Secondary Logo

Journal Logo

The effect of comments about shoe construction on impact forces during walking


Medicine & Science in Sports & Exercise: July 2000 - Volume 32 - Issue 7 - p 1258-1264
Applied Sciences: Biodynamics

McCAW, S. T., M. E. HEIL, and J. HAMILL. The effect of comments about shoe construction on impact forces during walking. Med. Sci. Sports Exerc., Vol. 32, No. 7, pp. 1258–1264, 2000. Comparisons of ground reaction forces (GRF) during gait are not typically conducted with blinding of the varied shoe characteristic, raising concerns related to the existence of a placebo effect, or a subject response based on a perceived expectation of change.

Purpose To determine whether investigator comments on shoe construction influenced GRF measures and ratings of shoe cushioning during walking.

Methods 19 female college students volunteered for a study presented as a test of a new shoe material. The study involved walking (2.5 m·s1) under three shoe conditions (SC). Shoes in SC2 and SC3 were harder than those in SC1, but shoes in SC1 and SC3 looked similar. Subjects in a mislead group (N = 9) were told SC1 and SC3 were baseline measures in a standard shoe, with SC2 misleadingly described as a shoe constructed of unique new material. A control group performed the same three conditions without investigator description. GRF data were collected for 10 trials for each subject in each condition, and subjects rated the perceived cushioning of each shoe. GRF data and perceived cushioning scores were analyzed using mixed-factor (group by shoe) ANOVA.

Results A significant shoe main effect was found for loading rate. Post hoc tests identified the difference between SC1 and both SC2 and SC3. The group main effect was not significant for any dependent variable.

Conclusions Results suggest that, as a group, GRF data and cushioning scores are not affected by investigator comments that do not match shoe construction characteristics. However, ratings of perceived shoe cushioning by some individual subjects reflected investigator comments and not the vertical GRF variables.

Illinois State University, Normal, IL; and University of Massachusetts, Amherst, MA

Submitted for publication January 1999.

Accepted for publication October 1999.

Address for correspondence: Steven T. McCaw, Ph.D., FACSM, 5120 Illinois State University, Normal, IL 61790-5120; E-mail:

Design of modern athletic footwear addresses the needs for a shoe to be comfortable while providing energy absorption and rearfoot stability (2). The latter factors, energy absorption and rearfoot stability, have been identified as important in the prevention and progression of a variety of gait-related injuries to the lower extremity and low back. Clinicians dealing with individuals with degenerative joint disease frequently advise such patients to continue or to adopt a walking program to maintain cardiovascular fitness while recommending use of shoes with superior cushioning effects to reduce the impulsive loading that has been associated with the onset and progression of joint degeneration (14,19). However, selection of the appropriate shoe is usually left to the patient, who may depend on advice from a variety of sources during selection.

Footwear designers incorporate a variety of midsole materials and mechanical systems to cushion shock while controlling rearfoot motion. To market shoes, advertising copy and the claims of salespeople often extol the benefits of the materials and design incorporated in a shoe. This marketing strategy is common, as evident from a perusal of a fitness magazine, in spite of a paucity of unbiased biomechanical testing of the validity of many of these claims, and a lack of understanding of how such claims may influence the gait behavior of a purchaser of a pair of shoes.

The role of subject perception to expectations of energy absorption by the shoes has often been overlooked when evaluating energy absorbing characteristics of shoes. Shoe design and material construction is frequently evaluated by comparing an experimental “energy-absorbing” shoe with a standard lab shoe. In addition, the instructions provided to the subject may follow the format of “this shoe is designed to absorb more of the shock at contact.” Shoes designed for energy absorption may include an insert at the heel, consisting of various materials, or simply a different hardness of the material used to construct the midsole. These modifications are visible and/or perceivable by the subjects.

