Background: Growing evidence suggests that prolonged uninterrupted sitting can be detrimental to health. Much sedentary behaviour research is reliant on self-reports of sitting time, and sitting-reduction interventions often focus on reducing motivation to sit. These approaches assume that people are consciously aware of their sitting time. Drawing on Action Identification Theory, this paper argues that people rarely identify the act of sitting as ‘sitting’ per se, and instead view it as an incidental component of more meaningful and purposeful typically-seated activities. Methods: Studies 1 and 2 explored whether people mentioned sitting in written descriptions of actions. Studies 3– 5 compared preferences for labelling a typically desk-based activity as ‘sitting’ versus alternative action identities. Studies 6 and 7 used card-sort tasks to indirectly assess the prioritisation of ‘sitting’ relative to other action descriptions when identifying similar actions. Results: Participants rarely spontaneously mentioned sitting when describing actions (Studies 1–2), and when assigning action labels to a seated activity, tended to offer descriptions based on higher-order goals and consequences of action, rather than sitting or other procedural elements (Studies 3–5). Participants primarily identified similarities in actions based not on sitting, but on activities performed while seated (e.g. reading; Studies 6–7). Conclusion: ‘Sitting’ is a less accessible cognitive representation of seated activities than are representations based on the purpose and implications of seated action. Findings suggest that self-report measures should focus on time spent in seated activities, rather than attempting to measure sitting time via direct recall. From an intervention perspective, findings speak to the importance of targeting behaviours that entail sitting, and of raising awareness of sitting as a potential precursor to attempting to reduce sitting time. Keywords: Sedentary behaviour, Sitting, Standing, Cognition, Action identification, Psychology, Office workers, Experimental Background many focusing on challenging motivation to sit . Prolonged sitting has been linked to adverse mental and Such research assumes that people are aware of their sit- physical health, and premature death [1–5]. This has ting, can reliably reflect on it and wish to reduce it. This spurred policy and research interest. National guidelines paper questions such assumptions. We argue that people assert the importance of limiting sitting time [6–8]. Re- mentally represent sitting not as a purposeful act, but ra- searchers have sought to describe and identify determi- ther an incidental by-product of pursuing more mean- nants of sitting patterns, often based on self-reported ingful actions. sitting [9, 10]. Various interventions have been trialled, Action Identification Theory  describes how people assign identities to behaviours: reading this paper, for example, could be identified as ‘reading a research * Correspondence: email@example.com report’ or ‘moving my eyes’. Action identities are hier- Department of Psychology, Institute of Psychiatry, Psychology and archically structured: higher levels capture general Neuroscience, King’s College London, Guy’s Campus, London SE1 1UL, UK Full list of author information is available at the end of the article © The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Gardner et al. International Journal of Behavioral Nutrition and Physical Activity (2019) 16:85 Page 2 of 11 understandings of why an action is done, and lower autobiographical events. To identify whether people levels represent more concrete details of how action is were inattentive to sitting per se, or to postural informa- done. Levels of representation are relative; ‘reading a re- tion more broadly, we also recorded mentions of stand- search report’ is a higher-level identity than ‘moving my ing. We predicted that: eyes’, but lower-level than ‘learning about new research’. Action identities generate and sustain action, and Hypothesis 1: When recalling previous events, people higher-level identities, which reveal the purpose and will not mention sitting or standing. likely consequences of action, tend to dominate because they offer optimal guides for action . Assuming the We assumed a failure to mention sitting or standing present paper is being read to attain a desired goal (e.g. would reflect that ‘sitting’ and ‘standing’ are not domin- to learn about new research), for example, ‘reading a re- ant representations, so are less cognitively accessible and search report’ allows more effective implementation and unlikely to be elicited . Support for our hypothesis monitoring of progress towards the goal than does ‘mov- could alternatively reflect a failure to encode postural in- ing my eyes’. People thus mentally ‘chunk’ instrumental formation into memory. To explore this, we assessed the actions into higher-order action units. Action identifica- clarity of each recollection, and of specific aspects, in- tion is a dynamic process, and people adopt lower-level cluding whether the participant was sitting or standing identities where pursuit of the higher-level action is dis- and location, others present, time, and clothes worn. rupted. For example, if the reader drops her glasses, a The latter was included because we expected that, like more procedurally-focused identity (‘picking up my posture, clothes worn would not be central to the mean- glasses’) will temporarily dominate until recovery of and ing of the event. reversion to the higher-level identity (‘reading a research report’) is achieved. Method Our thesis is that people rarely conceive of sitting as Participants and procedure ‘sitting’, and instead assign higher-order action labels Adults recruited from a UK-based online recruitment that convey the meaning of activities performed while platform  were paid £1 (~US$1.30) to complete a sitting. That is, sitting is ‘invisible’; people seldom view task involving describing autobiographical events. Eligi- ‘sitting’ as the purpose, nor do they value it as an out- bility criteria were age (≥18 years), and English as first come, of seated activity. If asked what they were doing, language. Of 178 adults that began the task, 28 did not an office worker sitting at their desk would likely offer a complete it, and four were ineligible. The final sample description oriented in work-related goal pursuit (e.g. comprised 146 participants (117 [80%] female; age 18- ‘working’), to which sitting is usually subservient . 