INT'L. J. AGING AND HUMAN DEVELOPMENT, Vol. 9(3), 1978-79
TRAINING RESPONSE SPEED IN YOUNG AND ELDERLY WOMEN
FRANCES W. HOYER Hutchings Psychiatric Center WILLIAM J. HOYER Syracuse University NANCY J. TREAT Pennsylvania State University Fayette Campus PAUL B. BALTES Pennsylvania State University
ABSTRACT
Effectiveness of response speed training on the performance of thirty adult women was assessed. Five young and five elderly female volunteers were tested in each of three experimental conditions: (1) one training session with cognitive feedback in which participants were given response rate information (control); (2) five training sessions with cognitive feedback (practice); or (3) five training sessions with cognitive feedback in which the number of S&H green stamp units earned was directly proportional to response rate (conjugate reinforcement). Dependent variables were (a) response speed on three paper-pencil tasks, and (b) posttraining performance on twelve intelligence subtests chosen as far transfer tasks. Response speed increased significantly with training in both age groups, but contrary to expectation, young adults showed greater training effects than elderly adults. No significant far transfer effects were obtained.
Several writers have noted that the study of intelligence in late adulthood and old age is guided by a decline model of aging [ l , 21. This emphasis is characteristic of the aging literature for a variety of psychological performance dimensions, (e.g., learning, perception, speed, and timing). As a corollary, there has been an extreme dearth of experimental-manipulative research aimed 247
0 1979. Baywood Publishing Co.. he.
doi: 10.2190/EF3C-FLPV-8Q7M-KD5Y http://baywood.com
248 I HOYER E T A L .
at demonstrating the potentials and modifiability of intellectual performance and timed behavior in elderly individuals. The notion that performance on measures of intelligence is more variable than traditionally assumed has been documented recently in several studies. For example, intellectual performance has been shown to be modifiable by conditions associated with the testing context [3], and by short-term response speed training [4] . Also, Woodruff studied experimentally the relationship between reaction time and EEG biofeedback, and found that biofeedback can be used to produce faster brain waves and possibly to facilitate reaction time [5]. In the Hoyer, Labouvie, and Baltes experiment, a two session nonreinforcement practice program and a two-session practice with continuous reinforcement program both produced some, improvement in response speed and in intelligence test scores of elderly women [4]. Similarly, the present research is an attempt to demonstrate an improvement in response speed as a function of training. Since there were only slight differences in the effectiveness of the two training programs, it was felt that the use of a more powerful reinforcement schedule might increase reinforcement effects. In previous research it has been found that a conjugate schedule of reinforcement, in which amount of reinforcement is directly proportional to the magnitude of the criterion behavior, is superior to traditional schedules in which only reinforcement frequency is varied [6] . A conjugate reinforcement schedule was employed in the present research. Another purpose of this study was to compare the response speed training effects on elderly persons with training effects on young persons. Whereas Hoyer, Labouvie, and Baltes assessed performance only in women over sixty, in the present investigation both elderly and young women were tested [4]. Presumably, young (college student) adults have extensive speeded test-taking experience in their everyday lives and thus may be closer to peak speeded performance levels. Elderly persons, however, do not currently have (and some have never had) extensive speeded test-taking experience, and they might be exhibiting correctable pretraining response speed deficits. Accordingly, it was predicted that training of response speed would be less effective for young adults than for elderly adults. Hoyer, Labouvie, and Baltes also examined the transfer of response speed training to standard abilities tests [4]. Their expectation that experience with testing materials and reinforcement of speeded performance would significantly improve intellectual performance of the elderly on abilities tests was not confirmed. The authors suggested that additional training sessions might enhance transfer effects. In the present study the number of sessions was increased from two to five with the expectation that transfer of training would be obtained.
METHOD A cross-sectional design was used with thirty white middle class adult women as participants. The fifteen elderly women who ranged in age from
RESPONSE SPEED TRAINING I 249
fifty-seven to eighty-six years (mean = 70.5 years) had an average educational level of 11.5 years. The fifteen young women who ranged in age from twenty to twenty-two years (mean = 21.3 years) were volunteers from undergraduate psychology classes with an average educational level of 14.0 years. Three paper-pencil tasks were used to give participants practice in speeded responding: cancellation (marking and counting target letters), marking (filling in designated blanks on a standard IBM answer sheet), and writing (copying words as rapidly as possible). Response speed was measured by the number of items completed in thirty-second intervals. Participants were randomly assigned within each age group to one of three experimental conditions: control, practice, and conjugate reinforcement. In all three groups, participants received cognitive feedback for response rate at thirty-second intervals; this treatment consisted of telling the participants the number of completed items after each thirty-second interval. Women in the control condition participated in one session with cognitive feedback, while those in the practice condition participated in five such sessions. Participants in the conjugate reinforcement condition participated in one cognitive feedback session and in four conjugate reinforcement sessions. During the conjugate reinforcement sessions these women received, in addition to cognitive feedback, the same number of S&H green stamps per thirty-second interval as items completed for that interval. Sessions for experimental participants were approximately forty-five minutes in length and were distributed over a fourteen-day period. Approximately one week following training, a posttraining session was held in order to assess transfer of response speed training to intellectual performance. All participants completed the SRA Primary Mental Abilities test [7], and six tests (identical pictures, finding “a” ’s, word beginnings, word endings, word beginnings and endings, number comparison) adapted from the French, Ekstrom, and Price [ 8 ] kit of reference tests.