The role expectancy plays in the effectiveness of clinical treatments is often examined through administration of a placebo. Placebos are used to alter response expectancies and may affect a change in those to whom they were administered (11). The essential ingredient to a placebo treatment is the belief by the subject that they are receiving treatment and the anticipation that some effect will occur. By use of words or actions, a subject must be led to believe that the placebo will cause the desired effect. Knowledge of shoe construction characteristics or the specific instructions given to a subject may cause adoption of a gait pattern in accordance with the expectation of the energy absorption provided by the shoe. A form of placebo testing was used by Robbins and Waked (15) to alter expectations of a cushioning material. To evaluate the influence of investigator comments regarding cushioning material, subjects stepped barefoot onto a force platform covered by identical shoe sole material that had been modified to appear different. Subjects had been provided with false advertising copy that provided different descriptions of the materials supposedly used in each condition. The advertising copy ranged from a warning that the material did not provide cushioning to the use of graphs, tables, and athlete endorsement suggesting superior impact absorption. The results indicated the impact ground reaction force varied as a function of the advertising message. When subjects were informed that a particular surface provided additional cushioning, impact ground reaction force data were higher (121% of body weight) than when subjects were provided with a warning message (110% of body weight). The results were interpreted as suggestive of subjects moderating impact in accordance with the expected cushioning of the material. That is, subjects were less inclined to use a landing strategy that would reduce impact force if they had been told that cushioning would be provided by the surface material. The study raises the question of how subjects would respond if the cushioning characteristics of a shoe, rather than a landing surface, were altered, because a shoe represents a more personalized aspect of the foot/ground interface.

In laboratory testing of shoes, subjects have been asked to quantify their perception of shoe cushioning. Using a 15-point rating scale, Hennig et al. (8) had subjects classify running shoes from very, very hard to very, very soft. Ground reaction force data were collected after subjects quantified shoe cushioning. The authors reported that the peak vertical force at footstrike was lower when subjects wore harder shoes. The authors suggest that subjects may alter the kinematic pattern of gait to reduce the peak ground reaction force encountered during the support phase in such a way as to reduce both peak and impulse values. However, an increase in loading rate magnitude was associated with a rating of increased shoe hardness. In accordance with expectation theory, making judgments about a shoe before testing may have led subjects to unknowingly alter lower limb mechanics to adjust for perceived cushioning.

In a study of perception of shoe cushioning when running, Hennig and Milani (7) reported that most subjects have the ability to judge impact severity in experimental situations, but it is unknown to what extent these perceptions are affected by a subject’s expectation of the testing environment. Robbins and Waked (15) demonstrated that subjects may be deceived into accepting higher forces during landing if provided with misleading information regarding the energy absorbing characteristics of the landing. Because of research suggesting a possible cause of lower impact peaks in harder shoes to be an intrinsic avoidance mechanism (1), it may be theorized based on expectancy theory that influencing someone to believe that one shoe can absorb more energy than another may result in different ground reaction force values. Contrary and sometimes confusing findings in the literature emphasize the need for further study in the area of perception and energy absorption. The purpose of this study was, therefore, to determine the effect of investigator comments regarding shoe construction on the ground reaction force measured during walking.

Back to Top | Article Outline



Nineteen apparently healthy female college-age students were recruited as subjects in this study. The only requirements for participation were that subject was a regular participant in physical activity, had not suffered an injury to the lower extremity in the past year, and wore a size 7 shoe. Before participation, each subject read and signed an informed consent in accordance with university procedures and the American College of Sports Medicine. Subjects were randomly assigned to either the control (N = 10) or mislead (N = 9) groups. Unpaired t-tests (α = 0.05) indicated no significant differences in basic demographics between the two groups (Table 1).

Table 1

Table 1

Back to Top | Article Outline


Three pairs of walking shoes were specially manufactured for this study. The materials and appearance of the uppers of all three pairs were the same, and the midsole of each pair was made from ethylene-vinyl acetate (EVA). However, to provide for deception of the subjects, the midsoles of two pairs were white but differed in midsole hardness, whereas the midsole of the third pair had been dyed blue and was similar in midsole hardness to one of the shoes with the white midsole. When the shoes were provided to our laboratory, the investigators were unaware of the midsole characteristics of the different shoes and did not know which of the white midsole shoes matched with the blue insole shoe.