70y, mean = 34). People adopt lower-level identities for difficult or novel actions , but sitting is a simple and familiar act . Documenting how people think about sitting could offer Data collection and analysis new avenues for understanding and reducing sedentary Participants were asked to describe in as much detail as behaviour. possible “three recent experiences … things that you have Action representations can be elicited in various ways, done, or have happened to you, within the last three such as eliciting descriptions of actions, directly asses- months”. For each event, they also reported recency sing preferences for one action identity over others, or (today, yesterday, last week, a few weeks ago, last month, indirectly assessing the prioritisation of identities in cat- a few months ago), and clarity of the overall memory egorisation tasks [12, 14, 15]. We used various methods (‘my memory of this event involves [1 = little or no, 7 = to assess how people mentally represent sitting. Studies a lot of] visual detail’) and of discrete aspects (‘my mem- 1 and 2 explored whether people mentioned sitting ory for [the time of day when / how many people were when freely describing their own and others’ actions. present when / the location in which / the clothes I was Studies 3–5 descriptively analysed preferences for label- wearing when / whether I was standing or sitting down ling sitting as ‘sitting’ versus other action identities. when] this event took place’ [1 = is vague, 7 = is clear/ Studies 6 and 7 used card-sort tasks to document the ac- distinct]). cessibility of ‘sitting’, relative to alternative action labels, Verbs within each description were categorised into when categorising similar actions. sitting (e.g., ‘sit’, ‘seated’, ‘sat’), standing (e.g., ‘stand’, ‘standing’, ‘stood’) or discrete other actions (e.g. Study 1 ‘shopping’). We descriptively analysed the frequency This study investigated the accessibility of ‘sitting’ as an of verbs and each action. Clarity ratings were nested action representation by documenting the frequency within person and within events, so intraclass correla- with which people mentioned sitting when recalling tions were calculated for descriptive purposes, and Gardner et al. International Journal of Behavioral Nutrition and Physical Activity (2019) 16:85 Page 3 of 11 comparisons tested using planned comparisons in salience of posture in participants’ descriptions of photos multilevel models . of others performing activities while sitting or standing. Following Study 1, we predicted: Results and discussion In total, 2445 verbs were coded within the 438 descrip- Hypothesis 2: When describing others’ actions, people tions (mean 5.58 verbs per description, SD = 4.19), which will be less likely to mention sitting or standing than were of events that typically occurred ‘a few weeks ago’ other action identities. or more recently (n = 186). The verb frequency intraclass correlation (.64) indicated within-participant consistency in the number of actions recounted per event. Method Contrary to Hypothesis 1, sitting was mentioned in 23 Participants and procedure (5%) and standing in 5 descriptions (1%). The most com- Participants were directed to an online task via an email mon ‘other’ activity (‘going to [a destination]’) featured circular to staff and students in a UK university, an ad- in 178 descriptions (41%). vertisement on an undergraduate research participation Recollected events were rated as visually clear (mini- pool system, and a social media post. Undergraduates re- mum per-event mean = 4.36, SD = 1.38; see Additional ceived course credits, but no other incentives were pro- file 1). Intraclass correlations revealed that 33% of vari- vided. Eligibility criteria were as in Study 1. ability in overall visual clarity was explained at the event The task featured sixteen photos of individuals sitting level and 25% at the person level, suggesting consistency or standing, for each of which participants had to write a in clarity within participants and across events. Across description. Three illustrative examples, spanning differ- events, participants reported clear memory of the time, ent identification levels, were provided (e.g. ‘people location and others present (minimum per-event watching a live band’, ‘people putting their hands in the mean = 6.07, SD = 1.40), and clearer recollection of air’). Of 122 people who began the task, 19 were ineli- whether sitting or standing than of the overall event gible, and 33 did not complete, leaving a final sample of (z = 14.12, p < .01), or the clothes worn (z = 9.64, 70 (58 [83%] female; aged 18–57 years, mean = 27). p < .01). There were no differences between perceived clarity of sitting or standing and event location (z = − Materials 1.73, p = .48), presence of others (z = 1.24, p = .80), or Photos were selected from a public photo-sharing web- time (z = − 2.55, p =.11). site  where they met the following criteria: colour, Participants rarely mentioned sitting or standing no obvious editing, depicting at least one person with when describing past events, though when prompted, open eyes and unambiguously standing or sitting while remembered whether they were sitting or standing. performing another activity. Photos depicting famous This suggests that, while people committed postural people, babies, more than 10 people, or nudity were ex- information to memory, descriptions focused on the cluded. Eight photos depicted sitting and eight standing, purpose or consequences of actions – for example, to of which three showed ‘active standing’ (i.e., walking, arrive at a destination – rather than subservient ele- running). mentssuch asposture. Although participants expressed confidence in recal- Data collection and analysis ling posture, other elements of the event may have pro- Participants were instructed to “describe the action or vided cues to posture; visiting the cinema, for example, actions that you see in [each] image, in no more than entails sitting. Autobiographical events are inherently one sentence”. Verbs were coded and categorised as in idiosyncratic, and variation in duration and number of Study 1. To ensure the most salient action identity was discrete actions within such events may have influenced elicited, where descriptions exceeded one sentence, only the likelihood of posture being recounted. We were un- the first sentence was coded. Factually inaccurate or able to verify the accuracy of recollections; autobio- verb-free descriptions (e.g. ‘man in a park’) were ex- graphical memories