RESULTS Training effects were assessed by comparing mean numbers of items completed in thirty-second intervals. Means for elderly women in the control group on cancellation, marking, and writing, respectively, were 5.17, 7.17, and 12.80, while corresponding means for college students in the control group were 7.77, 16.68, and 18.68. Separate two (age) by five (session) by two (training group : practice vs. conjugate reinforcement) analyses of variance were computed for cancellation, marking, and writing performance. In addition, one two (age) by three (training group: control vs. practice vs. reinforcement) analysis of variance was performed for each of the twelve posttest intellectual transfer measures. The results of these analyses are summarized in Tables 1 and 2. For all three response speed training rash, significant improvement was obtained over sessions for the young and the elderly women. No differences
250 I HOYER E T A L .
Table 1. Summary of Analyses of Variance on Training Tasks ~~
Training Task Source
d.t
Age (A) Treatment (TI Sessions (S) AXT AX S TX S AXTXS
1, 16 1, 16 4, 64 1, 16 4, 64 4, 64 4, 64
ap
Cancellation Marking
Writing
63.02a
104.14a
30.92a
126.79’ 1.44 30.88a 4.21
210.43a 1.84 69.6ga
35.34a
1.61
-
5.68’ 2.81 6.25’
< .005.
Note: All entries are F ratios larger than 1 .OO.
Table 2. Summary of Analyses of Variance on Cognitive Transfer Tasks Source Transfer variable
Verbal meaning Space Reasoning Number Word fluency Z PMA Word beginnings & endings Word beginnings Word endings Z Word fluency Number comparison Identical pictures Finding A s Z perceptual speed Tota I performance
Age
Treatment
9.04a 6.6ga 75.38’ 2.15 7.67’ 17.17’
2.12
3.20a 2.15 2.24 19.26’ .65.74’ 10.07’ 36.08’ 21.81
’
Age X Treatment
1.27 1.93
1.86
1.01
1.56 1.26
1.06
1.22
2.59 2.50 1.15 1.10
1.24 1.31
p < .05.
p
TRAINING RESPONSE SPEED IN YOUNG AND ELDERLY WOMEN
FRANCES W. HOYER Hutchings Psychiatric Center WILLIAM J. HOYER Syracuse University NANCY J. TREAT Pennsylvania State University Fayette Campus PAUL B. BALTES Pennsylvania State University
ABSTRACT
Effectiveness of response speed training on the performance of thirty adult women was assessed. Five young and five elderly female volunteers were tested in each of three experimental conditions: (1) one training session with cognitive feedback in which participants were given response rate information (control); (2) five training sessions with cognitive feedback (practice); or (3) five training sessions with cognitive feedback in which the number of S&H green stamp units earned was directly proportional to response rate (conjugate reinforcement). Dependent variables were (a) response speed on three paper-pencil tasks, and (b) posttraining performance on twelve intelligence subtests chosen as far transfer tasks. Response speed increased significantly with training in both age groups, but contrary to expectation, young adults showed greater training effects than elderly adults. No significant far transfer effects were obtained.
Several writers have noted that the study of intelligence in late adulthood and old age is guided by a decline model of aging [ l , 21. This emphasis is characteristic of the aging literature for a variety of psychological performance dimensions, (e.g., learning, perception, speed, and timing). As a corollary, there has been an extreme dearth of experimental-manipulative research aimed 247
0 1979. Baywood Publishing Co.. he.
