This article was downloaded by: [New York University] On: 26 May 2015, At: 05:53 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Research Quarterly for Exercise and Sport Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/urqe20

Age-Related Differences in Response Programming a

Marie A. Reilly & Waneen W. Spirduso a

Department of Medical Allied Health, Division of Physical Therapy , University of North Carolina , Chapel Hill , North Carolina , 27599 , USA Published online: 26 Feb 2013.

To cite this article: Marie A. Reilly & Waneen W. Spirduso (1991) Age-Related Differences in Response Programming, Research Quarterly for Exercise and Sport, 62:2, 178-186, DOI: 10.1080/02701367.1991.10608708 To link to this article: http://dx.doi.org/10.1080/02701367.1991.10608708

PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions

Research Quarterly for ExerciseandSport © 1991 bythe American Alliancefor Health, Physical Education, Recreation and Dance Vol.62, No. 2. pp. 178-186

Age-Related Differences in Response Programming

Downloaded by [New York University] at 05:53 26 May 2015

Marie A. Reilly and Waneen W Spirduso Age-related differential effects on reaction time (R1) performance for movement complexity and response-response(R-R) compatibility were examined in children, adolescents, and young adults. A two-choiceRT paradigm involved three different finger responses, and each finger movement response was paired with every other movement response. Movement complexity was manipulated by varying the digits activated and was measured as the mean RTfor a particular movement across all choice pairs. R-R compatibility was manipulated by altering the pairing ofchoice alternatives and was determined by the mean RT comparison for each of the movements according to the paired choice alternative. Simple RTs were also obtained for all finger movement responsesfor comparison with the RTs achieved in choice situations. Age-related differences werefound for both movement complexity and R-R compatibility. Mean RT and response consistency improved with age. Although higher overall speeds werefound with age, adolescents were not significantly slower than young adults. Adolescents did, however, make significantly fewer response errors on movements differing in complexity. Bilateral versus unilateral control and number offingers involved in the task werefound to affect both movement complexity and the compatibility between response pairs. The relationship between the alternative and choice response was found to be a robust factor affecting R-R compatibility. Choice responses were significantly slower than simple responses, and the rank ordering of movement responses was identical within the two paradigms.

Key words: reaction time, response complexity, response programming, age differences, response compatibility

C

omparison studies have documented that children are consistently slower than adults in performing quick responses. Although both noncentral and central processes contribute to decrements in speed, central processing plays a dominant role (Thomas, 1980). A fairly ubiquitous pattern of response processing time decreasing steadily from middle childhood to late childhood and leveling off in adolescence has been found in investigations of processing speed. Bisanz, Danner, and Resnick (1979) demonstrated that eight-year-olds were significantly slower than older children and adults in matching tasks, suggesting that younger children employ less efficien tvisual search. Similar findings have been reported by Keating, Keniston, Manis, and Bobbitt (1980) for both visual and memory search tasks. Kail (1988) demonstrated the rate of developmental change was similar for mental addition as well as the aforementioned tasks. He hypothesized that age differences are due to a central mechanism that limits performance on speeded tasks.

Marie A.Reilly is affiliated with the Department of Medical Allied Health, Division of Physical Therapy, University of North Carolina, Chapel Hill, North Carolina, 27599. Waneen W Spirduso is an associate professor inthe Department of Kinesiology andHealth Education at the University of Texas atAustin. Requests for reprints should be sent to Marie Reilly. Submitted: March 7, 7990 Revision accepted: November 5, 7990

178

A three-stage model (Schmidt, 1988) that includes stimulus identification, response selection or determination, and response programming is a useful construct to examine age differences in central information processing. Any of these proposed stages of processing may be affected by developmentally related performance differences. A chronometric approach can be used to identify those stages of central processing affected by age. Although much is known about the early stages ofthe information processing model in children, little is known about later stages. Features of stimulus identification such as feature abstraction, identification, and encoding of the stimulus have been found to affect the processing time of children more than adults (Wickens,1974). In con trast, little is known about the stages in which children must select and program responses. Fairweather and Hutt (1978) suggested the response selection stage was the primary site ofdevelopmen tal change in information processing. In a developmental study of differences in response processing, Clark (1982) proposed that stimulus-response (S-R) characteristics affect the response selection stage and response-response (R-R) compatibility affects the response programming stage. Her comparison of kindergartners, fourth graders, and young adults on a two-choice reaction time (RT) task resulted in age-related differential effects on RT performance for S-R compatibility. Age effects, however, were not demonstrated for (R-R) compatibility. Clark (1982) referred to the ability to distinguish between two response pairs as responsediscriminability, but Kornblum (1965) had earlier used the term R-R compatibility to describe his finding that some paired responses

