Printer Friendly

The effect of different response formats on ratings of exerciser stereotypes.

In social science research, questionnaires can be used to describe respondents' personality characteristics and assessment of variables related to a topic of interest. However, response mistakes and bias can occur when using this method (Piedmont, McCrae, Riemann, & Angleitner, 2000). As Schwarz (1999) noted, questionnaire characteristics including question wording, context, and format can influence responses. Item content and format can influence the psychometric properties of a scale, such as construct and criterion validity (Rauthmann, 2011). Changing the format (e.g., question wording) can change the responses obtained. For example, Hall and Roggenbuck (2002) found that respondents were more likely to choose higher ratings when presented with a closed format (i.e., scale with anchored points) as opposed to a semiopen format (i.e., fill in the blank), and the scores obtained from these formats were different. Similarly, Courneya, James, Rhodes, and Blanchard (2003) found different effects on self-reported exercise frequency depending on type of response scale. When participants responded on a scale in which the response options ranged from less than 15 times in the past 30 days to more than 30 times, they reported three times more bouts of activity than when the response options ranged from zero to more than 15 times (Courneya et al., 2003). It is, therefore, likely that the framing of a scale results in some contextualization of response so that, for example, a person who thinks they exercise a "medium" amount will be more likely to pick the middle point on a scale.

It has also been noted that respondents' reports of their attitudes are influenced by a variety of variables beyond the attitudes themselves (Donnelly, Vaske, Whittaker, & Shelby, 2000). These authors reported that response formats had effects on prevalence norms, which they define as "the proportion of individuals in a population who can articulate a norm in a given evaluation context" (p. 403). In addition, Courneya, Conner, and Rhodes (2006) found that when compared to the standard 7-point scale, scores using four experimental scales (i.e., extreme 7- and 11-point, and phenomenal 7- and 11-point scales) were more variable for measures of the theory of planned behavior despite the five scales having similar levels of predictive theory validity.

A stereotype is defined as the beliefs about the characteristics, attributes, and behaviors of a member of an outgroup (Hilton & von Hippel, 1996). The fitness center environment and the physical activity domain are full of stereotypes (Harrison, 2001). In a pilot study to the current research at least eight exerciser stereotypes were found. A considerable number of researchers have examined exercise-based stereotypes and it has been found that people who are described as physically active and fit are rated more positively than those who are described as being irregular exercisers and unfit (Martin Ginis, Sinden, & Fleming, 2000). Others have found that exercisers, nonexerciser intenders, and nonexerciser nonintenders rated exercising targets more favorably compared with nonexercising targets (Rodgers, Hall, Wilson, & Berry, 2009). It has also been shown that active participants were implicitly biased for exercisers and against sedentary individuals (Berry, Spence, & Clark, 2011). Thus, the activity level of respondents may influence questionnaire ratings of exerciser stereotypes.

Rodgers and colleagues (2009) found that gender was not related to ratings of exercisers and nonexercisers. However, Leary (1992) noted the influence of gender on impressions formed. He suggested that impressions formed of the same target may differ by societal subgroup. For example, whereas a man may view another man taking an aerobics class in a negative way, a woman may view the same man in a positive way. Martin Ginis and Leary (2006) reported that exercise status was more influential on women's impressions of another woman than it was on men's impressions of the same woman. Shields, Brawley, and Martin Ginis (2007) also showed that the impact of exercise status on impressions formed of the exercising individual was moderated by the gender of the judges, with men rating an exercising target more positively than did women.

It is evident that question response formats may influence questionnaire ratings and the type of response format used can increase or decrease the validity and reliability of questionnaires. However, few researchers have examined the influence of gender or other moderating variables on responses to various questionnaire formats. In the present study to assess exercise stereotypes, we unintentionally made an error in a questionnaire and then took advantage of this to study the effects of using two different response formats for the same question. We examined the effects of different response formats on ratings of exercise stereotypes and also examined differences between gender and activity level. Based on previous research, it was hypothesized that the ratings on the two response formats of definitely would not like to do this/definitely would like to do this (NL anchor) vs. definitely false/definitely true (FT anchor) would be different between men and women and between active and inactive individuals. Because these two response formats have not been explored together in previous research, we did not develop a directional hypothesis.



