Journal Logo


Self-administered questionnaire compared with interview to assess past-year physical activity


Author Information
Medicine & Science in Sports & Exercise: June 2000 - Volume 32 - Issue 6 - p 1119-1124
  • Free


A sedentary lifestyle is recognized as an important risk factor for the development of chronic diseases such as coronary heart disease, hypertension, obesity, type II diabetes, certain types of cancer, and osteoporosis (4,21,29). A major issue in epidemiologic studies intending to assess the relationships between physical activity and disease outcomes is the choice of a method for measuring physical activity. There are many dimensions of physical activity, and each may be associated with a specific aspect of health (14). Common categories, based upon the context in which physical activity occurs, are leisure time activities (including competitive sports or exercise training) and occupational and/or household activities (12). At the lower end of the activity spectrum, television watching represents a typical sedentary behavior (5). To better delineate the associations between habitual physical activity level and health outcomes, the development of methods assessing these different dimensions of physical activity is an important research objective.

Questionnaires represent the most widely used method to assess usual physical activity in population studies, as they are generally well accepted by study participants and are easy to administer to a large number of subjects at a low cost (3,12,14,17,19,23,29,30,31). Numerous questionnaires measuring physical activity over various time periods have been developed, some of which have been tested for validity and reproducibility (17,23,30). However, few of these questionnaires are both comprehensive and easy to use in longitudinal studies, i.e., designed in a self-administered format that could be administered to large number of subjects to monitor changes in physical activity patterns over time. Moreover, to the best of our knowledge, questionnaires that have been designed or translated in French for this purpose are scarce (30).

The Modifiable Activity Questionnaire (MAQ), developed by Kriska, assesses current (past 12 months) physical activity during both work and leisure time, as well as some inactivity indices (9,10,13). Validation and reliability testing of the MAQ were previously published (9,27). An interesting feature of the MAQ is that it was designed for easy modification to maximize the ability to assess physical activity in a variety of populations (13). However, in its original version, this questionnaire was interviewer-administered (9,10,13). With the objective to apply this tool in a self-administered format to a large cohort of adults, a French version of the MAQ was designed after minor cultural adaptation. The aim of the present study was to compare physical activity and inactivity data obtained by self-administration and personal interview, and to determine the concordance between these two modes of administration when using the French version of the MAQ.



Subjects were recruited among the 12,535 volunteers enrolled in the SU.VI.MAX study (“SUpplémentation en VItamines et Minéraux AntioXydants”) (6,7). The SU.VI.MAX study is a randomized double-blind, placebo-controlled, prevention trial primarily designed to test the efficacy of a daily supplementation with antioxidant vitamins and minerals, at nutritional doses, in reducing the incidence of cardiovascular diseases and cancers. Another objective of the study is to contribute to a better understanding of the relationships between nutrition and disease risk, by constituting a large database on dietary intakes, eating habits, and health events in a national French sample of adult subjects. The SU.VI.MAX study is ongoing since October 1994 with a planned follow-up of 8 yr. Participants undergo a yearly visit (every other year, either clinical examination or blood sampling). All subjects gave their informed written consent to the study which was approved by the ad hoc ethical committee, i.e., “Comité Consultatif de Protection des Personnes dans la Recherche Biomédicale” (CCPPRB n° 706 Paris-Cochin, France) (7).

All consecutive subjects examined in May-June 1997 at the Parisian center of the SU.VI.MAX study (Hôpital St. Lazare, AP-HP, Paris, France) at the time of the third yearly evaluation were asked to complete both modes of the questionnaire. Body weight and height were measured during the clinic visit. Body mass index (BMI) was calculated as body weight (kg) divided by height squared (m2). Educational level obtained from a questionnaire was coded in three categories according to the highest attained certification and degrees (elementary school, secondary school, university, or equivalent).

Questionnaire and physical activity assessment.

The most recent published version of the MAQ was used (13). Translation into French and minor adaptations to fit in the current context in terms of usual leisure and occupational activity performed by the French people were carried out by the authors. The MAQ was initially designed to be interviewer-administered, so in the present French version instructions were included in the body of the document to make it easier to understand and to fill out. The same questions were used for the interview and self-administered administration.

