HUMAN FACTORS, 1972,14(1), 45-50

Hand Steadiness During Unrestricted Linear Arm Movements PHILIP G. MEAD and PHILIP B. SAMPSON, Tufts University, Medford, Massachusetts

Factors influencing hand steadineu were inuestigated in 66 subjects performing unrestricted linear arm mouements. Results demonstrated that most hand tremor occurs in the updown plane, with right-left and in-out tremor occurring less frequently. /? addition, arm mowments made by subjects uiewinga target by means of a mirror were found t o produce more errors than those made by direct uiewing. Other literature is cited, and implications are discussed for the utility of a hand-ateadinerr procedure in dental practice.


Hand steadiness has frequently been investi- gated in tasks which require the subject to maintain a stylus in a fixed position, often in a small hole. Movements which occur while attempting to hold the hand steady are a measure of hand steadiness, and appear to be related to several factors, among them the size of the hole (Whipple, 1914) and whether the arm is supported (Herbert, 1957). Hand tremor has also been studied while the arm is in motion during pursuit movements (Corrigan and Brogden, 1948), positioning movements (Brown, Knauft, and Rosenbaum, 1948), and free movements (Herbert, 1957). Much con- flicting evidence, however, has resulted from these and other studies. Some of the conflicts have been attributed to the different conditions under which the experiment was performed. Herbert (1957) has pointed out that divergent results may reflect the fact that different types of responses were required in the various studies, and has suggested that the following variables be accurately defined: (a) short vs. long arm movements, (b) pursuit vs. free arm movements, (c) movements involving friction vs. unrestrained arm movements, (d) arm move- ments with or without visual feedback, (e) arm movements with or without performance feed- back, ( f ) fast vs. slow arm movements, (9) arm movements involving various body postures,

and (h) arm movements involving various target configurations.

The present study is concerned with hand steadiness during slow, unrestricted, linear movements of the unsupported arm. Several variables and hypotheses were investigated:

1. Position of the subject during the arm movement. It was expected that hand tremor would be minimized if the subject were seated rather than standing because of body sway and fatigue in the latter case.

2. Direction and orientation of the target. Herbert (1957) has indicated that short, linear arm movements in the up-down direction are the most accurate, while those in the right-left direction are least accurate. Similar results were predicted in the immediate investigation for slightly longer arm movements.

3. Mode of target observation. Subjects in the present experiment viewed the target directly as well as through a mirror while performing the task, with higher accuracy predicted in the former condition.

4. Type of subject. Two types of subjects participated in the present study: dentists and nondentists. Because dentists frequently per- form fine arm movements with the aid of a mirror, it was predicted that they would show fewer errors with mirror viewing than non- dentists.

The study had several objectives. First, it was felt that if a reliable index of hand


at EASTERN MICHIGAN UNIV on February 19, 2015hfs.sagepub.comDownloaded from

46-February, 1972 H U M A N FACTORS

steadiness could be achieved by using the proposed equipment, such a test might be valuable in assessing motor fatigue during various dental procedures. Second, it was sug- gested that such a procedure might be a useful selection device for dental students. Third, certain recommendations for the design of dental equipment might arise from the hand steadiness test. And, fourth, findings resulting from the present study might add support to those of other studies cited in the arm move- ment literature.



A total of 66 subjects took part in the experiment, half of the group were dentists and half were nondentists. Ages ranged from 22 to 67 years and all subjects were right-handed males.

Appora t us

The equipment consisted of a rectangular frame 5 ft. high, 2 ft. wide, and 2 ft. deep with inside runners spaced 2 in. apart up and down both sides. A cabinet containing the hand steadiness target was situated inside this frame on the runnets a t a height corresponding to the subject’s position, and was easily adjustable. Inside the cabinet was a circular illuminated metal disk containing a narrow groove 0.5 in. wide and 12 in. long which could be adjusted to project from right to left or from front to back in either the horizontal or vertical plane.

covered with heavy plastic, and was affixed to the base of a dental chair containing a motor suitable for height adjustment via a foot switch.