ROES: June 1991

Downloaded by [New York University] at 05:53 26 May 2015

Reilly andSpirduso

were more difficult to program than others. He demonstrated RTwas affected by the type ofmovementresponse serving as the paired alternatives in a two-choice RT experiment. Reaction time has been found to be faster for a finger paired with another finger on the opposite hand than when paired with a finger on the same hand (Kornblum, 1965; Schulman & McConkie, 1973); however, other researchers reported conflicting results (Reeve & Proctor, 1988). In her developmental study of kindergarten children, fourth graders, and young adults, Clark (1982) found that the choice RTs (CRTs) for same hand and different hand pairings were equivalent. Kornblum (1965) suggested same hand responses involve more inhibition, and/or competition, between responses than contralateral hand responses. However, in an immature population the observation that isolated movements may be difficult to perform (Stern, Oster, & Newport, 1980) is another factor that must be considered. The ability to inhibit movements in order to perform a differentiated movement occurs through maturational changes in the central nervous system (Kinsbourne, 1973). Isolated single finger movemen ts without associated movements occurring in other fingers increases steadily from the age of six (Kinsbourne, 1973). By nine years of age a marked decrease of these overflow movements occurs in normal children (Cohen, Taft, Mahadeviah, & Birch, 1967). Because Clark's (1982) youngest subjects were only five and six years of age, the results of her study may have been confounded by the children's inability to isolate single finger movements even in the absence of a choice paradigm. The final stage ofprocessing, response programming, may involve both response program determination and response execution (Kerr, 1978). Response programming is the stage where the movement response is structured and activated. It is, therefore, concerned with the planning and output of the movement once the stimulus has been coded and identified and the appropriate response determined. Henry (1980) suggested that during response programming subroutines are coordinated, neuromotor details are organized, and neural impulses are appropriately channeled to initiate the motor response. According to Ivry (1986), the term programming should not be limited to the computer programming concept of constructing the program but should en tail all the even ts preceding response initiation including the additional phase of program implementation. Response complexity has been found to be a robust factor influencing the time delay in response programming. Research iden tifying complexityvariables has been ongoing since Henry and Rogers (1960) noted the prolonged latency between the stimulus and movement initiation for movements of greater complexity. An underlying hypothesis of response complexity is that the more complex or difficult the movement is, the longer the program governing it and the more time needed to

ROES: June 1991

access it. Although investigators have extensively varied response demands to study the effects of movement complexity on programming time, an area of continued interest is the amount ofchange in movement complexity necessary to affect RT. Until recently the number of movement parts was considered the crucial element influencing programming time (Christina, Fischman, Lambert, & Moore, 1985). Sidaway, Christina, and Shea (1988) have reinterpreted data from several studies investigating this specific movement complexity effect and proposed that programming time is affected by the cumulative constraints placed on the motor system. They suggest that, with a greater demand for high accuracy content responses, programming time will increase due to an increase in neural organization and inhibition of unwanted motor activity. Movement complexity and R-R compatibility are both factors affecting the later stages of information processing (Light, 1988) . In the present study, the factors of movement complexity and R-R compatibility were used to examine age effects in children, adolescents, and young adults. Response speed was investigated within an information processing model, considering factors influencing both response selection and response programming. Movement complexity was varied by altering the number of controlled digits and the number of controlled sides used in a RT task. To determine age group differences in programming time related to movement complexity, the time necessary to respond with each ofthree different finger movements was analyzed for both a simple RT (SRT) paradigm and a CRT paradigm. Movement complexity was measured as the mean RT for a particular movement that in the case of the choice paradigm was averaged across all choice pairs. The factors of movement complexity considered were the number of digits activated and whether the response was unilateral or bilateral. To determine R-R compatibility, each finger was paired with every other movement response in a two-choice RT task, and R-R compatibility was manipulated by altering the pairing of choice alternatives. A response was considered to be more compatible with another response if it could be activated faster when paired with the particular response than when paired with other responses.

Method Subjects The subjects were 60 female volunteers from three age categories: children (8-9 years, M =8.9), adolescents (12-13 years, M = 13.0), and young adults (20-29 years, M = 22.0). Each group was composed of 20 volunteer subjects. Informed consent was obtained from all sub-

179

Reilly andSpirduso

jects and their guardians (in the case of child subjects). The children and adolescents were either participants in a summer camp program or students from a private school. The young adults were undergraduate university students. All subjects were right-handed nonsmokers. They were not on medication that alters response speed and were in good health. Health status was obtained by self-report from the adults and by parental report for the children.

Downloaded by [New York University] at 05:53 26 May 2015

Apparatus RT was measured using finger activated microswitches connected to a Standard Electric Time digital timer read to the nearest millisecond. The interval timer remained constant at 5 s, and on each trial a buzzer sounded a warning signal which initiated a variable foreperiod (1-3 s) prior to activation of the light stimulus. The stimulus light display was wall mounted at eye level and at a distance approximately 3 ft from the subject. Red and yellow lights measuring 2 cm across and vertically embedded in a flat black background signaled the appropriate movement for each trial.