Participants were 203 Canadians (101 male, 85 female, and 17 unidentified sex) with a mean age of 26.52 (SD = 8.52) years who volunteered to participate in this study. Participants included 103 university students recruited from two first-year health classes, one first-year mathematics class, and one first-year sociology class; the rest were exercisers recruited from nine different fitness centers.

Materials and Procedure

The questionnaire was developed for this research and included 17 questions measuring attitudes, subjective and descriptive norms, self-efficacy, perceived behavioral control, past behavior, behavioral willingness, and behavioral intentions with respect to eight exercise stereotypes. When we were composing the questionnaire we made an error, in that one question was asked twice in each exerciser stereotype questionnaire. The question appeared in the behavioral willingness section, and again in the behavioral intentions section, but was associated with a different response format in each. Thus, respondents were asked to rate the same question with two different response formats.

The repeated question was "To what extent would you like to exercise in a gym solely comprising [stereotype]?" Both response formats consisted of a 7-point Likert scale. However, in the first presentation, the scale anchors ranged from 1 = definitely would not like to do this to 7 = definitely would like to do this (NL anchors). This anchor was used in previous studies to assess behavioral expectations (Gibbons, Gerrard, Blanton, & Russell, 1998; Thornton, Gibbons, & Gerrard, 2002). In the second instance, the scale anchors ranged from 1 = definitely false to 7 = definitely true (FT anchors). This anchor was suggested by Fishbein and Ajzen (2010) to assess perceived behavioral control.

Exerciser Stereotypes. Each exercise stereotype was determined based on its descriptive characteristics obtained from pilot research with 209 undergraduate students who were asked to list as many types of exercisers as they could think of with five descriptive words for each. By examining the descriptive words associated with each of these exercisers, eight commonly mentioned exercisers were identified (see Table 1). For example, the athlete stereotype was described as "Alex is a competitive athlete and is very focused and determined when training. Alex is motivated, fit, and healthy".

Activity Levels. Activity levels were measured using the Godin Leisure-Time Exercise Questionnaire (GLETQ; Godin & Shepard, 1997). Individuals answered three questions regarding how many times over the previous week they had engaged in strenuous, moderate, and mild physical activity for at least 15 minutes in their spare time. These values were multiplied by their estimated value (i.e., nine, five, and three respectively) in metabolic equivalent (MET). MET scores were then used to, classify participants as either active or inactive. Male participants who reported participating in 38 METs a week or more and females who reported at least 35 METs a week were classified as active (Garcia Bengoechea, Spence, & McGannon, 2005). The GLETQ is a valid and reliable measure of physical activity (Gionet & Godin, 1987; Godin & Shephard, 1985; Jacobs, Ainsworth, Hartman, & Leon, 1993), with a demonstrated one-month test-retest reliability of 0.62 and concurrent validity coefficients of 0.32 with an objective indicator (accelerometer), 0.56 with maximum oxygen consumption (V[O.sub.2]Max), and 0.43 with percentage body fat (Jacobs et al., 1993).

Procedure. The details of the study were explained to participants, who provided informed consent prior to participation. Participants completed a demographic information survey and completed eight brief question sets about eight different exercise stereotypes, in random order.

Data Analysis

Paired samples t tests were used to compare scores between NL and FT categories for each exerciser stereotype. In order to assess the effects of gender (2 levels) and activity level (2 levels) on responses to the two options (NL or FT) of 8 exercise stereotypes, multivariate analysis of variance (MANOVA) was used. Since the correlations between NL and FT formats were highly positive (r = .88), a composite score (an average of NL and FT scores) was created for use in the MANOVA to avoid problems of multicollinearity (Tabachnick & Fidell, 2007). Significant interaction effects were followed up with independent samples t tests. All analyses were performed using SPSS version 18.0 for Windows.