Subjects were required to move a stylus (15 in. long with a 4-in., 90” bend a t the tip) inside the groove without touching either side. The stylus was electrically wired to a tone generator, which produced a mild audible tone for each “hit” on either side of the groove, and to two electronic counters which recorded the number of hits per trial. Finally, an electrical timer, actuated at the beginning of each trial, was employed to time each arm movement.

In front of the hand steadiness apparatus, a seat was provided which consisted of a me d iu m-so f t , f oam-ru b ber-filled surface

Each subject was read the following instruc- tions:

“The equipment you will be using has been built to test your handeye coordination in a variety of situations. Your task will be the same throughout the experiment; namely, you will be asked to move the pointer along the slot outward and back while trying to touch the edges as little as possible. If you touch either side you will hear a tone, and this is to tell you that you are making contact with the edge. We would like to ask you to complete each trial in approxi- mately 5 seconds, so before you begin there will be a few practice trials.. . [Each subject then received 5 practice trials] . . . Now that you are familiar with the apparatus, here is how the experiment will be run. Hold the wand against the relay at the near end of the slot. You will hear a click. When I say ‘ready, begin,’ move the wand along the groove, trying not to touch either edge. Touch the far end and then proceed back to the start. When you amve at the starting point, hold the wand against the relay while 1 record the data. We will then repeat the task twice more and move on to a new trial. Are there any questions?

There will be several different ways in which we will conduct this task. Some will have you standing, mme sitting, and some perching; this last position being one in which you will lean against the seat such that some but not all of your weight is absorbed by the seat. At times we will use a mirror, other times not, and occasionally we will swivel the target 90 degrees. Your task, however, will remain the same- that of moving the wand in the goove out and back while attempting to touch the side as little as possible.”

Next, subjects were positioned in front of the cabinet at a distance such that with the stylus resting a t the end of the groove the subject’s forearm was horizontal and his elbow was directly opposite the rib cage. The stylus contained a wooden grip which the subject grasped in a manner similar to grasping a tennis racket. Eight different target configurations were presented with the subject in each of 3 positions: standing, perching, or sitting. I t took each subject approximately 0.5 hr. to complete the total of 24 trials. Order effects were con- trolled by randomizing the sequence in which trials were presented to subjects. The diagrams

at EASTERN MICHIGAN UNIV on February 19, 2015hfs.sagepub.comDownloaded from

P H I L I P G. MEAD A N D P H I L I P B. S A M P S O N February, 1072-47

in Figure 1 illustrate the cabinet portion of the apparatus together with the various arm move- ments and direction of tremor associated with each movement. Three directions of arm move- ments and three types of hand tremor were investigated for both direct and mirror viewing.


The major results are presented in Table 1 in the form of mean errors (“hits’? for all subjects for every condition. Initial inspection of the data revealed that certain conditions produced more hand tremor than others. Specifically, tremor was much greater for in-and-out movements than for either up-down or right- left movements. Also, when the mirror was used to view the target, subjects appeared to make more errors regardless of postural posi- tion or dental affiliation than when direct viewing was allowed (Mean hits with mirror = 146.5; Mean hits without mirror = 123.8). To examine these and other possible rela

A. B.













tionships, a 5-way analysis of variance was performed on the data in the form of a split-plot factorial design (type SPF-2.3222), as outlined in Kirk (1960). The results are shown in Table 2, where the main effects and signifi- cant interactions are reported. Several treat- ment conditions reached significance.

(1) IIand steadiness differed quite signifi- cantly when trials with and without the mirror were compared, as indicated Ly an I: ratio of 27.93. Inspection of the mean errors in these two conditions in Table 1 reveals that the mirror condition is the one that contributed most frequently toward higher error scores.