Procedure The testing of each subject required two experimental sessions (approximately 30-40 min) on two consecutive days: a practice session followed by a test session on the second day. Subjects were seated at a table such that their forearms were resting comfortably, and index fingers and thumbs were positioned on each microswitch. Prior to testing, each of the three digit movements used to close the microswitches was demonstrated. The appropriate pairing of light stimulus and digit response also was explained .and demonstrated to each subject. To ensure understanding of the task and to rule out color blindness prior to practice trials, the subjects were asked to indicate the color of the illuminated light stimulus and what response they would make. A CRT paradigm was utilized in which the subject activated one of the two finger movements that were demonstrated and practiced prior to each block of trials. In total, three different finger combinations were used as movement responses to close the microswitches. The digit combinations to make the responses were a flexion motion of the right index finger (RI), simultaneous approximation of the right index finger and thumb into a pinch (RP), and simultaneous flexion of both index fingers (BI). When the stimulus light was activated, the subject responded as quickly as possible by pressing the appropriate switch or switches paired to the light stimulus (see Figure 1). To determine the effect of response-response compatibility on the latency required to activate a finger combination, each combination was paired with every

180

other finger combination response in blocks of50 trials. Each finger combination in a pair was randomly distributed 25 times over the block of 50 trials. A total of 150 trials was presented in three blocks to each subject. Errors in response selection were marked, and those trials were repeated at the end of each block sequence. Thus, three dependent variables were analyzed: mean latencies, standard deviations, and errors calculated for each movement response under every paired condition. The standard deviation scores, employed as an estimate of the subjects' response consistency, were calculated as the subjects' standard deviation about their own RT mean for each response type. In total there were six movement response pairings. The first response listed in the pair (upper case) represents the movement performed, and the second response (lowercase) represents the movement that was presented as an alternate choice. The effect of response complexity was determined by averaging the reaction time means, standard deviations, and choice errors for each response (RI, RP, BI) across choice pairs so the movement with which they were paired was not a factor in determining movement complexity. The data for each subject for each response were computed for only the second day of testing and were averaged across subjects in each age group. Reaction times that were two standard deviations above or below the mean were considered outliers. Subsequent analysis demonstrated the number of outliers per movement per block for each age group averaged to be one or less; therefore, outliers were discarded and means recalculated. Following the CRT trials, which provided ample practice on SRT, the subject reacted in a block on 0 trials to a SRT stimulus for each of the three movements.

~

RP

81

Figure 1. Diagrammatic representation ofthe finger responses:

right index finger trigger movement (RII, right pinch (RPI, and bilateral index finger trigger movement (BI).

ROES: June 1991

Reilly andSpirduso

Downloaded by [New York University] at 05:53 26 May 2015

Design andAnalyses A doubly multivariate analysis (DM MANOVA) was used to analyze each of the factors, compatibility and complexity, for the dependent variables considered together. The Pillai-Bartlett trace criterion was used as an omnibus test of significance (Olsen, 1976). Hierarchical tests of the age by within-subjects factors were made first, followed by tests of the main effects on the dependent variables collectively. In the event the omnibus tests were significant, follow-up MANOVAs and repeated measures ANOVAswith probability levels adjusted to protectagainst inflated alphas were performed (Schutz & Gessaroli, 1987). Compatibility. To follow up significant omnibus tests, a MANOVA was conducted for the one between factor (age: 8, 13,20 years) and the repeated measures factor with 6 levels of movement response (Rl-rp, RI-bi, RP-ri, RP-bi, BI-ri, BI-rp) to determine whether significant differences existed among the age groups and repeated measures on each of the three dependent variables of means, standard deviations, and choice errors. Again, the Pillai-Bartlett trace was used as a criterion for rejecting null hypotheses. Complexity. Because no errors were recorded on the SRT tests, only two dependent variables, the means and standard deviations (within-subject consistency), were analyzed by the DM MANOVA. Follow-up ANOVAs for repeated measures were calculated on each of these dependent variables. For both compatibility and complexity analyses, probability levels were Bonferroni-adjusted at .05/3, or p « .017. A separate ANOVA was completed on the SRT error scores, also using p « .017 as the criterion level for rejection of the hypotheses.

Results The omnibus test of factors relating to compatibility revealed that age interactions with response selection failed to reach significance-that is, significant age and response selection differences were indicated, but they were similar across the three ages studied. Consequently, interpretation of the follow-up analyses for the CRT means and standard deviations were made only for the main effects of age and response selection. Similarly, omnibus tests of age interactions with the factors of reaction type (SRT, CRT) and response type (RI, RP, BI) were not significant. The only multivariate interaction test to reach significance was reaction type with response type. That is, for complexity, the DM MANOVArevealed significant differences existed in speed ofresponse andl or consistency among the response types (RI, RP, BI), depending on whether the reaction was a simple or choice reaction protocol. The F tests in the follow-up

ROES: June 1991

repeated measures ANOVAs for means, standard deviations, and errors were examined only for the main effects of age, reaction type, and response type and for the interaction ofreaction type by response type.