The MAQ was described in details elsewhere (13). As indicated in the instructions for this questionnaire, we developed a list of leisure activities common to the population in question. Compared with the leisure activities listed in the original version (13), some activities were added (motocross, Frisbee, boxing), whereas others were deleted (softball/baseball, tai chi, wood chopping, water/coal hauling, rock climbing, fencing, jumping rope, snow shoeing). Subjects had to identify all activities performed at least 10 times over the past 12 months. Then, detailed information was collected about the frequency and duration of each leisure activity. Instead of checking the months each activity was performed over the past year as in the original version (13), a single question asked the total number of months each activity was done over the past year. Hours per week for all activities performed during the past-year period were summed to obtain an indicator expressed in hours per week of leisure activity. After multiplying the number of hours per week of each leisure activity by its estimated metabolic cost (MET), an energy expenditure indicator was also obtained expressed in MET-hours per week of leisure activity. A MET is the ratio of the working metabolic rate of an activity divided by the resting metabolic rate (1). One MET represents the metabolic rate of an individual at rest (sitting quietly) and is set at 3.5 mL of oxygen consumed per kg body mass per minute, or approximately 1 kcal·kg1·h1. A 10-MET activity would require 10 times the resting metabolic rate. The average metabolic cost for each activity was drawn from lists available in the literature (1,2,16,20). Television watching was measured using a single question that read “In general, how many hours per day do you spend watching television?” (h·d1).

The assessment of occupational activity was based on the number of hours that the individual participated in physically demanding activity during an average workday, for each job held over the past-year period (13). Hours per week of light, moderate, and hard intensity activity averaged over the past year were summed to obtain an indicator expressed in hours per week of occupational activity. The number of hours in each of the three categories of occupational activity (light, moderate, and hard) were multiplied by an average group MET value (2, 4, and 7 METs, respectively) and then summed, resulting in a final occupational activity estimate expressed in MET-hours per week. Occupational activities given as examples for each of the three categories were similar to those listed in the original version of the MAQ (13).

Total physical activity over the past-year period was determined as the sum of the leisure and occupational activity indicators (expressed in hours of activity per week or MET-hours per week).

Study design.

Subjects were randomly assigned to group 1 (self-administration first) or to group 2 (interview first). Subjects in group 1 received the questionnaire at home by mail; they were requested to fill it out about 10 d before their yearly scheduled visit at the clinical center and to bring it with them on this occasion. During this visit, they were asked to answer the same questionnaire during an individual meeting with the interviewer. Subjects in group 1 were not informed initially about the interview so that they would not try to memorize their original answers. Subjects in group 2 were first administered the questionnaire through a personal interview during the yearly visit at the clinical center; after the examination, they were requested to answer the same questionnaire in the self-administered mode within about 10 d after the interview, and to return it by mail in a preposted return envelope. Upon arrival at the clinical center, subjects in group 2 were not informed about the self-administered questionnaire they received later in the day to minimize a learning effect. All interviews were carried out by the same interviewer.

Statistical analysis.

Because the distributions of variables were neither normal nor lognormal, nonparametric statistics were used. Physical activity levels determined by self- and interviewer-administered questionnaires were compared by analysis of variance on ranks in a random effect model. Three main effects, each with two levels, were considered: mode of administration (self-administration, interview), order of administration (self-administration first, interview first), and period (mode·order) (24,28). Concordance (individual subject agreement) between the two modes of administration for activity variables was assessed using a nonparametric intraclass correlation coefficient (26). Because the reporting of leisure walking for exercise was previously shown to be unreliable in several populations, data were analyzed both with and without inclusion of this activity, as recommended (8,9,13). The two items taken into account for this purpose were “walking for leisure” (i.e., at a slow pace) and “walking rapidly,” as listed in the leisure section of the questionnaire. All statistical analyses were performed using SAS software (SAS Institute, Cary, NC). Statistical significance was judged at P < 0.05.


Of 165 potential subjects examined during the study period at the clinical center, 132 subjects did answer to both modes of the questionnaire. Among these individuals, 16 subjects answered the same day to both modalities and were excluded due to a potential learning effect. For the same reason, 11 subjects for whom the date of administration of the questionnaire was missing were excluded. Twenty-one self-reported questionnaires were found incomplete. Therefore, completed questionnaires using the two methods of administration were available for 84 subjects. Characteristics of the subjects are shown in Table 1. Educational level was elementary school: 22%; secondary school or technical degree: 38%; and university or equivalent: 40%. The mean (± SD) delay between the two questionnaire administrations was 8.0 d (± 5.3) for subjects with self-administration first (group 1) and 7.8 d (± 6.3) for subjects with interview first (group 2) (NS).

Table 1
Table 1:
Characteristics of the subjects [mean ± SD (min; max)].

