the change in
reaction time were due to
the interestingness of the slides - subjects should still show slower reaction
times on the second
viewing of the slides. But they don't.
Future research should continue to investigate this question of the
automaticity of the
allocation of capacity to encoding. It would also be interesting to investigate
the effects of
valence (controlling for arousal) in television messages on reaction times and
memory both in
comparison to one another and in comparison to neutral messages. In addition,
the question of
how production pacing interacts with arousal to alter reaction times may further
theoretical question of exactly what is being measured by secondary task
reaction times. References
Bohannon, J. (1988). Flashbulb memories for the space shuttle disaster: A
tale of two
theories. Cognition, 29, 179-196.
Bolls, P., Potter, R.F., Lang, A. (1996). The effects of production pacing and
on encoding, storage, and retrieval of television messages. Paper to be
presented to the
Midwest Artificial Intelligence and Cognitive Science Conference.
Bradley, M. M., Greenwald, M. K., Petry, M. C., & Lang, P. J. (1992).
pictures: Pleasure and arousal in memory. Journal of Experimental
Bradley, M. M. (1994). Emotional memory: A dimensional Analysis. In S. Van
Goozen, N.E. Van de Poll, & J. A. Sergeant (Eds.), Emotions: essays on
(pp. 97-134). Hillsdale, NJ: Lawrence Erlbaum, Associates.
Bradley, M., Greenwald, M., & Hamm, A. (1993). Affective picture processing.
Birbaumer & A. Ohman (Eds.), The structure of emotion:
cognitive, and clinical aspects. Toronto: Hogrefe & Huber.
Brosius, H-B. (1993). The effects of emotional pictures in television news.
Communication Research, 20 (1), 105-124.
Christianson, S., & Loftus, E. (1991). Remembering emotional events: The
detailed information. Cognition and Emotion, 5, 81-108.
Friestad, M. & Thorson, E. (1993). Remembering Ads: The effects of encoding
strategies, retrieval cues, and emotional response. Journal of consumer
Geiger, S. & Reeves, B. (1993). The effects of scene changes and semantic
attention to television. Communication Research, 20, 155-175.
Greenwald, M., Cook, E., & Lang, P. (1989). Affective judgment and
response: Dimensional covariation in the evaluation of pictorial stimuli.
psychophysiology, 3, 51-64.
Kawahara, K., Bolls, P., Hansell, R., & Lang, A. (1996). The effects of
and content arousal on viewers allocation of capacity to encoding and
storage of television
messages. Paper to be presented to the International Communication
Information Systems Division, Chicago, IL. May.
Lang, A. & Bolls, P. (1995). Memory for emotional television messages:
Arousal, valence, and
capacity. Presented to the Information Systems Division of the
Communication Association. Albuquerque, NM, May.
Lang, A., Bolls, P., & Kawahara, K. (1996). The effects of arousing message
structural complexity on television viewers' level of arousal and
allocation of processing
resources. Paper to be presented to the Midwest Artificial Intelligence
Lang, A. (1995). Defining audio/video redundancy from a limited capacity
processing perspective. Communication Research, 22, 86-115.
Lang, A. (1989). Effects of structural features and emotional impact on
televised political commercials. Presented to the American Academy of
Lang, A. (1991). Emotion, formal features, and memory for televised political
In F. Biocca (Ed.) Television and political advertising (vol. I):
pp. 221-244. Hillsdale, NJ: Lawrence Erlbaum, Associates.
Lang, A. (1992). A limited capacity theory of television viewing. Paper
presented to the
Information Systems Division of the International Communication
Lang, A., & Basil, M. (1996). What do secondary task reaction times measure
Accepted to the Information Systems Division, International Communication
Association, Chicago: IL.
Lang, A. & Bolls, P. (1995). Memory for emotional television message. Top four
presented to the Information Systems Division of the International
Association. Albuquerque, NM.
Lang, A., Dhillon, K., & Dong, Q. (1995). The effects of emotional arousal and
on television viewers' cognitive capacity and memory. Journal of
Electronic Media, 39, 313-327.
Lang, A., and Sumner, J. (1990). Emotion, arousal, and memory for public
announcements: Murky but interesting? Presented to the International
Association, Dublin, Ireland.
Lang A., & Friestad, M. (1993). Emotion, hemispheric specialization, and
verbal memory for television messages. Communication Research, 20,
Lang, A., Geiger, S., Strickwerda M., & Sumner, J. (1993). The effects of
unrelated cuts on viewers' memory for television: A limited capacity
theory of television
viewing. Communication Research, 20, 4-29.
Lang, P. (1979). A bio-informational theory of emotional imagery.
Lang, P, Ohman, A., &Vaid, D. (1988). The international affective picture
[photographic slides]. Gainesville, FL: The Center for Research in
University of Florida.
Lang, P., Bradley, M., & Cuthbert, B. (1990). Emotion, attention, and the
reflex. Psychological Review, 97, 377-398.
Lang, P., Bradley, M., & Cuthbert, M. (in press). Motivated attention:
and action. In P. Lang, R. Simons, & M. Balaban (Eds.) Attention and
Sensory and Motivational Processes. Hillsdale, NJ: Lawrence Erlbaum
Lang, P.J. & Greenwald, M. (1985). The international affective picture system
technical report. Gainesville: University of Florida, Center for Research
Matlin, M., & Stang, D. (1978). The pollyana principle: Selectivity in
and thought. Cambridge, MA: Schenkman.
Newhagen, J. & Reeves, B. (1991). Emotion and memory responses for negative
political advertising: A study of television commercials. In F. Biocca
and political advertising (vol. I): Psychological processes. pp. 197-220.
Lawrence Erlbaum, Associates.
Newhagen, J., & Reeves, B. (1992). The evening's bad news: Effects of
television news images on memory. Journal of Communication, 42 (2),
Reeves, B., Lang, A., Thorson, E., & Rothschild, M. (1989). Emotional
and hemispheric specialization. Human Communication Research, 15, 493-508.
Reeves, B., Newhagen, J., Maibach, E., Basil, M., & Kurz, K. (1991). Negative
television messages: Effects of message type and message context on
memory. Animal Behavioral Scientist, 34, 679-694.
Reeves, B., Thorson, E. & Schleuder, J. (1986). Attention to television:
theories and chronometric measures. In Bryant & Zillmann (Eds.),
Media Effects. Hillsdale, NJ: Lawrence Erlbaum, Inc.
Thorson, E. & Friestad (1989). The effects of emotion on episodic memory for
commercials. In P. Cafferatta & A. Tybout (Eds.) Cognitive and affective
advertising. pp. 305-326. Lexington, MA: Lexington Books.
Thorson, E., Reeves, B. & Schleuder, J. (1986). Attention to local and global
television messages. In M. L. McLaughlin (Ed.), Communication Yearbook 10.
Hills, CA: Sage.