Age Effects The Pillai-Bartlett approximated F test was highly significant for complexity age main effects, F (4, 114) = 7.61, p « .001. Follow-up ANOVAs, with adjusted plevels to protect against alpha inflation, revealed strong age effects for both mean latencies, F (2,57) = 20.08, p« .001, and response consistency, F (2,57) = 8.21, p « .001. The post hoc tests showed that for both dependent variables, the children were significantlyslower (Bonferroni/Dunn, critical difference [BID, cd] = 33.02, P< .001) and less consistent (p < .001) than the adults. Adolescents and adults did not differ, except on errors made on all the responses taken together, F (2,57) = 5.47, p« .01. Twentyyear-olds made more errors than the adolescents (M=5.27 vs. M = 2.95; cd = 2.23, P< .007) The results for compatibility were similar. The age group main effectwashighlysignificant,F(6, 112) =3.29, P < .005, indicating that differences existed among the age groups on one or more of the means, standard deviation, and error dependent variables. Differences in the three dependen t variables also existed on the withinsubject response selection factor, F (15,43) =9.88, p« .001. Follow-up repeated measures MANOVAs revealed robust age differences on means, F (2,57) = 9.18, p « .001, and standard deviations, F (2,57 = 6.12, P < .004, but not on errors. Thus, the children were significantly slower than adolescents (BID, cd = 44.01, P < .001) and adults (P < .001), but the adolescents and adults were not significantly different. The children were also significantly less consistent than adolescents (BID, cd = 16.94, p< .025) and adults (p< .002), but adolescents and adults were not different. Errors were few and were not significantly different across the age groups. In summary, for both complexity and compatibility, means and standard deviations of children (8 years) were significantly slower than both adolescents (13 years) and young adults, but the two latter age groups did not differ. No age differences were significant for the compatibility factor; however, the young adults made significantly more errors than the other two age groups.

Movement Complexity Effects The multivariate analysis revealed that both RT type, F (2,56) = 91.67, p« .001, and response type (RI, RP, BI), F (4, 54) = 21.97, P < .001, combined to influence responsespeed, consistency, or both. The RT type by response type was significant, F (4,54) = 4.66, p« .003, and the follow-up repeated measures adjusted ANOVAs revealed that a significant interaction existed in response

181

Downloaded by [New York University] at 05:53 26 May 2015

Reilly andSpirduso

consistency, F (2, 114) = 9.94, P< .001, but not in mean latencies. Not surprisingly, SRT was significantly faster than CRT,F(I, 57) = 163.71, P< .001, and more consistent, F (1,57) = 160.54, P< .001. The rank order ofresponse speed from fastest to slowestRTs for all age groups was RI, RP, BI, F (2,114) = 61.79, P< .001, and all were significantly different from each other (P < .001). Although the SRTs were faster than the CRTs, the rank ordering ofthe response types from fastest to slowest (RI, RP, BI) was the same for simple and choice paradigms (see Figure 2). Scheffe contrasts demonstrated BI was a significantly more complex movement to program than either RI or RP (P < .001); the unilateral movements RI and RP were faster to program than the bilateral movement BI (P < .001), and the one-digit movements were faster to program than the two-digit movements RP and BI (peOOl). For all groups, as the latency for each of the CRT response types increased, implying that greater response complexity required more programming time, the response consistency for each movement decreased. As in mean latencies, the response consistency order was RI, RP, BI. However, although consistency was significantly different among the three response types on CRT, no difference in consistency was found between the RI and RP response types in SRT. Thus, for movements in the CRT paradigm, each response type was significan tlymore difficult to program consistently, butin the SRTparadigm, only the bilateral response (BI) was inconsistently programmed. The significant complexity effect for choice errors, F (2,114) = 52.11, P < .001, is illustrated in Figure 3.

For all groups significantly fewer errors were made for RI and RP responses than for the BI response (P< .05).

R-R Compatibility Effects Highly significant differences existed among the movement pairs, F (15,43) = 9.88, P< .001, but, as indicated earlier, these differences were similar for all age groups. Follow-up tests showed the rank order of compatibility pairs from fastest to slowest RTs was approximately the same for all age groups. The post hoc Scheffe contrasts revealed three subgroups: a fast movement grouping (RP-ri, RI-rp, RI-bi) in which RTs were not significantly different from each other; an intermediate movement grouping (BI-ri and RP-bi), which were not significantly different from each other but which were significantly slower than the three fast movement groups; and a movement (BI-rp), which took significantly longer to produce than all other movement pairs (see Figure 4). The rank order of movement pairs, from most consistent to least consistent, was the same for all age groups (RI-bi, RI-rp, RP-ri, RP-bi, BI-ri, BI-rp). Although this ordering was slightly different from that for RT means, the responses appear in the same three significant movement groupings found for the RTs mean comparisons. The main effect for consistency of movement pairs was significant, F (5,285) = 15.49, p-

..