One pair of white shoes was randomly assigned to be shoe 1 and the other was assigned to be shoe 3. After completion of the ground reaction force data collection, the shoes were tested to determine their impact characteristics. The three pairs of shoes were tested using an Exeter Research Impact Testing System. The test included 25 preimpacts with a mass of 8 kg dropped from a height of 0.05 m followed by 10 impact trials under the same conditions. The measured parameters were peak acceleration and the time to peak acceleration. Results of this testing are presented in Figure 1. Higher peak acceleration values and shorter times to peak acceleration reflect reduced cushioning of a material. Figure 1 indicates that shoe 1 exhibited greater cushioning compared with shoe 2 and shoe 3, but that all shoes provided cushioning typical of commercially available walking shoes. The acceleration values, ranging from 8.67 to 9.64 g, were lower than the values of 16, 50, and 28 g reported by Cavanagh et al. (4) for crepe-soled shoes, rubber army boot soles, and leather soles of street shoes, respectively, lower than the value range of 9.9–15.5 g for running shoes reported by Hennig and Milani (7) but higher than the value range of 7.58–8.49 g for running shoes reported by Milani et al. (12).

Figure 1

Figure 1

Back to Top | Article Outline

Deception Message

Subjects in the control group were not provided with any details regarding shoe construction and were simply informed that they would rate the cushioning provided by three pairs of walking shoes. Subjects in the mislead group were provided with written and oral messages intended to suggest that the second pair of shoes to be worn were constructed with special energy absorbing material (Appendix A). The different midsole appearance of the pairs of shoes was highlighted to the subject while the statement was being read and discussed.

Back to Top | Article Outline


Force platform.

Ground reaction force data were collected using a force platform (Advanced Medical Technologies, Inc., Model OR-6–5-6, Newton, MA) sampling at 960 Hz. The force platform was installed flush with the floor in the middle of a 10-m walkway. Walking speed was controlled by two sets of photocells located 1 m before and after the force platform. Subjects were provided multiple practice trials to become familiar with the data collection protocol. A fast walking speed of 2.5 m·s1 (5.6 mph) was used for all trials in recognition of the relatively high fitness of the subjects. Only walking trials within 5% of the target speed through all timing intervals were accepted for analysis. As a further control for normal gait, visual inspection was used to detect overt targeting to the force platform. Ten trials were collected in each pair of shoes. The 2- to 3-min break required to change shoes was deemed adequate to negate any possible fatigue effects, which were minimal considering the walking pace used in the study and the regular activity level of the subjects.

A set of four variables describing the initial impact phase of gait were measured from the vertical ground reaction force curve of each trial (Fig. 2) using custom software. A threshold of 10 N in the vertical ground reaction force was used to identify the onset of foot contact. Loading rate was calculated as the rate of force increase between 25 N and 0.75·body weight, a modification of the calculation proposed for the analysis of ground reaction force in running by Munro et al. (13). Ground reaction force data were converted to units of BW.

Figure 2

Figure 2

Back to Top | Article Outline

Perception scale.

The 15 point perception scale proposed by Hennig et al. (8) was used to quantify a subject’s perception of shoe cushioning. The scale had been adapted from the Borg scale (3) commonly used to measure perceived exertion. Subjects were instructed to rate the perceived cushioning of each pair of shoes following completion of the 10 walking trials.

Back to Top | Article Outline

Statistical Analysis

For each of the four ground reaction force variables, a ten-trial mean value was calculated for each subject in each of the three shoe conditions. These values, along with the subject’s rating of perceived shoe cushioning, were analyzed using a two-way mixed factor ANOVA, with group representing a between subjects factor and shoe representing a within-subjects factor. When appropriate, the Tukey honestly significant difference was used as the post hoc analysis to identify the source of a significant F-ratio.

Back to Top | Article Outline


Presence of Impact Force

The impact peak occurring in the first 20 ms of ground contact was not consistently present across all subjects, nor within individual subjects across all conditions (Table 2). The presence of the impact peak reflects the characteristics of the shoe construction, gait speed, the walking pattern of the subject, and the sampling rate used to collect ground reaction force data. Keller et al. (10) suggest that the inconsistency in foot strike pattern is overtly manifest at about 2.5 m·s1, the speed used in the present study. Because of the inconsistency in its presence, no significant F-ratios for interaction or main effects were obtained for any of the dependent variables quantifying the impact peak. This finding supports the statement by Munro et al. (13) that the vertical ground reaction force during the initial period of ground contact is best described using loading rate, a variable that can be calculated regardless of the presence or absence of the impact peak.