NL and FT Anchors

Paired t tests demonstrated the significant difference of mean scores between NL and FT formats for judgmental young women and overweight exerciser stereotypes (see Table 2).

Gender and Activity Levels

The MANOVA showed that there was a significant interaction effect of gender by activity levels for the runner stereotype F(1, 23.158) = 6.954,p < .05); but not for other exercise stereotypes (all, p > .05; see Table 3). The Box's M test was not significant F(108, 31235.73) = 1.044, p = .357).

As can be seen in Table 3 there were significant effects of gender for rating athletes, elderly exercisers, judgmental young women, overweight exercisers, and yoga practitioners (all p < .05); but not for jocks or weightlifters (all p > .05). Specifically, males gave significantly higher ratings than females for athletes and judgmental young women. However, males gave significantly lower ratings than females for elderly exercisers, overweight exercisers, and yoga practitioners. The main effect of activity level was not significant.

Independent samples t tests were used to follow up the interaction effect of gender by activity levels for the runner stereotype. The test results showed that active females' scores were higher than active males' scores (t = 3.599, p < .01) but there was no difference between inactive females and inactive males (p > .05). Active females had higher scores than inactive females (t = 2.192, p < .05), but active and inactive males had similar scores.


Our aim in this study was to examine the effects of different response formats on ratings of the likelihood of exercising, based on eight different exercise stereotypes. The findings suggest that response format may influence ratings of exercise stereotypes. Gender and activity levels were also found to affect the response formats for rating for the stereotypes.

First, the results showed effects of the response formats on ratings of judgmental young women and overweight exerciser stereotypes. This finding is similar to that of previous researchers, who have found that there are response format effects on prevalence norms (Donnelly et al., 2000), effects of different response scales on self-reported exercise behavior (Courneya et al., 2003), and measurement scale effects in the exercise domain (Courneya et al., 2006). In our study we found that respondents recorded lower scores on the NL scales compared to the FT scales when asked whether they would like to exercise in a gymnasium with a clientele composed solely of the two stereotype groups of judgmental young women or overweight exercisers. We consider it is of interest that the effect of the response format on the judgmental young women and overweight exercisers was more significant for these stereotype groups than it was for the other groups. This seems to offer some evidence that, although the two scales were generating different response patterns, these might be because of bona fide conceptual differences in the anchors. It appears, for example, that not liking the idea of being near some stereotypes is the same as not wanting to exercise with individuals representing the stereotype, whereas for other stereotypes these two constructs can be separated. Thus, the effects of not liking the idea of being near a particular group might not necessarily translate into not being willing to be in the same exercise space as that group. For example, the jock stereotype received the lowest ratings on NL and also on FT, indicating that the respondents did not want to be around jocks. However, whereas the respondents' scores were toward the "not like" end of the NL scale for judgmental young women and the overweight exercisers, their responses also indicated that they did not mind being in the same environment as these two stereotype groups through their higher "true" (FT) ratings. From a motivational standpoint then, the FT rating scale might be more meaningful than the NL in understanding the motivational influence of the stereotypes on exercise behavior.

In any research context, it behooves the researcher to determine-following our example-not only whether respondents do not like some people whom they perceive as belonging to a particular group or stereotype, or some contexts or some activities, as well as whether that dislike is actually related to whether or not the respondents will engage with the people, the context, or the activity. The differential pattern of dissociation of NL and FT suggest they were conceptually distinct to our respondents, but that distinction did not have any motivational effects for the majority of stereotypes assessed here.

Second, our results showed gender effects on stereotype ratings and for the rating scales. This result illustrates that men's and women's perceptions about exercisers are different. This finding is in line with that of Peterson (2003), who reported that gender tends to influence the choice of exercise activities. Gender had influence on impressions formed (Leary, 1992; Martin Ginis & Leary, 2006; Shields et al., 2007). Constantinou, Manson, and Silverman (2009) reported that perceptions about a safe environment and a safe activity were the main factors that impacted girls' participation in, and attitude towards, physical activity and sports. This relates to our finding that women had lower scores than men for exercising with athletes and judgmental young women; but they had higher scores than men for exercising with the stereotype groups of yoga practitioners, overweight exercisers, and elderly exercisers. There are some important gender differences in people's willingness to exercise with various exercise stereotypes. Generally, our results suggested that women were more discriminating in their responses to the scale format and it may be a useful line of research to determine whether women, in general, are more responsive to wording differences in questions and scale anchors.