420

II

400

i5. E

380

• U

'".. ....E

• ...••

...... II

.!

•

III

• • l2l

360

0

~

340

m

320

8

CBI CRP CRI SBI SRP SRI

280 260

CRI

6

.. .•. ......

300

CBI CRP

4

2

240 220

0

Childr.n

AdoI.sconts

Adults

"'. Figure 2. Meanreaction time±standard errors(ms) of children (M= 8.9 years), adolescents (M =13.0 years), andadults (M= 22.0 years) for eachresponse type undersimple (S) and choice (C) reaction timeconditions. Response typeswere: right index finger (RI), right pinch (RP), andbilateral indexfingers(BI). Example: responding with the rightindexfinger as a simple RT response = SRI; with a right pinch selected from a choice of bilateral indexor right pinch =CRP.

182

Childron

Adolosconts

Adults

Age

Figure 3. Mean number of movement errors± standard errorsof children (M= 8.9 years), adolescents (M= 13.0 years), and adults (M =22.0 years) on choice movements: right index(RI), right pinch (RP), and bilateralindex(BI). Bars belowthe abscissa bracket nonsignificant differences among the movements for eachage group. Movement error =selectionof the incorrect response in the CRT paradigm.

ROES: June 1991

Reilly and Spirduso

Downloaded by [New York University] at 05:53 26 May 2015

differ in consistency on the movement pairs RP-bi and BI-ri; however, subjects were less consistent on these two movement pairs than the above three movement pairs (RI-bi, RI-rp, RP-ri ). Performance on all five movement pairs was significantly more consistent than it was on the slowest RT movement pair, Bl-rp, The number of errors made was significantly different among the movement response pairs, F (5, 285) = 36.35, p« .001. All groups made more errors in movement pairs where the bilateral movement was the RT choice (Bl-rp and BI-ri). The post hoc contrasts further indicated that for all groups these two movement pairs were not significantly different from each other (p » .05). As mentioned earlier, the age groups were not significantly different in the number of errors produced, nor was the interaction of age and response selection significant.

Discussion Movement Complexity In general, the present findings indicate the speed of response programming of movements differing in complexity is slower in children than in adults. This is in agreement with previous research (Fairweather & Hutt, 1978; Thomas, 1980). Since these age differences in processing speed appear to follow a pattern similar to the patterns found by Kail on other tasks (1988), it is tempting to suggest they reflect a general developmen tal change. However, the number ofage groups in this study was limited, and the design was cross sectional. A specific

i... :;j ...E

. ...... .e ..,:".. U

420

II1II

400

EJ II1II

3BO

(l3

0

360

0

0

BI-rp BI-r; RP-bi RI-b; RP-ri RI-rp

340

E

320

j:: 300

ell:

2BO 260 240

Childr."

Adol.sc."ts

Adults

Age

Figure 4. Mean reaction time ± standard errors (ms)of children (M =8.9 years), adolescents (M= 13.0 years),and adults (M =22.0 years) for movements right index (RI), right pinch (RP), and bilateral index(BI) when paired with each alternative in a choice response. Bars below the abscissa bracket nonsignificant differences among the movements for each age group. Least compatible pairs produce longertimes, and most compatible pairs produce shortest times.

ROES: June 1991

growth function could not therefore be determined. Possible causes of slower RT in children may be inefficient organization of processes to perform particular tasks (Chi,1977), limitation in the quantity of mental effort and attention, or less efficiently allocated available resources (Kail, 1988). Age did not interact with complexity. Therefore, one could conclude, as did Larish and Stelmach (1982) for aging individuals, responses of children are programmed in the same way as young adults but are simply slower. The mean RTs of the adolescents and the adults indicate a trend toward faster latencies in the adult group, but differences did not reach significance. Although a less common finding, nonsignificant response time differences between adolescents and young adults have also been reported (Clark, 1982). It appears the small changes in task complexity utilized in this study were easily accommodated by young adolescents. As has been reported numerous times, bilateralversus unilateral control appears to be a definitive movement complexity factor (Light, 1988; Stern et al.,1980). The bilateral response was slower to program than either unilateral response for both SRT and CRT paradigms. In an investigation of RT in measurement of hemispheric lateralization, Stern et al. (1980) found bilateral RT was longer than single-handed RT for the same movement for their adult age groups. In contrast to the findings for the adults, single-handed RT was longer than bilateralRT for subjects 7 years of age and not significantly different for9 and 11 year olds. The authors suggested the decision time in the unilateral task may have been longer for the youngest children due to central nervous system immaturity, since it included time for response inhibition, which would not be necessary in the bilateral tasks. They contended, as was demonstrated in the present investigation, with increasing age, control of the motor system improves to allow specifically directed responses. Since movement of the digits is under the control of the contralateral hemisphere, coordinated bilateral communication would take longer than single-handed movement controlled by one hemisphere. Although less distinct, the number of controlled digits also appears to be a factor ofmovement complexity. In general, a response with a smaller number ofdigits was faster to program. The unilateral one-digitresponse (RI) was faster to program than the unilateral two-digit response (RP), a finding also reported for older adults (Light, 1988). Consistentwith Light's (1988) findings, RI was also faster than the bilateral two-digit response (BI). However, the two-digit responses in this study (RP and BI) were also significantly different from one another for programming time. In a previous study (Light, 1988), a four-digit movement (two bilateral pinch responses) was found faster to program than a two-digit movement (bilateral index). Light's study suggests the number of digits controlled is not a complexity factor. These con-