Table 2 shows past-year leisure, occupational, and total (leisure and occupational combined) activity levels, and hours of television watched per day, obtained with the self-administered and interviewer-administered modes of the questionnaire. Order and period effects were not significant. A significant mode effect was found only for past-year leisure physical activity [walking included: h·wk1 (P = 0.11), MET-h·wk1 (P = 0.04); walking excluded: h·wk1 (P = 0.02), MET-h·wk1 (P = 0.02)]. In general, higher levels of past-year leisure physical activity were reported with the interview compared with self-administered questionnaire (see Table 2).

Table 2
Table 2:
Past-year physical activity and hours of television watched (TV) according to mode of administration of the Modifiable Activity Questionnaire [median (Q1; Q3)].

Table 3 shows the concordance (within-subject agreement) between self-administered and interviewer-administered mode of the questionnaire for past-year leisure, occupational, and total (leisure and occupational combined) activity levels, and hours of television watched per day. Intraclass correlation coefficients were highest for past-year leisure physical activity (about 0.90 with and without inclusion of walking for exercise) and television watching (0.97). For past-year occupational and total (leisure and occupational combined) physical activity, intraclass correlation coefficients were clustered around 0.83.

Table 3
Table 3:
Concordance between self-administered and interviewer-administered mode of the Modifiable Activity Questionnaire for past-year physical activity and hours of television watched (TV).a


The potential interest of collecting accurate physical activity data in large-scale population studies using a self-administered questionnaire, including a reduction of study costs and research staff involved, led to the present study. After translation and adaptation into French, we developed a self-administered version of the MAQ, a physical activity questionnaire initially designed to be interviewer-administered. A high agreement was found between self-administered and interviewer-administered mode of the questionnaire for past-year leisure, occupational and total (leisure and occupational combined) activity levels, as well as for an index of inactivity (hours of television watched per day).

Retrospective quantitative history, on which the MAQ is based, represents the most comprehensive form of physical activity recall survey (29). Recall surveys are generally thought to influence behavior to a lesser extent and to require less effort by the respondents than either diaries or logs. In addition, surveys focusing on a 1-yr time frame are more likely to reflect usual activity patterns than those recording activities over a few days or over the past week (12). The accuracy (i.e., both reliability and validity) of the MAQ was previously reported (9,27). Relationships between test and 1–3 wk retest were used to assess reproducibility (9). In adults over 20 yr of age, Spearman correlation coefficients were 0.88–0.92 for past-year leisure physical activity, 0.88 for past-year occupational physical activity, and 0.89 for past-year total (leisure and occupational combined) physical activity. Validity was demonstrated through comparisons with activity monitors (9) and measurement of total energy expenditure by the doubly labeled water method (27). Past-year leisure physical activity was shown to be significantly related to the past-week activity measured by the Caltrac monitor (Spearman r = 0.69, P < 0.05) (9). In a study by Schulz et al. (27), past-year leisure and total physical activity were significantly related to total energy expenditure assessed by doubly-labeled water (Spearman r = 0.56 and 0.74, respectively, both P < 0.05). In the interviewer-assisted mode, the MAQ has been used in a number of studies in various populations (5,11,15,18,22,33).

As shown by the intraclass correlation coefficients reported in the present study, a high level of concordance was found between self-administered and interviewer-administered mode of the questionnaire for past-year leisure, occupational, and total (leisure and occupational combined) activity indicators, and for the number of hours of television watched per day. Interestingly, in this study, when physical activity values were weighted by their estimated metabolic cost and expressed as MET-hours per week of activity, the observed intraclass correlation coefficients remained of the same order of magnitude as when expressed as hours per week of activity. Similarly, including or excluding leisure walking for exercise did not substantially modify the level of the correlations. The high agreement between the two modes of administration of the questionnaire suggests that the self-administered version of the MAQ would be a valuable instrument for assessing activity and inactivity variables in self-administered conditions for epidemiological purposes.

There are very few studies in the literature that have examined this issue, that is comparing the self- and interviewer-administered mode of the same physical activity questionnaire. During the initial development of the Tecumseh Occupational Physical Activity Questionnaire (25), an attempt was made to design a self-administered questionnaire and to compare it versus data obtained by a personal interview (32). In that study (32), 100 men and 100 women answered and mailed a self-administered questionnaire about the type and frequency of leisure and occupational physical activities over the past year. One week after the return of the self-administered questionnaire, subjects were contacted for an interview to assess the accuracy of the original responses using a supplementary interviewer-administered physical activity questionnaire. Similar values were obtained when classifying subjects based on responses to the self-administered questionnaire or based on responses to both the self- and interviewer-administered questionnaires, indicating consistency.