Our last finding relates to the effects of activity levels on rating scores to exercise with runners but not for other exerciser stereotypes. Although active and inactive respondents had similar scores for a number of the stereotype groups, in our study we found that, compared with inactive females, scores recorded by active females with regard to exercising with the runner stereotype group were higher. This result was similar to that of a previous study in which it was found that exercisers and nonexercisers rated various exercise stereotypes similarly (Martin Ginis, Latimer, & Jung, 2003). These authors also reported that both exercisers and nonexercisers view exercisers more positively than nonexercisers (Martin Ginis et al., 2000; Martin Ginis & Leary, 2006; Rodgers et al., 2009).

A limitation in our study was the self-report method of data collection for physical activity level, which was dichotomized into active and inactive groups. Further, participants were recruited from university classes and fitness centers, limiting generalizability. It may be that groups selected on a different basis, such as culture, weight status, or socioeconomic status, would respond differently to our questionnaires. Therefore, this research should be replicated with other groups. Finally, in this study we took advantage of an inadvertent error, rather than undertaking a planned comparison of different response formats. Further research is needed in which this topic is purposefully examined.

In conclusion, our findings in this study highlight the need for very careful consideration of the response options used in questionnaires. Given that there were significant differences between ratings on two response options according to gender and activity group it does appear that scale anchors can have different connotations for respondents. Furthermore, an analysis of the extant literature might reveal a need to take into consideration the response scales used, even when the items are identical, to help account for any disparity or equivocality in current understandings, particularly of the motivational influence of being around various types of exercisers. In general, if something is reported as "significant", the perception is that a true difference exists. Thus, if the response option in a questionnaire has an influence on the significance of findings, the results of research could - and should - be called into question.



Berry, T. R., Spence, J. C., & Clark, M. E. (2011). Exercise is in! Implicit exercise and sedentary-lifestyle bias held by in-groups. Journal of Applied Social Psychology, 41, 2985-2998.

Constantinou, P., Manson, M., & Silverman, S. (2009). Female students' perceptions about gender-role stereotypes and their influence on attitude toward physical education. Physical Educator, 66, 85-96.

Courneya, K. S., Conner, M., & Rhodes, R. E. (2006). Effects of different measurement scales on the variability and predictive validity of the "two-component" model of the theory of planned behavior in the exercise domain. Psychology & Health, 21, 557-570.

Courneya, K. S., Jones, L. W., Rhodes, R. E., & Blanchard, C. M. (2003). Effect of response scales on self-reported exercise frequency. American Journal of Health Behavior, 27, 613-622. http://

Donnelly, M. P., Vaske, J. J., Whittaker, D., & Shelby, B. (2000). Toward an understanding of norm prevalence: A comparative analysis of 20 years of research. Environmental Management, 25, 403-414.

Fishbein, M., & Ajzen, I. (2010). Predicting and changing behavior: The reasoned action approach. New York: Psychology Press.

Garcia Bengoechea, E., Spence, J. C., & McGannon, K. (2005). Gender differences in perceived environmental correlates of physical activity. International Journal of Behavioral Nutrition and Physical Activity, 2, 2-12.

Gibbons, F. X., Gerrard, M., Blanton, H., & Russell, D. W. (1998). Reasoned action and social reaction: Willingness and intentions as independent predictors of health risk. Journal of Personality and Social Psychology, 74, 1164-1180.

Gionet, N. J., & Godin, G. (1989). Self-reported exercise behavior of employees: A validity study. Journal of Occupational Medicine, 31, 969-973.