are often inaccurate or incomplete cluded. Paired-samples t-tests compared, for each photo, . Study 2 explored mentions of sitting in descriptions the proportion of responses citing sitting or standing of a standardised set of stimuli. versus the most commonly-elicited ‘other’ activity. Study 2 Results and discussion Action identification principles apply equally to one’s Both sitting (t = 4.65, p = .002) and standing (t = 5.03, own and others’ behaviours; people tend to identify p = .002) were less commonly elicited than other action others’ actions by inferring the actors’ thoughts, emo- identities, supporting Hypothesis 2. For seven of eight tions and intentions . This study assessed the sitting photos, other identities dominated: the Gardner et al. International Journal of Behavioral Nutrition and Physical Activity (2019) 16:85 Page 4 of 11 proportion of citations of sitting across photos ranged Results and discussion from 3 to 37%, whereas alternative action citations Most participants (168; 75%) preferred ‘getting work ranged from 26 to 100%. Where sitting was mentioned, done’ over ‘sitting down’ as a label for ‘using the office it was commonly alongside other action identities (e.g. computer’ (χ = 54.76, p < .001), and 177 (79%) preferred ‘sitting and drawing’). Similarly, standing was less fre- ‘starting the day’ over ‘standing up’ as a label for ‘getting quently cited for seven of eight photos, and was typically out of bed’ (χ = 73.96, p < .001). Hypothesis 3 was sup- described with other actions (e.g. ‘standing watching ported. This cannot be attributed to postural informa- football’). These findings support our assumption that tion lacking salience, because sitting and standing were ‘sitting’ and ‘standing’ are not salient representations. explicitly offered as possible action labels. Studies 1 and 2 assessed mental representations Higher-level action identities were preferred for the within freely-generated descriptions. Participants may three BIF items (maximum p = .002). Our findings imply have excluded postural information because they that, compared to alternative action identities, ‘sitting’ perceived it to have insufficient communicative was deemed less applicable because it does not con- value; nobody mentioned ‘living’ or ‘breathing’ in ei- vey the purpose or implications of seated activities. ther study, for example. Studies 3–5 circumvented We did not however directly assess the meaningful- this problem by examining explicit preferences for ness of sitting. Study 4 tested whether sitting and posture-based action representations. standing are indeed perceived as mechanistic rather than purposeful action identities. Study 3 This study directly assessed preferences for labelling ac- Study 4 tivities as ‘sitting’ (or ‘standing’) versus an alternative ac- This study investigated the level of abstraction at which tion identity. We predicted that: people portray ‘sitting’ relative to alternative action la- bels for an archetypal sedentary behaviour (i.e. desk- Hypothesis 3: When describing seated or standing based activity ). Following Wegner and colleagues activities, people will be unlikely to assign an action , participants rated the extent to which action iden- identity based on sitting or standing. tities described desk-based activity, and factor analysis identified response clusters corresponding to identifica- tion levels. We argue that sitting is primarily seen as in- strumental in pursuing more meaningful actions, so Method predicted that: Participants, procedure and materials 225 adults recruited via a US-based online research re- Hypothesis 4: People will portray the act of ‘sitting’ at cruitment platform  completed an online task. Age the same level of abstraction as other procedural (≥18 years) was the only eligibility criterion. Due to re- action identities. searcher error, no demographic information was col- lected. The survey comprised five items, based on the Behavioural Identification Form (BIF), a validated action identification tool . Each item presented an action Method (e.g. ‘locking the front door’), and two valid alternative Participants, procedure and materials action descriptions, one based on more concrete ele- Office workers were recruited via a UK-based online re- ments (i.e. lower-level identity; e.g. ‘putting the key in cruitment platform . Eligibility criteria were age the lock’), and one addressing the presumed purpose of (≥18 years) and, to ensure personal relevance of the focal action (higher-level; ‘securing the house’ ). Partici- action, working full-time in office-based employment. pants were required to ‘choose the identification that best Participants completed an online task in which they describes the behaviour for you’. Three filler items were were rated how well each of 20 activities (including ‘sit- randomly selected from the BIF. The two focal items re- ting’) accurately described ‘what you personally do at a lated to sitting (‘using the office computer’; response op- desk’ (1 = not well at all, 7 = extremely well ). The 20 tions: ‘sitting down’ vs ‘getting work done’) and standing activities were independently generated by a separate (‘getting out of bed’; ‘standing up’ vs ‘starting the day’). panel of eight university-based office workers. Of 150 people who completed the survey, 11 did not Analysis meet eligibility criteria. Our final sample comprised 139 Chi-squared goodness-of-fit tests assessed preferred de- participants (81 [58%] female; age range 22–71 years, scriptions for each action. mean = 39). Gardner et al. International Journal of Behavioral Nutrition and Physical Activity (2019) 16:85 Page 5 of 11 Analysis ‘doing my job’). The third factor, except one item (‘moving Direct oblimin principal component analysis identified my hands’), related to economic implications of work (e.g., discrete factors underlying responses, with observed ei- ‘contributing to the economy’, ‘earning money’), and the genvalues compared to randomly-generated thresholds fourth to information processing (e.g., ‘organising infor- . Only items that loaded at ≥.40 were deemed indica- mation’). These data support Hypothesis 4 by suggesting tive of factors. Bartlett’s test of sphericity was significant that sitting was viewed as one of several procedural (i.e. (χ (190) = 1165.83, p < .001), and sampling adequacy relatively low-level) desk-based activities, distinct from was high (KMO = .82), indicating acceptability of higher-level identities that convey the broader social or or- analysis. ganisational functions of such activities. Although