183

Downloaded by [New York University] at 05:53 26 May 2015

Reilly andSpirduso

flicting results are not surprising since bilateral responding is a factor of complexity that may confound the issue of the number of digits involved in the response. Because there has been much debate as to the type of RT paradigm appropriate to study complexity (Henry,1980; Ivry, 1986; Klapp,1980; Marteniuk & MacKenzie,1981), both SRT and CRT pardigms were analyzed for this study. The results obtained were the same for both SRT and CRT. The rank ordering of the movements was identical within the SRT and CRT paradigms (RI, RP, BI). Klapp (1980) argued that differences in programming could not be seen in SRT because it is theoretically possible to preprogram the SRT response prior to the stimulus, thus allowing the program to be triggered at stimulus onset. Results of this study indicate if subjects preprogrammed the finger movements, they were able to do this in the CRT as well as the SRT paradigm. Although Schmidt (1988) suggested complexity factors operate in both CRT and SRT paradigms, controversy still exists over which paradigm is appropriate. In defense of his choice of using a SRT paradigm in his pioneeringwork, Henry (1980) explained the use ofCRT would cause increases in latency such that the time needed to select the movement could not be separated from actual program selection and processing of the response.Ivry (1986) suggested SRT and CRT illuminate different aspects ofthe response preparation period and could be used in a complementary manner when examining complexity effects. The results from the present study indicate that either SRT or CRT may be used. Marteniuk and MacKenzie (1981) provided numerous examples from the literature where equivocal results for the CRT versus SRT controversy were found.They concluded CRT falsely estimates the effects of processing involved in movement organization and also suggested when there is a choice of responses in an experimental task the issue of compatibility must be considered. In agreement with previous studies (Clark, 1982; Fairweather & Hutt, 1978), response consistency was also found to be age dependent. Children were less consistent than adolescents and adults, but there was no significant difference between the two older groups. Overall, maturational studies support this finding since they demonstrate that performance becomes progressively less variable with increasing age (Clark, 1982; Fairweather & Hutt, 1978; Stern et aI., 1980; Thomas,1980). In terms of movement complexity, consistency of programming time was significantly greater for the bilateral response than for the two unilateral responses, which were not different. A surprising finding in the present study was that the three age groups were not different in the number of errors made in each of the response types. All three groups made significantly more errors in the bilateral response than the two unilateral responses. This finding

184

was similar to Clark's (1982) study. She found that 6-yearolds and 10-year-olds did not differ from a young adult group on the number of errors made on a CRT task.

R-R Compatibility In general, response programming speed of choice movement pairs that differed in compatibility increased with age. As was true with movement complexity, children were significantly slower than adolescents and young adults, but adolescents and young adults were not significantly different. These results do not agree with those ofClark (1982), who found no significant response speed differences among 6-year-olds, 10-year-olds, and young adults. Clark (1982), however, examined R-Rcompatibility (response discriminability in her study) between only two movement pairs, whereas in the present study compatibility was analyzed by providing six different twochoice movement pairs. In the present study movement initiation time of a specific movement was dependent on the alternative movement choice. In general, if a unilateral movement was the alternative choice in a movement pair, a specific response type was programmed faster than if a bilateral movement was the alternative choice. These findings are in agreement with the movement complexity results. In terestingly, when a unilateral response was paired with its analogous movement on the contralateral hand (RIbi), it was not found to be significantly faster than when it was paired with another finger response on the same hand (Rl-rp). This finding is congruent with Clark's (1982) findings for response discriminability for single finger response pairings. Investigators have reported that CRT is faster for a finger response paired with another finger on the contralateral hand than when paired with a finger on the same hand (Kornblum, 1965; Schulman & McConkie, 1973). These studies and Clark's investigation differed from the present study in that the alternate movement was bilateral in this study and unilateral in all the other investigations. As indicated in the discussion ofmovemen t complexity, bilaterality is a confounding variable that is likely to increase programming time. When Heuer (1982) investigated R-R compatibility for movement responses with different characteristics, he proposed CRT is shorter if the response alternatives involve similar central processes and structures and relationships between responses may be determined by the degree of structural and functional interference. Since all participants in Heuer's study were adults, the conflicting findings may also be related to age differences. In the present study and in Clark's (1982) developmental study, age did not interact with compatibility level; therefore, differences between response pairs could not be compared within each age group. The ability to discriminate structural and functional differences in