Godin, G., & Shephard, R. J. (1985). A simple method to assess exercise behavior in the community. Canadian Journal of Applied Sport Sciences, 10, 141-146.

Godin, G., & Shephard, R. J. (1997). Godin Leisure-Time Exercise Questionnaire. Medicine and Science in Sports and Exercise, 29, 36-38.

Hall, T. E., & Roggenbuck, J. W. (2002). Response format effects in questions about norms: implications for the reliability and validity of the normative approach. Leisure Sciences: An Interdisciplinary Journal, 24, 325-337.

Harrison, L., Jr. (2001). Understanding the influences of stereotypes: Implications for the African American in sport and physical activity. Quest, 53, 97-114.

Hilton, J. L., & von Hippel, W. (1996). Stereotypes. Annual Review of Psychology, 47, 237-271.

Jacobs, D. R., Jr., Ainsworth, B. E., Hartman, T. J., & Leon, A. S. (1993). A simultaneous evaluation of 10 commonly used physical activity questionnaires. Medicine and Science in Sports and Exercise, 25, 81-91.

Leary, M. R. (1992). Self-presentational processes in exercise and sport. Journal of Sport and Exercise Psychology, 14, 339-351.

Martin Ginis, K. A., Latimer, A. E., & Jung, M. E. (2003). No pain no gain? Examining the generalizability of the exerciser stereotype to moderately active and excessively active targets. Social Behavior and Personality: An international journal, 31, 283-290.

Martin Ginis, K. A., & Leary, M. R. (2006). Single, physically active, female: The effects of information about exercise participation and body weight on perceptions of young women. Social Behavior and Personality: An international journal, 34, 979-990.

Martin Ginis, K. A., Sinden, A. R., & Fleming, J. C. (2000). Inactivity may be hazardous to your image: The effects of exercise participation on impression formation. Journal of Sport & Exercise Psychology, 22, 283-291.

Peterson, J. A. (2004). Ten ways gender can affect exercise attitudes and performance. ACSM's Health & Fitness Journal, 8, 44.

Piedmont, R. L., McCrae, R. R., Riemann, R., & Angleitner, A. (2000). On the invalidity of validity scales: Evidence from self-reports and observer ratings in volunteer samples. Journal of Personality and Social Psychology, 78, 582-593.

Rauthmann, J. F. (2011). Not only item content but also item format is important: Taxonomizing item format approaches. Social Behavior and Personality: An international journal, 39, 119-128.

Rodgers, W. M., Hall, C. R., Wilson, P. M., & Berry, T. R. (2009). Do nonexercisers also share the positive exerciser stereotype? An elicitation and comparison of beliefs about exercisers. Journal of Sport & Exercise Psychology, 31, 3-17.

Schwarz, N. (1999). Self-reports: How the questions shape the answers. American Psychology, 54, 93-105.

Shields, C. A., Brawley, L. R., & Martin Ginis, K. A. (2007). Interactive effects of exercise status and observer gender on the impressions formed of men. Sex Roles, 56, 231-237.

Tabachnick, B. G., & Fidell, L. S. (2007). Using multivariate statistics (5th ed.). Needham Heights, MA: Allyn & Bacon.

Thornton, B., Gibbons, F. X., & Gerrard, M. (2002). Risk perception and prototype perception: Independent processes predicting risk behavior. Personality and Social Psychology Bulletin, 28, 986-999.


University of Alberta

Sonthaya Sriramatr, Tanya R. Berry, Wendy Rodgers, and Sean Stolp, Faculty of Physical Education and Recreation, University of Alberta.