mean scores indicated that ‘sitting’ was viewed Results and discussion as descriptive of desk-based activity, the task did not expli- All 20 action were typically viewed as descriptive (range citly assess the priority of ‘sitting’ as an action representa- of means: 4.12–6.24), and ‘sitting’ was particularly de- tion relative to the nineteen alternative action identities. scriptive (mean = 5.88, SD = 1.40). Four factors were ex- Study 5 explored preferences for identifying desk-based tracted (see Table 1). The first, which explained most action as ‘sitting’ compared to alternative action identities. variance in responses, appeared to capture procedural actions (e.g. ‘typing’, ‘looking at the monitor’, ‘pressing Study 5 buttons’), and included ‘sitting’. The second factor re- Study 3 documented, in a binary choice task, preferences lated to meeting work responsibilities (e.g. ‘working’, for describing desk-based activity using labels reflecting Table 1 Study 4: Principal component analysis of possible descriptors of ‘what I do at a desk’ Mean Factor loadings (≥.40) applicability Action identity Factor 1 Factor 2 Factor 3 Factor 4 rating (SD) Looking at the monitor 5.76 (1.36) .78 Sitting 5.88 (1.40) .68 Pressing buttons 4.97 (1.81) .64 Moving the cursor 5.35 (1.69) .59 Using the internet 5.29 (1.52) .54 Using my computer 6.24 (1.17) .50 Typing 5.48 (1.35) .47 Reading information 5.53 (1.27) .40 Working 6.06 (1.14) .83 Doing my job 5.98 (1.07) .83 Getting my work done 5.83 (1.04) .72 Contributing to the economy 4.12 (1.73) −.59 Moving my hands 4.91 (1.85) −.55 Honouring my employment contract 5.38 (1.62) −.48 Earning money 5.84 (1.21) −.45 Furthering my career 4.30 (1.61) −.43 Organising information 5.42 (1.45) −.74 Processing information 5.78 (1.19) −.62 Making progress 5.16 (1.34) −.42 Earning a living 5.96 (1.15) Eigenvalue 6.13 2.60 1.58 1.38 % variance explained 30.64 13.01 7.91 6.88 Inter-factor correlations: Factor 1 .09 −.30 −.33 Factor 2 −.26 −.23 Factor 3 .28 Gardner et al. International Journal of Behavioral Nutrition and Physical Activity (2019) 16:85 Page 6 of 11 work-related goals (i.e. ‘getting work done’), rather than Results and discussion as ‘sitting’, which was represented in Study 4 as a mech- The activity typically ranked as most descriptive was anistic action. Study 5 extended these findings by exam- ‘working’ (mean 3.78, median 3; Table 2), and rank- ining the prioritisation of ‘sitting’ versus multiple ings differed across items (p < .001). ‘Sitting’ was alternative actions, drawn from the four factors extracted assigned a low rank (mean 6.72, median 8), and was in Study 4. Office workers ranked action identities ac- viewed as less descriptive than ‘working’ (mean 3.78, cording to perceived descriptiveness of desk-based activ- median 3; T = 1.01, p < .001), but not ‘organising in- ity. We predicted that: formation’ (mean 5.91, median 6; T = 0.34, p = .13) nor ‘honouring my employment contract’ (mean 7.25, Hypothesis 5: People will rate ‘sitting’ as less median 8; T = − 0.24, p = .70). Hypothesis 5 was par- descriptive than identities that focus on the purpose tially supported. and consequences of desk-based activity. Taken together with Studies 3 and 4, findings in- dicate that ‘sitting’, while seen as highly applicable to desk-based activity, lacks priority as a representa- tion of such activity. Preferred action identities re- Method lated to work responsibilities (e.g. ‘working’, ‘doing Participants, procedure and materials my job’). One-hundred and forty-nine office workers (77 [52%] fe- Studies 3–5 directly assessed action representations male, aged 18–68 years, mean = 38) were recruited via a and may have primed responses; participants may not UK-based online recruitment site . Eligibility criteria have identified desk-based activity as ‘sitting’, or indeed were age (≥18 years), and employed full- or part-time in assigned any other lower-level identity, had these iden- professional, managerial or administrative roles. Partici- tities not been made salient by data collection materials. pants were paid £0.85 (~US$1.10) on completing an on- Studies 6 and 7 adopted indirect methods to elicit action line task in which they ranked 10 randomly-ordered representations. action identities according to “how well they describe what you personally do at your desk” (1 = most, 10 = least Study 6 descriptive). Action identities were a subset of the 20 This study conceptually replicated and extended used in Study 4, capturing each of the factors extracted Studies 3 and 5 by assessing indirectly the accessibil- in Study 4 (see Table 2). ity of ‘sitting’ and ‘standing’ compared to alternative identity labels. Participants viewed three photos, two of which depicted a person sitting (or standing), and Analysis two depicted a person engaging in a presumed Friedman’s ANOVA with follow-up Wilcoxon tests higher-level action (e.g. reading) and were asked to compared mean rankings for ‘sitting’ versus identities select two photos depicting the same action. We pre- deemed most representative of the three non-procedural dicted that: factors from Study 4 (‘working’, ‘honouring my employ- ment contract’, ‘organising information’). Representative- Hypothesis 6: People will identify similarities between ness of the latter three items was based on higher actions based on common higher-level action identities loadings and descriptiveness ratings in Study 4. rather than ‘sitting’ or ‘standing’. Table 2 Study 5: Mean and median descriptiveness rankings of desk-based action identities Action identity Factor label (from Study 4) Mean rank (SD) Median rank Working Work responsibilities 3.78 (2.71) 3 Doing my job Work responsibilities 4.27 (2.77) 3 Getting my work done Work responsibilities 4.38 (2.51) 4 Using my computer Procedural activities 4.67 (2.40) 5 Processing information Information processing 5.24 (2.39) 5 Organising information Information processing 5.91 (2.47) 6 Typing Procedural activities 6.28 (2.35) 7 Looking at my monitor Procedural activities 6.50 (2.65) 7 Sitting Procedural activities 6.72 (3.63) 8 Honouring my employment contract Economic impact 7.25 (2.84) 8 Gardner et al. International Journal of Behavioral Nutrition and Physical Activity (2019) 16:85 Page 7 of 11 Method posture would only be identified in later pairings. Participants, procedure and materials Thus, we predicted that: Adults recruited via a US-based crowdsourcing website  were paid ~£0.30 (US$0.40) on completing a brief Hypothesis 7: People will (a) be unlikely to identify online task