(2) Target orientation conditions (horizontal versus vertical targets) produced an I: ratio of 10.99, also highly significant, indicating that more tremor was induced during arm move- ments with vertical targets than with horizontal targets; however, such a comparison may not be meaningful because type of arm movement and associated tremor are confounded within t h e two target orientations. (Ilorizontal targets














TREMOR: IN-OUT Figure 1. Sketches of the hand steadiness apparatus illustrating the types of a m movements and associated tremor that were investigated. (Note: Lower row represents conditions in which a mirror was used.)

at EASTERN MICHIGAN UNIV on February 19, 2015hfs.sagepub.comDownloaded from

4R-February, 1972 H U M A N FACTORS


Mean Errors (“Hits”) for All Conditions of the Hand Steadiness Experiment ~~

A. Target: Vertical, 90′

Arm Movement: In-Out Tremor: Up-Down

(Grand Mean = 239.81

B. Target: Horizontal, 0″

Arm Movement: ln-Out Tremor: Right-Left

(Grand Mean = 208.5) ~~ -~

With Mirror Without Mirror

Non Non Dentist Dentist Dentist Dentist

With Mirror Without Mirror

Non Non Dentist Dentist Dentist Dentist

~ ~~ ~~~~~ –

Standing 237.9 270.1 220.7 265.8 21 7.2 236.5 184.9 215.5 Perching 265.7 254.3 213.2 248.3 213.8 219.1 171.1 185.9 Sitting 223.6 245.7 206.8 225.7 236.2 222.8 191.3 207.3

Mean 242.4 256.7 213.6 246.6 222.4 226.1 182.4 202.9

C. Target: Vertical, 0″

Arm Movement: Up-Down Tremor: In-Out

(Grand Mean – 58.7) 0.

Target: Horizontal, 90″ Arm Movement: Right-Left

Tremor: In-Out (Grand Mean = 41.8)

~~ ~~~~~ ~ ~~

With Mirror Without Mirror

Non Non Dentist Dentist Dentist Dentist

With Mirror Without Mirror

Non Non Dentist Dentist Dentist Dentist

Standing 56.4 63.2 53.6 38.1 Perching 81.3 54.6 62.1 41.1 Sitting 73.7 57.3 66.4 50.1

51.6 39.0 40.4 29.2 50.3 41.2 41.2 31.8 57.0 46.2 37.2 35.8


Mean 70.5 58.4 60.7 45.3 53.0 42.1 39.6 32.3


Summary of Analysis of Variance for Hand Steadiness Data

Source of Variance df MS F

Between Subjects Dental Affiliation (A) Subjects Within Groups (S)

Within Subjects Postural Position (P) P X S Mirror (M) M X S Target Direction (D) D X S Target Orientation (0) o x s D X O D X O X S


65 1


1518 2

128 1

64 1

64 1

64 1


5,322.1 1 0.12 43.095.43

2.1 23.08 L.29 7.41 1.98

131,956.44 27.93′ 4.724.98

17.653.97 1.12 15,739.07

208,401.50 10.99′ 18,963.50

12,054,726.03 51 1.36′ 23,573.68



at EASTERN MICHIGAN UNIV on February 19, 2015hfs.sagepub.comDownloaded from

PHILIP G . MEAD A N D P I I I L I P B . S A h l P S O N February, 1972-49

measured in-out and right-left tremor while vertical targets measured in-out and up-down tremor.)

(3) A highly significant interaction was found between target direction and target orientation (F = 511.36) indicating that when type of tremor is taken into account while comparisons are made across different target configurations, realistic performance differ- ences emerge. For example, it can be seen in Table 1 that when the u down vertical target condition is turned 90 , many more errors occur because the direction of tremor being measured changes from in-out to up-down.



The literature in motor skills learning appears to be handicapped to a large extent by the wide variability found in the types of tasks employed. In arm movement research, for example, it has been pointed out that authors often neglect specifying whether a particular task involves slow, tension arm movements or ballistic movements, whether or not the target is visible, whether or not the task is paced, and so on. These and other factors must be care- fully indicated if generalities are to emerge from findings in such types of motor skills research. The immediate experiment is equally vulnerable to these problems in that studies closely resembling it do not exist a t the present time. Still, certain aspects of the hand steadi- ness task can be discussed in relation to other similar experiments. Corrigan and Brogden (1948) have shown that errors occurring during continuous arm movements in the horizontal plane are in part a function of the angle in which the movement is made. Their task involved moving a stylus in a narrow track at a constant velocity. All movements were short, paced, linear responses varying only in the angle in which they occurred. It was found that the least number of errors resulted from move- ments at angles of 135″ and 315″ (1:30 and 7:30 positions) while most errors occurred a t angles of 45″ and 225″. These findings are consistent with those of the present experi- ment, where more errors resulted from in-out

arm movements than from right-left move- ments; however, one must realize that in Corrigan and Brogden’s study the subjects were paced and the stylus was restricted in move- ment while in the present task subjects re- sponded at their own pace (although a 5-sec. trial duration was encouraged) with an un- restricted stylus. As a result, the similarities between the two studies must be viewed with caution.