Correspondence concerning this article should be addressed to: Tanya R. Berry, Faculty of Physical Education and Recreation, University of Alberta, E-488 Van Vliet Centre, Edmonton AB, T6G 2H9, Canada. Email:
Table 1. Exercise Stereotypes and Descriptive Characteristics

Stereotypes            Descriptive characteristics

Athlete                Fit, focused, determined,
                       competitive, healthy, motivated
Elderly exerciser      Healthy, happy, fit
Judgmental             No sweating, judgmental of
  young women          other people, wearing tight
                       clothes and makeup
Jock                   Male, competitive, show-off
Overweight exerciser   Unhealthy, self-conscious, overweight
Runner                 Fit, lean, having endurance, healthy
Weightlifter           Bulky, intimidating, male, muscular
Yoga practitioner      Relaxed, flexible, spiritual, female

Table 2. Paired t Test Between NL and FT Anchors
for Each Exerciser Stereotype

                                  Compared means

Stereotypes       N        NL            FT          t       p
                         M (SD)        M (SD)

Athlete          191   4.26 (2.07)   4.29 (2.11)   -.503    .616
Elderly          190   2.64 (1.71)   2.69 (1.69)   -.800    .425
JYW              193   2.36 (1.68)   2.65 (1.81)   -3.414   .001
Jock             187   2.68 (1.95)   2.79 (1.95)   -1.585   .115
Overweight       190   2.84 (1.77)   3.03 (1.93)   -2.198   .029
Runner           194   3.72 (1.94)   3.75 (1.67)   -.449    .193
Weightlifter     196   2.74 (1.79)   2.77 (1.83)   -.470    .654
Yoga             189   3.85 (2.10)   3.91 (2.05)   -1.088   .278

Note: JYW = Judgmental young women.

Table 3. MANOVA Results for Gender by Activity for Eight
Exerciser Stereotypes

                      Active       Inactive        Total
                      M (SD)        M (SD)        M (SD)

Athlete             4.58 (2.02)   4.68 (1.98)   4.61 (1.99)
Elderly             2.15 (1.50)   2.40 (1.40)   2.24 (1.47)
JYW                 2.84 (2.00)   2.67 (1.57)   2.78 (1.86)
Jock                2.76 (1.96)   3.12 (2.05)   2.88 (1.99)
Overweight          2.41 (1.51)   2.80 (1.72)   2.54 (1.59)
Runner              3.23 (1.87)   3.72 (1.70)   3.39 (1.82)
Weightlifter        2.91 (1.65)   2.82 (1.81)   2.88 (1.69)
Yoga practitioner   3.36 (1.96)   3.07 (1.77)   3.26 (1.90)

                      Active       Inactive        Total
                      M (SD)        M (SD)        M (SD)

Athlete             4.29 (2.05)   3.48 (2.08)   3.98 (2.09)
Elderly             3.21 (1.79)   2.70 (1.46)   3.02 (1.68)
JYW                 2.07 (1.37)   2.09 (1.27)   2.08 (1.33)
Jock                2.69 (1.86)   2.39 (1.62)   2.57 (1.77)
Overweight          3.52 (1.96)   3.17 (1.78)   3.39 (1.89)
Runner              4.50 (1.94)   3.48 (1.69)   4.12 (1.91)
Weightlifter        2.79 (1.77)   2.04 (1.47)   2.51 (1.69)
Yoga practitioner   4.38 (2.10)   4.00 (1.87)   4.24 (2.02)

                         p             p             p

                     Gender X       Gender       Activity
                     activity                     levels

Athlete                .176          .028          .299
Elderly                .147          .009          .617
JYW                    .717          .015          .789
Jock                   .300          .204          .932
Overweight             .197          .011          .959
Runner                 .014          .090          .383
Weightlifter           .244          .107          .130
Yoga practitioner      .902          .003          .300

Note: JYW = Judgmental young women.
COPYRIGHT 2012 Scientific Journal Publishers, Ltd.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2012 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Sriramatr, Sonthaya; Berry, Tanya R.; Rodgers, Wendy; Stolp, Sean
Publication:Social Behavior and Personality: An International Journal
Article Type:Report
Geographic Code:1CANA
Date:Nov 1, 2012
Previous Article:The effects of time perspective and salience of possible monetary losses on intertemporal choice.
Next Article:Comparison of social anxieties among Han, Tibetan, and Muslim university students in China.

Terms of use | Privacy policy | Copyright © 2022 Farlex, Inc. | Feedback | For webmasters |