In our study, results of the analysis of variance indicate a significant mode effect for the past-year leisure physical activity indicator. In general, higher levels of leisure physical activity were reported during the interview compared with the self-administered questionnaire. Because our investigation was a study of comparability and not validity, it cannot be estimated which mode of administration was the most accurate. This observation stresses the impact of the interview process on the data collected. The presence of an interviewer may provide the individual with a more structured framework for his/her response and may help the subject to give more information. However, given the high agreement between self-administered and interviewer-administered modes of the MAQ we found, it appears unlikely that ranking individuals by their physical activity level would be affected by this phenomenon.

A limitation of our study is that the sample size was reduced due to strict exclusion criteria that were applied to missing data. First, we considered the indication of the date of administration as critical to the present study. Second, the calculation of summary indices for leisure and occupational physical activity in the MAQ requires all items to be properly filled. Such reduction in sample size would not have happened if the questionnaire was interviewer-administered. In future studies, a clear statement indicating the importance of all questions to be answered should therefore be added at the beginning of the self-administered form. Another concern is that subjects in this study were participants in a clinical trial. Because typically higher educated subjects participate in such trials, they could be far from representative of the general population. However, a previous report on baseline characteristics of the participants in the SU.VI.MAX study (7) showed that the study sample was close to the national population in terms of socioeconomic status.

In conclusion, results of this study indicate that the self-administered French version of the Modifiable Activity Questionnaire is a useful tool for estimating physical activity and inactivity that could be used in large-scale population studies investigating the relationships between physical activity and health outcomes.

We thank all the volunteers who participated in this study. We also thank Myriem Aïssa for assistance in data collection.

The French translation/adaptation of the MAQ is available upon request.