involving selecting two of three photos similarities between actions based on ‘sitting’ or depicting people engaging in the same activity. Age (≥18 ‘standing’, and (b) the first similarity identified will years) was the only eligibility criterion. Colour photo- not be based on ‘sitting’ or ‘standing’. graphs were selected from a photo-sharing website  on the basis that they depicted one adult, unambiguously standing or sitting and performing another activity. Par- Method ticipants were randomly assigned to view one of four Participants, procedure and materials sets of three photos (see Table 3). Each set featured two Adults were recruited via a UK online platform to photos depicting a person performing different higher- complete an online task that involved selecting, from a level actions (e.g. eating, smoking) in the same posture set of 12 photos, pairs depicting people doing the same (sitting or standing), and two showing people doing the action. Age (≥18 years) was the only eligibility criterion. same higher-level action in different postures. Photos were created especially for the study, using four Of 268 participants, one completed the task incor- models (two female), and were verified in a pilot study rectly, and nine chose ineligible pairings (e.g. ‘sitting of 40 participants to be affectively neutral on both and eating’, ‘standing and reading’), leaving a final valence and activation [26, 27]. Six photos depicted sit- sample of 258 (112 [43%] female; age range 18–71 ting, and six standing. Five other activities were depicted years, mean = 36). across the photos: painting, reading, talking on the phone, taking a ‘selfie’, and using a tablet computer Analysis (Additional file 1). To mask the study purpose, the fre- Chi-squared tests compared the frequency with which quency with which actions were depicted varied, with photos were paired according to posture versus higher- one action (talking on phone) featuring in only one level actions. photo, and one (painting) in four photos. Participants were asked to select a pair of photos and describe in free-text ‘the thing that both of the people in Results and discussion these two photos are doing’. After identifying five pair- Across all photo sets, participantsweremorelikely ings, participants could exit the task at any point. Partic- to perceive similarities based on higher-level actions ipants were paid £1 (~US$1.30) on completion, and to (N = 238; 92.2%) than on sitting or standing (N =20; 2 1 incentivise continuation, an additional £25 (~US$33) 7.8%; χ = 184.20, p < .001). Thesamepattern was cash was offered to the person who identified most valid observed within each set of photos (all p’s< .001). pairings. Supporting Hypothesis 6, participants were consist- Of 165 participants that began the survey, 14 did not ently more attentive to higher-level identities than to complete it, and three were excluded because their free- sitting or standing when identifying similarities be- text responses were not written in English. The final tween actions. The binary nature of the task, how- sample comprised 148 participants (82 [55%] female; ever, precludes assessment of the priority assigned to aged 18–77 years, mean = 32). ‘sitting’ or ‘standing’ relative to multiple alternative identities. Analysis Written descriptions were coded to identify verbs, which Study 7 were coded and categorised as in Study 1. Inaccurate de- This study extended Study 6, via a card-sort task in- scriptions, or descriptions lacking verbs and not identifi- volving identifying multiple action similarities within able as relating to action (e.g. ‘same brushes in photos of others, to assess the priority of ‘sitting’ and background’), were treated as invalid. Where multiple ‘standing’ among other action identities. Participants were incentivised to identify as many pairings of We varied the task instructions, randomly assigning participants to be people ‘doing the same thing’ as possible within a set asked to identify people doing the same ‘behaviour’ (n = 70), ‘thing’ of 12 photos with multiple similarities. We assumed (n = 79), ‘activity’ (n = 57), or ‘action’ (n = 52). The same pattern of preferences for higher-order action similarities was found across all in- that the order in which similar actions were identified struction variants (all p’s < .001). The same pattern of results also held reflected cognitive accessibility (16), such that people regardless of task completion time, comparing those who paired pho- would first attend to similarities corresponding to tos below versus at or above median completion time (7.13 s; dominant action identities, and any similarities in p’s < .001). Gardner et al. International Journal of Behavioral Nutrition and Physical Activity (2019) 16:85 Page 8 of 11 Table 3 Study 6: Frequency with which action similarities identified Set N Photo Posture Higher-level action Higher-level pairings (photos A-B), N (%) Posture-based pairings (photos A-C), N (%) χ 1 67 65 (97.0%) 2 (3.0%) 59.24*** A Standing Talking on phone B Sitting Talking on phone C Standing Painting 2 62 61 (98.4%) 1 (1.6%) 58.07*** A Sitting Playing guitar B Standing Playing guitar C Sitting Looking at laptop 3 56 52 (92.9%) 4 (7.1%) 41.14*** A Sitting Reading B Standing Reading C Sitting Taking a ‘selfie’ 4 73 60 (82.2%) 13 (17.8%) 30.26*** A Sitting Eating B Standing Eating C Sitting Smoking Total 258 238 (92.2%) 20 (7.8%) 184.20*** ***p < .001 verbs were cited, only the first was coded. Data were de- a deprioritized, less accessible representation of seated scriptively analysed. activity. Similar results were observed for representations of standing, indicating a broader deprioritization of pos- Results and discussion tural information in cognitive representations of every- Participants selected an average of 6.5 photo pairings day activities. Vallacher and Wegner  distinguish (SD = 1.47, range = 5–9, median = 6). Of 951 written de- between ‘behaviour’, which describes movement, and ‘ac- scriptions, 4 were invalid. Contrary to Hypothesis 7a, 81 tion’, which describes purposeful movement. From this participants (55%) identified similarities based on sitting perspective, our results suggest that while sitting is a be- or standing at least once. Posture was mentioned in 255 haviour of interest to researchers, it is rarely an action (27%) descriptions (137 sitting, 118 standing). from the actor’s perspective, instead being represented Descriptions of the first identified pairing showed that as a by-product of engaging in more meaningful seated only 22 (15%) referred