Table 2

Table 2

Back to Top | Article Outline

Loading Rate

Descriptive statistics of loading rate are presented graphically in Figure 3. The results of the mixed factor ANOVA indicated no significant interaction (F2,34 = 1.13, P = 0.334) and no significant group main effect (F1,17 = 0.0001, P = 0.991). However, loading rate was the only dependent variable to demonstrate a significant shoe main effect (F2,34 = 116.00, P < 0.0005). The post hoc Tukey test revealed that the loading rates for both harder shoes (shoe 2: 35.2 ± 9.1 BW·s1; shoe 3: 35.1 ± 8.6 BW·s1) were significantly greater than the softer shoe (shoe 1: 30.9 ± 7.6 BW·s1). The difference in loading rate between the hard and soft shoe conditions represents an effect size of 0.51, which represents a moderate effect (17).

Figure 3

Figure 3

Back to Top | Article Outline

Perceived shoe cushioning

Descriptive statistics of the perceived shoe cushioning are presented graphically in Figure 4. The results of the mixed factor ANOVA indicated no significant interaction (F2,34 = 1.33, P = 0.278), no significant group main effect (F1,17 = 1.23, P = 0.283), and no significant shoe main effect (F2,34 = 3.16, P < 0.055). The mean perceived cushioning scores for the control group are in accordance with the mean values for the loading rate. The control group exhibited a higher mean perception of cushioning score for the soft shoe (shoe 1: 9.9 ± 1.4) than for either of the hard shoes (shoe 2: 8.0 ± 2.1; shoe 3: 7.6 ± 2.2). These results contrast with the pattern of reported mean perception of cushioning values for the mislead group. These subjects reported a mean perception of cushioning score of 9.6 ± 2.7 for shoe 1, the soft shoe, and 8.6 ± 2.6 for shoe 3, a harder shoe. The mislead group reported the highest mean perception of cushioning score of 9.8 ± 2.5 for shoe 2, a shoe misleadingly described to them as a softer shoe but actually harder than shoe 1 and similar to shoe 3.

Figure 4

Figure 4

Back to Top | Article Outline


The role of investigator comments regarding shoe characteristics on perception of loads imposed during the initial period of the stance phase of walking was the focus of this research. In this study of ground reaction force variables and perception of shoe cushioning during walking, one group of subjects were intentionally mislead regarding the energy absorbing characteristics of the shoes while the other group was given no information on shoe characteristics. Only after ground reaction force data were collected from each pair of shoes did each subject provide a rating of perceived cushioning.

The loading rate values for all three shoes were greater than the value of 14.6 ± 3.7 BW·s1 reported by Keller et al. (10) for subjects walking at 2.5 m·s1 in running shoes. The higher values in our study could reflect differences in shoe construction, but is more likely attributable to differences in calculation of loading rate. Keller et al. (10) calculated loading rate by dividing the first maximum force by the time interval between initial foot contact and the occurrence of the vertical thrust maximum force. This method of calculation would tend to underestimate the loading rate because it ignores the nonlinear portion of the rise in vertical force from initial contact to the vertical thrust maximum force. Loading rate in the present study was calculated according to a procedure modified from Munro et al. (13) and provides a better reflection of the peak rate of force application during gait.

No significant differences were identified between groups in the loading rate values. Both groups walked in the identical shoes, with the conditions differing only in the investigator comments regarding shoe construction that were provided to the subjects. For both groups, the values of loading rate were significantly higher in the hard shoes compared to the soft shoes. Although subjects may adapt the walking pattern in response to expected cushioning provided by shoes (15), loading rate values in running have shown a consistent trend to increase with increased midsole hardness (8,12). Similar results in this study of walking suggest that shoe midsole characteristics are the primary determinant of the rate of loading imposed during the initial period of ground contact. Coupled with our finding that the presence of the impact peak was inconsistent among the subjects while walking in any of the pairs of shoes, these results support the statement by Munro et al. (13) that the ground reaction forces during the initial period of ground contact are most appropriately described using loading rate.