RDES: June 1991

Downloaded by [New York University] at 05:53 26 May 2015

Reilly andSpirduso

choice response pairs may be developmentally based and determined by maturational processes, as is the case for movement differentiation and motor inhibition (Cohen et al., 1967; Kinsbourne, 1973). Clark (1982) described the discrimination between two responses as the ability to distinguish between motor commands within a set. In the present study the six movement pairs were divided into three significantly different clusters as a result ofRT analysis. The movement pairs can be differentiated depending on whether the correct choice movement response was unilateral or bilateral and on the number of fingers in the movement pair. A third factor that appears to influence programming time strongly was whether the two movement responses were associated subsets (i.e., whether a response was a partial movemen t of the other response) . Therefore, in this study R-R compatibility appeared to contribute to movement complexity so that at least three interacting factors influenced the programming time. For the movement grouping with the fastest RTs (RP-ri, Ri-rp, RI-bi) the correct choice response was unilateral, three fingers were involved, and for two of the pairs the correct choice movement was a subset of the alternate movement (Ri-rp, RI-bi ). In the case of the third pair the alternate movement was a subset of the correct choice movement (RP-ri). The intermediate grouping consisted of response pairs BI-ri and RP-bi. The BI-ri correct response was bilateral and involved three fingers, butitalso contained a subset; the alternate movement (ri) was a subset of the correct choice movement (bi). In contrast, for the second response pair (RP-bi), the correct response was unilateral but involved four fingers and no subsets. The slowest movement grouping occurred when BI was paired with rp. In this pairing the correct choice movement was bilateral. It involved four digits, and there were no subsets; thus, this movement pair was the least compatible. Response pairs with the shortest RTs were most consistent and those with the longest RTs were the least consistent. Although the rank order for consistency was slightly different from that of RT means, the analysis revealed three distinct groupings that were identical to the RT mean data: three fastest and most consistent pairings (RI-bi, Rl-rp, RP-ri), two intermediate pairings (RP-bi, BI-ri) , and one slowest and least consistent pairing (BI-rp). The finding that fast compatible responses are programmed more consistently than slow uncompatible movement pairs is in agreement with Heuer's (1982) findings. Although the primary concern of this study was to determine the effects ofage on movement complexity and R-R compatibility, the stages of information processing associated with these factors deserve comment. Ithas been well established that movement complexity affects the

ROES: June 1991

final stage of processing, response programming (Schmidt,1988). Since R-R compatibility is less understood, the stage or stages affected by this factor continue to be of interest. When Schulman and McConkie (1973) investigated S-R and R-R compatibility, they proposed S-R effects influenced only the response selection stage, but R-R effects may be involved in either the response selection or the response programmingstage. Clark (1982) rejected this proposal, suggesting the R-R effects are localized to the response programming stage. She argued the results of Schulman and McConkie's study (1973) were congruent with Sternberg's (1969) Additive Factors logic since the combined effects ofthe two variables in the study were additive and did not interact. Examination of detailed models of information processing indicate it is more likely that R-R compatibility is a variable that can affect both information processing stages, response determination and response programming. In Theios's (1975) five-stage model, response selection or determination is considered to be the stage where a cognitive response is selected. His model, however, articulates two stages where decisions are made, and it is in the subsequent response decision stage, response program selection, the motor elements to organize and execute the particular motor response are selected. This second phase decision stage is considered to be contained in the final information processing stage, response programming (Kerr,1978). With this perspective, R-R could be considered a factor of both the response determination and response programming stages. In summary, the findings of this study provide evidence that age influences movemen t complexity and R-R compatibility. Both speed ofresponse and response consistencywere greater in adolescen ts and young ad ults than in children. Although an overall RT increment was found with increased age, adolescents reached young adult performance levels indicating that minor changes in complexity may be easily distinguishable by early adolescence. The rank ordering of response difficulty was not different across age groups; therefore, it appears children are just slower and less efficient in the utilization of control processes than young adults. Gradations of movement complexity and R-R compatibility were found to affect response time, thus altering the overall difficulty ofa task. Complexity factors operating in this study were clearly identified by using both a SRT and a CRT paradigm. Bilateral versus unilateral control and number of fingers involved in the movement appear to be factors that affect both movement complexity and compatibility between response pairs in a binary CRT task. R-R compatibility also appears to be strongly influenced by the relationship between the alternative and choice response pairs.