IIerbert (1957) studied the speed and accuracy of linear arm movements in six differ- ent directions (right, left, in, out, up, down) in a task involving quick, short arm movements. Subjects grasped a cylindrical balsa grip and moved it a total of 2.75 in. into a small cylinder a t each of six locations. In general, the study demonstrated that up-down movements were most accurate, push-pull movements next, and right-left movements least accurate. These findings are not entirely in accord with the present results, where accuracy was best for right-left movements and worst for up-down movements. Once again, however, several aspects of Herbert’s study are different from those of the immediate investigation, among them target length and type of arm movement. Hand tremor reported here reflects hand steadi- ness during slow, tension movements over a 24-in. path, while the above study represents accuracy of arm movementv during rapid, short positioning trials. As such, any disparity between the t w o experiments may be due in part to methodological differences rather than contrary information. It is possible that hand tremor in the present apparatus could reflect the speed in which subjects performed each trial such that fast movements might have produced a greater number of errors. To test this possibility, a random sample of error and time scores was selected on all trials from 20 subjects and subjected to correlational analysis. None of the results reached statistical signifi- cance, indicating that the variability in the data was primarily a reflection of the treatment conditions outlined above rather than the time each subject took to perform the task.

Certain practical considerations were main- tained in the development of the hand steadi-

at EASTERN MICHIGAN UNIV on February 19, 2015hfs.sagepub.comDownloaded from

50-February, 1972 HUMAN F A C T O R S

ness apparatus and accompanying study. Among them was whether such a task might be a reasonable selection device for dental students, and whether hand tremor would be an accurate reflection of motor fatigue during dental treatment. In addition, i t was felt that perhaps certain equipment design changes might be suggested if hand tremor appeared to be sensitive to the various conditions of the present task. While it is difficult to evaluate the utility of the hand steadiness apparatus as a selection device for dental students due to a lack of performance differences found here between dentists and nondentists, it seems reasonable to state that the accuracy of arm movements is significantly reduced when in- direct (mirror) viewing is required. As such, a feasible suggestion to dentists as well as dental equipment manufacturers might be to attempt to design equipment so that direct access to the oral cavity is possible. An additional point to be made is that whereas emphasis is currently placed on the importance of seating for dentistry, the present study found no empirical basis for atlsuming that sitting is any better than standing, a t least with respect to hand steadi- nem. It may be, then, that the manufacturer should concentrate on other aspects of the dental process.


This research was supported by Grant DII 00035-04 from the Office of Dental Health, U.S. Public Health Service and represents one of ‘several studies performed by the Biodental Engineering Research Project directed by Percy 11. IIilI, Chairman, Department of Engineering Graphics and Design, Tufts University, bled- ford, Massachusetts, 02155. The authors also wish to express their appreciation to Mr. Allan €1. Clemow and Mr. Raymond Lorion for their help in data collection.

REFERENCES Brown, J. S., Knauft, E. B., and Rosenbaum, G. The

accuracy of positioning reactions as a function of their direction and extent. American Journal of

Conigan, R. E. and Brogden, W. J. The effect of angle upon precision of linear pursuit movements.

Psychology, 1948.619167-182.

American J o u r ~ l of Psychology, 1948, 61, 502-510.

Herbert, M. J. The speed and accuracy with which six linear arm movements can be visually positioned from two different control locations, Report No. 260. U. S. Army Medical Research Laboratory, Fort Knox, Kentucky, 25, March 1957.

Kirk, R. E. Experimental design procedures for the behooioral sciences. Belmont, California: Brooks/Cole Publishing Company, 1968.

Whipple, C. M. Manual of mentol and physical tests. Baltimore: Warwick & York, 1914.

at EASTERN MICHIGAN UNIV on February 19, 2015hfs.sagepub.comDownloaded from