doi: 10.2190/EF3C-FLPV-8Q7M-KD5Y http://baywood.com
248 I HOYER E T A L .
at demonstrating the potentials and modifiability of intellectual performance and timed behavior in elderly individuals. The notion that performance on measures of intelligence is more variable than traditionally assumed has been documented recently in several studies. For example, intellectual performance has been shown to be modifiable by conditions associated with the testing context [3], and by short-term response speed training [4] . Also, Woodruff studied experimentally the relationship between reaction time and EEG biofeedback, and found that biofeedback can be used to produce faster brain waves and possibly to facilitate reaction time [5]. In the Hoyer, Labouvie, and Baltes experiment, a two session nonreinforcement practice program and a two-session practice with continuous reinforcement program both produced some, improvement in response speed and in intelligence test scores of elderly women [4]. Similarly, the present research is an attempt to demonstrate an improvement in response speed as a function of training. Since there were only slight differences in the effectiveness of the two training programs, it was felt that the use of a more powerful reinforcement schedule might increase reinforcement effects. In previous research it has been found that a conjugate schedule of reinforcement, in which amount of reinforcement is directly proportional to the magnitude of the criterion behavior, is superior to traditional schedules in which only reinforcement frequency is varied [6] . A conjugate reinforcement schedule was employed in the present research. Another purpose of this study was to compare the response speed training effects on elderly persons with training effects on young persons. Whereas Hoyer, Labouvie, and Baltes assessed performance only in women over sixty, in the present investigation both elderly and young women were tested [4]. Presumably, young (college student) adults have extensive speeded test-taking experience in their everyday lives and thus may be closer to peak speeded performance levels. Elderly persons, however, do not currently have (and some have never had) extensive speeded test-taking experience, and they might be exhibiting correctable pretraining response speed deficits. Accordingly, it was predicted that training of response speed would be less effective for young adults than for elderly adults. Hoyer, Labouvie, and Baltes also examined the transfer of response speed training to standard abilities tests [4]. Their expectation that experience with testing materials and reinforcement of speeded performance would significantly improve intellectual performance of the elderly on abilities tests was not confirmed. The authors suggested that additional training sessions might enhance transfer effects. In the present study the number of sessions was increased from two to five with the expectation that transfer of training would be obtained.
METHOD A cross-sectional design was used with thirty white middle class adult women as participants. The fifteen elderly women who ranged in age from
RESPONSE SPEED TRAINING I 249
fifty-seven to eighty-six years (mean = 70.5 years) had an average educational level of 11.5 years. The fifteen young women who ranged in age from twenty to twenty-two years (mean = 21.3 years) were volunteers from undergraduate psychology classes with an average educational level of 14.0 years. Three paper-pencil tasks were used to give participants practice in speeded responding: cancellation (marking and counting target letters), marking (filling in designated blanks on a standard IBM answer sheet), and writing (copying words as rapidly as possible). Response speed was measured by the number of items completed in thirty-second intervals. Participants were randomly assigned within each age group to one of three experimental conditions: control, practice, and conjugate reinforcement. In all three groups, participants received cognitive feedback for response rate at thirty-second intervals; this treatment consisted of telling the participants the number of completed items after each thirty-second interval. Women in the control condition participated in one session with cognitive feedback, while those in the practice condition participated in five such sessions. Participants in the conjugate reinforcement condition participated in one cognitive feedback session and in four conjugate reinforcement sessions. During the conjugate reinforcement sessions these women received, in addition to cognitive feedback, the same number of S&H green stamps per thirty-second interval as items completed for that interval. Sessions for experimental participants were approximately forty-five minutes in length and were distributed over a fourteen-day period. Approximately one week following training, a posttraining session was held in order to assess transfer of response speed training to intellectual performance. All participants completed the SRA Primary Mental Abilities test [7], and six tests (identical pictures, finding “a” ’s, word beginnings, word endings, word beginnings and endings, number comparison) adapted from the French, Ekstrom, and Price [ 8 ] kit of reference tests.
RESULTS Training effects were assessed by comparing mean numbers of items completed in thirty-second intervals. Means for elderly women in the control group on cancellation, marking, and writing, respectively, were 5.17, 7.17, and 12.80, while corresponding means for college students in the control group were 7.77, 16.68, and 18.68. Separate two (age) by five (session) by two (training group : practice vs. conjugate reinforcement) analyses of variance were computed for cancellation, marking, and writing performance. In addition, one two (age) by three (training group: control vs. practice vs. reinforcement) analysis of variance was performed for each of the twelve posttest intellectual transfer measures. The results of these analyses are summarized in Tables 1 and 2. For all three response speed training rash, significant improvement was obtained over sessions for the young and the elderly women. No differences
250 I HOYER E T A L .
Table 1. Summary of Analyses of Variance on Training Tasks ~~
Training Task Source
d.t
Age (A) Treatment (TI Sessions (S) AXT AX S TX S AXTXS
1, 16 1, 16 4, 64 1, 16 4, 64 4, 64 4, 64
ap
Cancellation Marking
Writing
63.02a
104.14a
30.92a
126.79’ 1.44 30.88a 4.21
210.43a 1.84 69.6ga
35.34a
1.61
-
5.68’ 2.81 6.25’
< .005.
Note: All entries are F ratios larger than 1 .OO.
Table 2. Summary of Analyses of Variance on Cognitive Transfer Tasks Source Transfer variable
Verbal meaning Space Reasoning Number Word fluency Z PMA Word beginnings & endings Word beginnings Word endings Z Word fluency Number comparison Identical pictures Finding A s Z perceptual speed Tota I performance
Age
Treatment
9.04a 6.6ga 75.38’ 2.15 7.67’ 17.17’
2.12
3.20a 2.15 2.24 19.26’ .65.74’ 10.07’ 36.08’ 21.81
’
Age X Treatment
1.27 1.93
1.86
1.01
1.56 1.26
1.06
1.22
2.59 2.50 1.15 1.10
1.24 1.31
p < .05.
p