to posture (13 sitting, 9 standing), actions. These findings may have important measure- offering some support for Hypothesis 7b. Among the 81 ment and intervention implications. participants who paired photos according to sitting or Although reliable objective measures of sitting time standing, posture was typically identified within the third are available , much empirical research into sitting pairing (mean rank = 3.19; range 1–7, median = 3). time relies on self-report , which assumes that While most people accounted for sitting or standing people can accurately reflect on sitting. Yet, in Study 1, when identifying similarities between actions, ‘sitting’ sitting was rarely cited in verbal descriptions of autobio- and ‘standing’ were of lesser priority than alternative ac- graphical events, suggesting that ‘sitting’ may not be reli- tion labels. ably encoded into or retrieved from memory. Although participants were confident in recollecting their posture General discussion during these events, we were unable to assess the accur- Any behaviour can be labelled in multiple ways; a person acy of their recollections. It is well-documented that on a bus, for example, may view her action as ‘commut- people misreport sitting time. One study, for example, ing’ or ‘sitting’ . Our participants typically repre- found that a direct self-report item (‘how long per work- sented seated episodes according to the activities ing day did you spend sitting?’) underestimated mean undertaken while sitting and viewed ‘sitting’ mostly as a monitor-assessed daily sitting time by 204mins  (see mechanistic description of how such activities are too ). While this may be partly attributable to self- enacted . Participants showed awareness of sitting, presentation biases , self-report accuracy may also be suggesting that sitting is not wholly ‘invisible’, but rather limited because people do not view episodes of seated Gardner et al. International Journal of Behavioral Nutrition and Physical Activity (2019) 16:85 Page 9 of 11 activity as ‘sitting’. Directly reporting sitting time thus seated activities serve important social functions (e.g. requires mental calculations to translate time spent in meeting friends for coffee), confer cognitive benefits (e.g. meaningful seated activities (e.g. ‘watching TV’) into sit- doing crosswords), or are instrumental to relaxation (e.g. ting time , a process susceptible to error and bias watching TV ). Similarly, office workers typically sit . This would explain why indirect measures, which to complete work tasks , and so software that deacti- infer sitting time from time spent in typically-seated ac- vates computer workstations at regular intervals to com- tivities, typically yield more accurate responses than dir- pel breaks from sitting disrupted participants’ workflow ect measures . Further work is needed to test can prompt frustration  (see too ). Environmen- whether viewing seated action as ‘sitting’ improves self- tal modification strategies, such as adjustable sit-stand report accuracy. Nonetheless, where objective sedentary workstations, which permit normally-seated tasks while behaviour assessment methods are unavailable, we rec- standing or moving, show acceptability and can reduce ommend that researchers adopt indirect self-report mea- sitting [11, 47, 48]. sures, which place the onus for sitting time estimation Limitations must be acknowledged. The demographic on the researcher rather than the participant. diversity and representativeness of users of the online For desk-based activity at least, participants viewed platforms from which we recruited has been questioned ‘sitting’, alongside ‘typing’ and ‘looking at the monitor’, . However, we have no reason to expect that demo- as a finer-grained procedural component incurred by graphics contributed to observed effects. Moreover, ac- completing work tasks. Action Identification Theory tion identification is a dynamic process, such that lower- proposes that people mentally represent actions accord- level representations such as ‘sitting’ may be prioritised ing to why and with what effect they are done, because in response to contextual changes, such as when no seat such representations (e.g. ‘working’) convey information is available. However, we would not expect ‘sitting’ to to guide goal-directed activity in a way that representa- become a dominant action representation other than in tions detailing how action is done do not . By impli- response to momentary contextual disruptions. Add- cation, our data suggest people sit not for the purpose of itionally, some people typically identify actions at finer- sitting, but because they are motivated to perform activ- grained levels of analysis rather than according to their ities that entails sitting. This qualifies research efforts broader meaning . ‘Sitting’ may therefore be a more that seek to understand motivations to sit . For ex- prominent representation for some people. However, the ample, in one study, children who expressed preferences lack of prioritisation of ‘sitting’ as an action identity ap- for seated tasks (e.g. playing video games) lost less peared to be a strong, robust effect: in Study 3, 75% of weight following a sedentary reduction intervention . participants expressed a preference for labelling seated The authors concluded that ‘the motivation to be seden- and standing activities according to higher-level mean- tary limits the effects of reducing sedentary behaviour ings rather than posture, and in Study 6, 92% identified on weight change’ (p1; emphasis added). Our results, similarities between photos based on purposeful acts ra- however, suggest that the motivation to engage in activ- ther than posture. Any effects of individual differences ities that involve sitting accounted for such effects. At- on identity preferences are likely to have been small. tempts to assess motivation to sit – rather than to Our studies assumed that people assign single-action pursue seated activities – via questionnaire methods risk labels to activities (e.g. ‘drawing’ ). Yet, in free-text capturing cognitions generated in response to question- descriptions, participants used multi-action labels (‘sit- naire items, rather than pre-existing cognitions . ting and drawing’), suggesting that they hold repre- Intervention developers should acknowledge that sit- sentations that combine multiple concurrent activities. ting per se is rarely consciously motivated. Interventions More recent theorising proposes that people store might seek to increase awareness