185

Reilly andSpirduso

Downloaded by [New York University] at 05:53 26 May 2015

References Bisanz.j., Danner F., & Resnick, L. B. (1979). Changes with age in measures ofprocessing efficiency. Child Development, 50, 132-14l. Chi, M.T. H. (1977) .Developmen tal change in mental rotation. Developmental Psychology, 13, 543-544. Christina, R W., Fischman, M. G., Lambert, A. L., & Moore, J. F. (1985). Simple reaction time as a function ofresponse complexity: Christina et al., revisited. Research QJ.tarterly for Exercise and Sport, 56, 316-322 Clark, J. E. (1982). Developmental differences in response programming. Journal ofMotor Behavior, 14, 247-254. Cohen, H. J., Taft, L. T., Mahadeviah, M. S., & Birch, H. G. (1967). Developmental changes in overflow in normal and aberrantly functioning children. Journal of Pediatrics, 71, 39-47. Fairweather, H., & Hutt, S. J. (1978). On the rate of gain of information in children. Journal ofExperimental Child Ps~ chology, 26,216-229. Henry, F. M. (1980). Use of simple reaction time in motor programming studies: A reply to Klapp, Wyatt, and Lingo. Journal ofMotor Behavior, 12,163-168. Henry, F. M., & Rogers, D. E. (1960). Increased response latency for complicated movements and a "memory drum" theory ofneuromotor reaction. Research QJ.tarterly, 31,448458. Heuer, H. (1982). Binary choice reaction time as a criterion of motor equivalence: Further evidence. ActaPsychologica, 50, 49-60. Ivry, R B. (1986). Force and timing components of the motor program. Journal ofMotor Behavior, 18, 449-479. Kail, R (1988). Developmental functions for speeds of cognitive processes. Journal ofExperimental Child Psychology, 45, 339-364. Keating, D. P., Keniston, A. H., Manis, F. R, & Bobbitt, B. L. (1980) . Development of the search-processing parameter. Child Development, 51, 39-44. Kerr, B. (1978). Task factors that influence selection and preparation for voluntary movements. In G. E. Stelmach (Ed.), Information processing in motor control and learning (pp. 56-69). New York: Academic Press. Kinsbourne, M. (1973). Minimal brain dysfunction as a neurodevelopmentallag. Annals ofthe New York Academy of Science, 265, 268-273. Klapp, S. T. (1980). The memory drum theory after 20 years: Comments on Henry's note. Journal ofMotor Behavior, 12, 169-17l. Kornblum, S. (1965). Response competition and/orinhibition in two choice reaction time. Psychonomic Science, 2, 55-56. Larish, D. D., & Stelmach, G. E. (1982). Preprogramming, programming, and reprogramming of aimed hand movements as a function of age. Journal ofMotor Behavior, 14, 322-340.

186

Light, K. E. (1988). Effects ofadult aging on response programming complexity and compatibility. Unpublished doctoral dissertation, The University of Texas, Austin. Marteniuk, R G., & MacKenzie, C. L. (1981). Methods in the study of motor programming: Is it just a matter of simple vs. choice reaction time? A comment on Klapp et al. (1979).JournalofMotorBehavior, 13,313-319. Olsen, C. L. (1976). On choosing a test statistic in multivariate analysis ofvariance. Psychological Bulletin, 83, 579-586. Reeve, T. G., & Proctor, R W. (1988). Determinants of twochoice patterns for same-hand and different-hand finger pairings. Journal ofMotor Behavior, 20, 317-340. Schmidt, RA. (1988). Motor control and learning: A behavioral emphasis (2nd ed.). Champaign, IL: Human Kinetics. Schulman, H. G., & McConkie, A. (1973). S-R compatibility, response discriminability, and response codes in choice reaction time.JournalofExperimentalPsychology, 98,375-378. Schutz, R. W., & Gessaroli, M. E. (1987). The analysis of repeated measures designs involving multiple dependent variables. ResearchQJ.tarterlyforExerciseandSport, 58, 132-149. Sidaway, B., Christina, R W., & Shea,J. B. (1988). A movement constraint interpretation ofthe response complexity effect on programming time. InA. M. Colley & J.R Beech (Eds.), Cognition and action in skilled behavior (pp. 87-102). Amsterdam: North Holland, Elsevier Science Publishers. Stern, AJ., Oster, P.J., & Newport, K. (1980). Reaction time measures, hemispheric specialization and age. In L. W. Poon (Ed.), Aging in the 1980s: Psychologicalissues (pp. 309326). Washington, DC: American Psychological Association. Sternberg, S. (1969). The discovery of processing stages: Extension ofDonder'smethod. ActaPsychologica, 30, 276-315. Theios, J. (1975). The components of response latency in simple human information processing tasks. In P. M. A. Rabbitt & S. Dornic (Eds.), Attention and performance, vol. 5 (pp. 418-439). New York: Academic Press. Thomas.]. R. (1980). Acquisition of motor skills: Information processing differences between children and adults. Research QJ.tarterly for Exercise and Sport, 51, 158-173. Wickens, C. D. (1974). Temporal limits of information processing: A developmental study. Psychological Bulletin, 11, 739-755.

Authors' Notes This research was conducted as part of a dissertation submitted by the first author in partial fulfillment of the requirements for the Ph.D. degree at The University of Texas at Austin.

ROES: June 1991