1. Ainsworth, B. E., W. L. Haskell, A. S. Leon, et al. Compendium of physical activities: classification of energy costs of human physical activities. Med. Sci. Sports Exerc. 25:71–80, 1993.
2. Bannister, E. W., and S. R. Brown. The relative energy requirements of physical activity. In: Exercise Physiology, H. B. Falls (Ed.). New York: Academic Press, 1968, pp. 267–322.
3. Boisvert, P., R. A. Washburn, H. J. Montoye, and L. Leger. Mesure et évaluation de l’activité physique par questionnaire: questionnaires utilisés dans la littérature anglo-saxonne. Sci. Sports 3:245–262, 1988.
4. Bouchard, C., R. J. Shephard, and T. Stephens (Eds.). Physical Activity, Fitness and Health: International Proceedings and Consensus Statement. Champaign, IL: Human Kinetics, 1994, pp. 569–916.
5. Fitzgerald, S. J., A. M. Kriska, M. A. Pereira, and M. P. De Courten. Associations among physical activity, television watching, and obesity in adult Pima Indians. Med. Sci. Sports Exerc. 29:910–915, 1997.
6. Hercberg, S., P. Galan, P. Preziosi, et al. Background and rationale behind the SU.VI.MAX study, a prevention trial using nutritional doses of a combination of antioxidant vitamins and minerals to reduce cardiovascular diseases and cancers. Int. J. Vit. Nutr. Res. 68:3–20, 1998.
7. Hercberg, S., P. Preziosi, S. Briançon, et al. A primary prevention trial using nutritional doses of antioxidant vitamins and minerals in cardiovascular diseases and cancers in a general population: “the SU.VI.MAX study”—design, methods and participant characteristics. Control. Clin. Trials 19:336–351, 1998.
8. Kriska, A. M., R. Black-Sandler, J. A. Cauley, R. E. Laporte, D. L. Hom, and G. Pambianco. The assessment of historical physical activity and its relation to adult bone parameters. Am. J. Epidemiol. 127:1053–1063, 1988.
9. Kriska, A. M., W. C. Knowler, R. E. LaPorte, et al. Development of questionnaire to examine relationship of physical activity and diabetes in Pima Indians. Diabetes Care 13:401–411, 1990.
10. Kriska, A. M., and P. H. Bennett. An epidemiologic perspective of the relationship between physical activity and NIDDM: from activity assessment to intervention. Diabetes Metab. Rev. 8:355–372, 1992.
11. Kriska, A. M., R. E. Laporte, D. J. Pettitt, et al. The association of physical activity with obesity, fat distribution and glucose intolerance in Pima Indians. Diabetologia 36:863–869, 1993.
12. Kriska, A. M., and C. J. Caspersen. Introduction to a collection of physical activity questionnaires. In: A collection of physical activity questionnaires for health-related research, M. A. Pereira, S. J. FitzGerald, E. W. Gregg, et al. Med. Sci. Sports Exerc. 29(Suppl. 6):S5–S9, 1997.
13. Kriska, A. M. Modifiable activity questionnaire. In: A collection of physical activity questionnaires for health-related research, M. A. Pereira, S. J. FitzGerald, E. W. Gregg, et al. Med. Sci. Sports Exerc. 29(Suppl. 6):S73–S78, 1997.
14. Laporte, R. E., H. J. Montoye, and C. J. Caspersen. Assessment of physical activity in epidemiologic research: problems and prospects. Public Health Rep. 100:131–146, 1985.
15. Lee, E. T., T. K. Welty, R. Fabsitz, and L. D. Cowan. The Strong Heart Study: a study of cardiovascular disease in American Indians-design and methods. Am. J. Epidemiol. 132:1141–1155, 1990.
16. McArdle, W. D., F. I. Katch, and V. L. Katch. Exercise Physiology: Energy, Nutrition, and Human Performance. Philadelphia: Lea & Febiger, 1988, pp. 642–649.
17. Montoye, H. J., H. C. G. Kemper, W. H. M. Saris, and R. A. Washburn. Measuring Physical Activity and Energy Expenditure. Champaign, IL: Human Kinetics, 1996, pp. 42–71.
18. Narayan, K. M., M. Hoskin, D. Kozak, et al. Randomized clinical trial of lifestyle interventions in Pima Indians: a pilot study. Diabet. Med. 15:66–72, 1998.
19. Paffenbarger, R. S., S. N. Blair, I-M. Lee, and R. T. Hyde. Measurement of physical activity to assess health in free-living populations. Med. Sci. Sports Exerc. 25:60–70, 1993.
20. Passmore, R., and J. V. G. A. Durnin. Human energy expenditure. Physiol. Rev. 35:801–840, 1955.
21. Pate, R. R., M. Pratt, S. N. Blair, et al. Physical activity and public health: A recommendation from the Centers for Disease Control and Prevention and the American College of Sports Medicine. JAMA 273:402–407, 1995.
22. Pereira, M. A., A. M. Kriska, M. L. Joswiak, et al. Physical inactivity and glucose intolerance in the multi-ethnic island of Mauritius. Med. Sci. Sports Exerc. 27:541–548, 1995.
23. Pereira, M. A., S. J. Fitzgerald, E. W. Gregg, et al. A collection of physical activity questionnaires for health-related research. Med. Sci. Sports Exerc. 29(Suppl. 6):S1–S205, 1997.
24. Ratkowsky, D. A., M. A. Evans, and J. R. Alldredge. Cross-over experiments. New York: Marcel Dekker, 1993, pp. 86–87.
25. Reiff, G. G., H. J. Montoye, R. D. Remington, J. A. Napier, H. L. Metzner, and F. H. Epstein. Assessment of physical activity by questionnaire and interview. J. Sports Med. Phys. Fitness 7:135–142, 1967.
26. Rothery, P. A nonparametric measure of intraclass correlation. Biometrika 66:629–639, 1979.
27. Schulz, L. O., I. T. Harper, C. J. Smith, A. M. Kriska, and E. Ravussin. Energy intake and physical activity in Pima Indians: comparison with energy expenditure measured by doubly-labeled water. Obes. Res. 2:541–548, 1994.
28. Senn, S. Cross-over Trials in Clinical Research. New York: John Wiley & Sons, 1993, p. 63.
29. U. S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion. Physical Activity and Health: A Report of the Surgeon General. Atlanta, 1996.
30. Vuillemin, A., G. Denis, F. Guillemin, and C. Jeandel. Revue des questionnaires d’évaluation de l’activité physique. Rev. Epidemiol. Sante Publique 46:49–55, 1998.
31. Washburn, R. A., and H. J. Montoye. The assessment of physical activity by questionnaire. Am. J. Epidemiol. 123:563–576, 1986.
32. Wessel, J. A., H. J. Montoye, and H. Mitchell. Physical activity assessment: by recall record. Am. J. Public Health. 55:1430–1436, 1965.
33. Yurgalevitch, S. M., A. M. Kriska, T. K. Welty, O. Go, D. C. Robbins, and B. V. Howard. Physical activity and lipids and lipoproteins in American Indians ages 45–74. Med. Sci. Sports Exerc. 30:543–549, 1998.


© 2000 Lippincott Williams & Wilkins, Inc.