of sitting patterns as a comprehensive representations incorporating informa- precursor to reduction. People often express surprise tion on multiple actions, alongside sensory informa- upon realizing their sitting time , suggesting they do tion, information about cognitions, affect and goals, not consciously attend to it. Raising awareness, by for and contexts [50, 51]. Nonetheless, our studies sug- example objectively monitoring and providing retro- gest that within such representations, ‘sitting’ may be spective feedback on time spent sitting, can motivate less heavily weighted or meaningful. people to reduce their sitting time [40–42]. Interventions might also reduce sitting indirectly, by targeting actions Conclusion that incur sitting. This will require acknowledging the Much research tacitly assumes that sitting is a meaning- meaning, purpose and function of seated actions, and ful action. Our studies challenge this claim; people rarely promoting sitting reduction in a way that impacts min- represented seated activity as ‘sitting’, instead viewing it imally on pursuit of such actions or desired conse- as an instrumental element of more meaningful activities quences [43, 44]. For example, for many older adults, performed while seated. Sedentary behaviour researchers Gardner et al. International Journal of Behavioral Nutrition and Physical Activity (2019) 16:85 Page 10 of 11 should recognize that sitting is often not a motivated ac- Received: 24 April 2019 Accepted: 25 September 2019 tion, but rather is incurred by and subservient to other activities. Developing acceptable and effective sitting- References reduction interventions may depend on reducing sitting 1. Biddle SJH, Bennie JA, Bauman AE, Chau JY, Dunstan D, Owen N, et al. Too in a way that respects the purpose and value that people much sitting and all-cause mortality: is there a causal link? BMC Pub Health. assigned to seated actions. 2016;16(1):635. 2. Biswas A, Oh PI, Faulkner GE, Bajaj RR, Silver MA, Mitchell MS, et al. Sedentary time and its association with risk for disease incidence, mortality, and hospitalization in adults: a systematic review and meta-analysis. Ann Supplementary information Intern Med. 2015;162(2):123–32. Supplementary information accompanies this paper at https://doi.org/10. 3. de Rezende LFM, Lopes MR, Rey-López JP, Matsudo VKR, Luiz O d. C. 1186/s12966-019-0851-0. Sedentary behavior and health outcomes: An overview of systematic reviews. PLoS ONE. 9(8):e105620. Additional file 1: Table S1. Study 1: Perceived clarity of memories of 4. Hamilton MT, Healy GN, Dunstan DW, Zderic TW, Owen N. Too little autobiographical events. exercise and too much sitting: inactivity physiology and the need for new Additional file 2: Table S2. Study 7: Description of stimuli. recommendations on sedentary behavior. Curr Cardio Risk Rep. 2008;2(4): 5. Puig-Ribera A, Martínez-Lemos I, Giné-Garriga M, González-Suárez ÁM, Bort- Abbreviations Roig J, Fortuño J, et al. Self-reported sitting time and physical activity: BIF: Behavioural Identification Form; SD: Standard deviation; TV: Television interactive associations with mental well-being and productivity in office employees. BMC Pub Health. 2015;15(1):72. 6. Piercy KL, Troiano RP, Ballard RM, Carlson SA, Fulton JE, Galuska DA, et al. Acknowledgements The physical activity guidelines for Americans. JAMA. 2018;320(19):2020–8. The authors thank Kristy-Lee Alfrey, Catherine Berkes, Charlotte Ferguson, 7. Tremblay MS, Leblanc AG, Janssen I, Kho ME, Hicks A, Murumets K, et al. and Renae Harland for assistance with data collection, and Mike Aitken Dea- Canadian sedentary behaviour guidelines for children and youth. Appl kin, Tim Rakow and Wijnand van Tilburg for study design and statistical Physiol Nutr Metab. 2011;36(1):59–64. advice. 8. UK Department of Health. Stay active: a report on physical activity for health from the four home countries’ chief medical officers. London: Authors’ contributions Department of Health; 2011. BG conceived of all studies, with input from SF, ALR and LS. BG oversaw data 9. Bennie JA, Chau JY, van der Ploeg HP, Stamatakis E, Do A, Bauman A. The collection and analysis of all studies. All authors contributed to design of one prevalence and correlates of sitting in European adults - a comparison of 32 or more studies. SKQ undertook data collection and analysis, under Eurobarometer-participating countries. Int J Behav Nutr Phys Act. 2013;10(1): supervision from BG, for Studies 3, 4, and 5. HW undertook data collection and analysis, under supervision from BG and SD, for Study 6. BG led on 10. Prapavessis H, Gaston A, DeJesus S. The theory of planned behavior as a drafting the manuscript, with individual sections contributed by SKQ and model for understanding sedentary behavior. Psych Sport Exerc. 2015;19: HW. The manuscript was subsequently iteratively refined by all authors. All 23–32. authors approved the final submission. 11. Gardner B, Smith L, Lorencatto F, Hamer M, Biddle SJH. How to reduce sitting time? A review of behaviour change strategies used in sedentary behaviour reduction interventions among adults. Health Psychol Rev. 2016; Funding 10(1):89–112. No external funding was received for this work. 12. Vallacher RR, Wegner DM. What do people think they’re doing? Action identification and human behavior. Psychol Rev. 1987;94(1):3–15. Availability of data and materials 13. Hadgraft NT, Brakenridge CL, LaMontagne AD, Fjeldsoe BS, Lynch BM, All datasets analysed within the current studies are available from the Dunstan DW, et al. Feasibility and acceptability of reducing workplace corresponding author on reasonable request. sitting time: a qualitative study with Australian office workers. BMC Pub Health. 2016;16(1):933. 14. Vallacher RR, Wegner DM, Frederick J. The presentation of self through Ethics approval and consent to participate action identification. Soc Cognition. 1987;5(3):301–22. All procedures for all studies were approved by a King’s College London 15. Wegner DM, Vallacher RR, Dizadji D. Do alcoholics know what They’re research ethics committee (Study 1: MR/17/18–249; Study 2: LRU-15/16– doing? Identifications of the act of drinking. Basic Appl Soc Psychol. 1989 2532; Study 3: MRS-17/18–8217; Studies 4 and 5: MRS-17/18–8195; Study 6: Sep 1;10(3):197–210. MR/16/17–29; Study 7: MR/18/19–10572). All participants provided fully in- 16. Higgins ET. Knowledge activation: accessibility, applicability, and salience. In: formed consent to participate. Higgins ET, Kruglanski W, editors. Social psychology: handbook of basic principles. New York, NY: The Guilford Press; 1996. p. 133–68. Consent for publication th 17. Prolific. www.prolific.ac. Accessed 24 March 2019. Not applicable. 