The implications of higher loading rate during gait are not entirely clear. Hennig and Lafortune (6) reported a high positive relationship between loading rate and peak tibial acceleration when running. In the literature on shoes and gait, an inherent assumption is that high levels of force and loading rate are deleterious (15,18). However, this simplistic approach overlooks the beneficial effects attributed to loading rate during gait relative to increases in bone mineral density (BMD) and the prevention of osteoporosis (5). Impact loads imposed during gait have been implicated in the onset and progression of degenerative joint disease and other overuse injuries (16); the same loading patterns have been identified as a beneficial stimulus for increasing BMD. This contradiction indicates a need for increased research regarding the clinical implications of impact loading during gait.

It was not unexpected that there was no significant main effect of group for the ratings of perceived shoe cushioning. It has been shown that, as a group, subjects are able to perceive differences in shoe cushioning when similar instructions are presented to all (7,8,12). In the current study, subjects in the mislead group were given comments by the investigators’ attributing superior energy absorption characteristics to the shoe midsole of one of the pairs of shoes. Analysis of the ground reaction force data identified no significant differences between the mislead and control groups for the loading rate. Milani et al. (12) reported a statistically significant relationship between perception of cushioning and loading rate. According to this relationship, the rating of perceived shoe cushioning should have been similar for both groups in this study if the rating of shoe cushioning was based entirely on perception of loading rate. However, the results of our study indicated that subjects tended to provided a higher rating of perceived shoe cushioning after having been told by the investigators that the shoe midsole incorporated a material designed to provide greater cushioning. Robbins and Waked (15) reported that subjects may be deceived by misleading claims regarding cushioning characteristics of the foot-ground interface, but these authors did not present or discuss individual subject data. The lack of a significant group effect in our study is not surprising because it would be unexpected for all subjects to consistently provide a rating of shoe cushioning that reflects the misleadingly described shoe characteristics while discounting personal perception of shoe cushioning. The standard deviation associated with the perception score mean values of the mislead group were higher than those for the control group, reflecting the inconsistent rating of shoe perception provided by members of the mislead group. Perception of cushioning scores for individual subjects in each group are presented in Table 3. These data show that some subjects, such as subjects 4, 5, and 6 in the mislead group, provided cushioning ratings for shoe 2 that were considerably higher than the ratings for either shoe 1 or shoe 3. The outcome of this study lends support to the quote attributed to Abraham Lincoln, 16th president of the United States, “You can fool some of the people all of the time, and all of the people some of the time, but you cannot fool all of the people all of the time” (9).

Table 3

Table 3

Several directions for future research are suggested by the outcome of this study. There is a need to measure the kinematics of gait, including rearfoot motion and segment orientation at touchdown, to determine whether subjects did alter gait patterns according to expectations of shoe cushioning. A longitudinal study is needed to identify if subjects will continue to report perception values in accordance with investigator comments, or if they will adjust values of perceived cushioning to concur with measures of loading rate. There is a need to identify the characteristics of subjects who are gullible to the comments of the investigator rather than the mechanical variables, because such individuals would be at the whim of improper advice when selecting shoes for participation in physical activity. Finally, of greatest importance, is a need to replicate the present study utilizing individuals with degenerative joint disease. Such individuals are commonly prescribed “cushioned shoes” by clinicians to reduce the load imposed during gait in an attempt to retard the progression of joint degeneration. A tendency to believe the comments related to the cushioning of prescribed shoes rather than the actual cushioning provided by the shoes could expose these individuals to a more rapid degeneration of the loaded joints.

In conclusion, the results of this study indicate that, as a group, ground reaction force data and measures of perceived shoe cushioning collected during level walking are not affected by information provided by the investigator that does not match characteristics of shoe construction. However, the measures of perceived shoe cushioning of some individual subjects were influenced according to the investigator comments and not the vertical ground reaction force variables.

Completion of this study was assisted by an ACSM Visiting Scholar Award to Steve McCaw to work with Joe Hamill at the University of Massachusetts in 1992, and by an Illinois State University Research Grant to STM (1993). The study was presented as a poster at the 1998 Meeting of the American College of Sports Medicine in Orlando.

Back to Top | Article Outline


1. Barnes, R. A., and P. D. Smith. The role of footwear in minimizing lower limb injury. J. Sports Sci. 12:341–353, 1993.
2. Bates, B. T. Testing and evaluation of running shoes. In: Biomechanics IX-B, D. A. Winter, R. W. Norman, R. P. Wells, K. C. Hayes, and A. E. Patla (Eds.). Champaign, IL: Human Kinetics, 1983, pp. 128–132.
3. Borg, G. Perceived exertion as an indicator of somatic stress. Scand. J. Rehabil. Med. 2:92–98, 1970.
4. Cavanagh, P. R., K. R. Williams, and T. E. Clarke. A comparison of ground reaction forces during walking barefoot and in shoes. In:Biomechanics VII-B, A. Morecki, K. Fidelus, K. Kedzior, and A. Wit (Eds.). Baltimore: University Park Press, pp. 151–156, 1981.
5. Gooch, K. J., and C. J. Tennant. Mechanical forces: Their effects on cells and tissues. Berlin: Springer, 1997, p. 55.
6. Hennig, E. M., and M. A. Lafortune. Relationships between ground reaction force and tibial bone acceleration parameters. Int. J. Sport Biomechanics 7:303–309, 1991.
7. Hennig, E. M., and T. L. Milani. In-shoe pressure distribution for running in various types of footwear. J. Appl. Biomech. 11:299–310, 1995.
8. Hennig, E. M., G. A. Valiant, and Q. Liu. Biomechanical variables and the perception of cushioning for running in various types of footwear. J. Appl. Biomech. 12:143–150, 1996.
9. Hertz, E. Lincoln Talks. New York: Viking Press, 1939, p. 138.
10. Keller, T. S., A. M. Weisberger, J. L. Ray, S. S. Hasan, R. G. Shiavi, and D. M. Spengler. Relationship between vertical ground reaction force and speed during walking, slow jogging, and running. Clin. Biomech. 11:253–259, 1996.
11. Kirsch, I. Changing Expectations. Pacific Grove, CA: Brooks/Cole, 1990, p. 43.
12. Milani, T. L., E. M. Hennig, and M. A. Lafortune. Perceptual and biomechanical variables for running in identical shoe constructions with varying midsole hardness. Clin. Biomech. 12:294–300, 1997.
13. Munro, C. F., D. I. Miller, and A. J. Fuglevand. Ground reaction forces in running: a reexamination. J. Biomech. 20:147–155, 1987.
14. Radin, E. L., I. L. Paul, and R. M. Rose. Role of mechanical factors in pathogenesis of primary osteoarthritis. Lancet 1 (7749):519–522, 1972.
15. Robbins, S., and E. Waked. Hazard of deceptive advertising of athletic footwear. Br. J. Sport. Med. 31:299–303, 1997.
16. Schwellnus, M. P., G. Jordaan, and T. D. Noakes. Prevention of common overuse injuries by the use of shock absorbing insoles: a prospective study. Am. J. Sports Med. 18:636–641, 1990.
17. Thomas, J. R., W. Salazar, and D. M. Landers. What is missing in p < .05 effect size. Res. Q. Exerc. Sport 62:344–348, 1991.
18. Wosk, J., and A. S. Voloshin. Low back pain: conservative treatment with artificial shock absorbers. Arch. Phys. Med. Rehabil. 66:145–148, 1985.
19. Yang, K. H., C. L. Riegger, V. L. Kish, J. J. O’Connor, and E. L. Radin. Gait pattern predictors of tibiofemoral osteoarthrosis. ASB Proceedings118–119, 1989.
Back to Top | Article Outline

Appendix A

The following message was read to each subject in the mislead group while highlighting specific features of the shoe:

You will be participating in a study to evaluate the effectiveness of a shoe designed with a revolutionary new shock absorbing material. The material is designed to protect the body from the potentially harmful shock forces associated with walking and running. In this study, you will be asked to walk across a force platform, which will measure the amount of force generated during walking. You will do this in three separate pairs of shoes. The first shoe is a standard walking shoe constructed with no special shock absorbing characteristics. The second shoe is constructed with an innovative material designed to absorb more of the shock. The third shoe is constructed exactly as the first shoe and will serve as a reference to measure the return of the shock forces back to normal levels. After each pair of shoes is tested you will be asked to rate the comfort of the shoe on a scale of 1–15.



© 2000 Lippincott Williams & Wilkins, Inc.