18. Bates D, Mächler M, Bolker B, Walker S. Fitting linear mixed-effects models using lme4. J Stat Softw. 2015;67(1):1–48. Competing interests 19. Conway MA. Autobiographical memory: an introduction. Milton Keynes: The authors declare that they have no competing interests. Open University Press; 1990. 20. Kozak MN, Marsh AA, Wegner DM. What do i think you’re doing? Action Author details identification and mind attribution. J Pers Soc Psychol. 2006;90(4):543–55. 1 th Department of Psychology, Institute of Psychiatry, Psychology and 21. Flickr. www.flickr.com. Accessed 24 March 2019. th Neuroscience, King’s College London, Guy’s Campus, London SE1 1UL, UK. 22. Mechanical Turk. www.mturk.com. Accessed 24 March 2019. 2 3 Carnegie School of Sport, Leeds Beckett University, Leeds, UK. Physical 23. Vallacher RR, Wegner DM. Levels of personal agency: individual variation in Activity Research Group, Appleton Institute, School of Health, Medical, and action identification. J Pers Soc Psychol. 1989;57(4):660–71. Applied Sciences, Central Queensland University, Rockhampton, Australia. 24. Ryde GC, Brown HE, Gilson ND, Brown WJ. Are we chained to our desks? Department of Experimental Psychology, Division of Psychology of Describing desk-based sitting using a novel measure of occupational sitting. Language Sciences, University College London, London, UK. The Cambridge J Phys Act Health. 2014;11(7):1318–23. Centre for Sport and Exercise Sciences, Department of Life Sciences, Anglia 25. Horn JL. A rationale and test for the number of factors in factor analysis. Ruskin University, Cambridge, UK. Psychometrika. 1965;30:179–85. Gardner et al. International Journal of Behavioral Nutrition and Physical Activity (2019) 16:85 Page 11 of 11 26. Bradley MM, Lang P. The International Affective Picture System (IAPS) in the 48. Shrestha N, Kukkonen-Harjula KT, Verbeek JH, Ijaz S, Hermans V, Pedisic Z. study of emotion and attention. In JA Coan, JJB Allen, editors. Series in Workplace interventions for reducing sitting at work. Cochrane Database affective science. Handbook of emotion elicitation and assessment. New Syst Rev. 2018;6:CD010912. York, NY: Oxford University Press; 2007. pp29–46. 49. Peer E, Brandimarte L, Samat S, Acquisti A. Beyond the Turk: alternative 27. Morris JD. SAM: the self-assessment manikin. An efficient cross-cultural platforms for crowdsourcing behavioral research. J Exp Soc Psychol. 2017;70: measurement of emotional response. J Advertising Res. 1995;35:63–9. 153–63. 50. Barsalou LW. Simulation, situated conceptualization, and prediction. Philos 28. Wegner DM, Vallacher RR. The trouble with action. Soc Cognition. 1987;5(3): Trans Royal Soc B: Biol Sci. 2009;364(1521):1281–9. 179–90. 51. Papies EK, Best M, Gelibter E, Barsalou LW. The role of simulations in 29. Grant PM, Ryan CG, Tigbe WW, Granat MH. The validation of a novel activity consumer experiences and behavior: insights from the grounded cognition monitor in the measurement of posture and motion during everyday theory of desire. J Cons Res. 2017;2(4):402–18. activities. Brit J Sports Med. 2006;40(12):992–7. 30. Healy GN, Clark BK, Winkler EAH, Gardiner PA, Brown WJ, Matthews CE. Measurement of adults’ sedentary time in population-based studies. Am J Publisher’sNote Prev Med. 2011;41(2):216–27. Springer Nature remains neutral with regard to jurisdictional claims in 31. Gupta N, Christiansen CS, Hanisch C, Bay H, Burr H, Holtermann A. Is published maps and institutional affiliations. questionnaire-based sitting time inaccurate and can it be improved? A cross-sectional investigation using accelerometer-based sitting time. BMJ Open. 2017;7(1):e013251. 32. Chastin SFM, Culhane B, Dall PM. Comparison of self-reported measure of sitting time (IPAQ) with objective measurement (activPAL). Physiol Meas. 2014;35(11):2319–28. 33. Hardy LL, Hills AP, Timperio A, Cliff D, Lubans D, Morgan PJ, et al. A hitchhiker’s guide to assessing sedentary behaviour among young people: deciding what method to use. J Sci Med Sport. 2013;16(1):28–35. 34. Chau JY, van der Ploeg HP, Dunn S, Kurko J, Bauman AE. A tool for measuring workers’ sitting time by domain: the workforce sitting questionnaire. Br J Sports Med. 2011;45(15):1216–22. 35. Gardner B, Abraham C. What drives car use? A grounded theory analysis of commuters’ reasons for driving. Transport Res Part F: Traffic Psychol Behav. 2007;10(3):187–200. 36. Dall PM, Coulter EH, Fitzsimons CF, Skelton DA, Chastin SFM. TAxonomy of self-reported sedentary behaviour tools (TASST) framework for development, comparison and evaluation of self-report tools: content analysis and systematic review. BMJ Open. 2017;7(4):e013844. 37. Epstein LH, Roemmich JN, Cavanaugh MD, Paluch RA. The motivation to be sedentary predicts weight change when sedentary behaviors are reduced. Int J Behav Nutr Phys Act. 2011;8(1):13. 38. Ogden J. Some problems with social cognition models: a pragmatic and conceptual analysis. Health Psychol. 2003;22(4):424–8. 39. Hadgraft NT, Willenberg L, LaMontagne AD, Malkoski K, Dunstan DW, Healy GN, et al. Reducing occupational sitting: workers’ perspectives on participation in a multi-component intervention. Int J Behav Nutr Phys Act. 2017;14(1):73. 40. De Greef K, Deforche B, Tudor-Locke C, De Bourdeaudhuij I. A cognitive- behavioural pedometer-based group intervention on physical activity and sedentary behaviour in individuals with type 2 diabetes. Health Educ Res. 2010;25(5):724–36. 41. Fitzsimons CF, Kirk A, Baker G, Michie F, Kane C, Mutrie N. Using an individualised consultation and activPAL™ feedback to reduce sedentary time in older Scottish adults: results of a feasibility and pilot study. Prev Med. 2013;57(5):718–20. 42. Dewitt S, Hall J, Smith L, Buckley JP, Biddle SJH, Mansfield L, Gardner B. Office workers’ experiences of attempts to reduce sitting time: an exploratory, mixed-methods uncontrolled intervention pilot study. BMC Pub Health. 2019;19:819. 43. Palmer VJ, Gray CM, Fitzsimons CF, Mutrie N, Wyke S, Deary IJ, et al. What do older people do when sitting and why? Implications for decreasing sedentary behaviour. Gerontologist. 2018. https://doi.org/10.1093/geront/gny020. 44. Such E, Mutrie N. Using organisational cultural theory to understand workplace interventions to reduce sedentary time. Int J Health Prom Educ. 2017;55(1):18–29. 45. Cooley D, Pedersen S, Mainsbridge C. Assessment of the impact of a workplace intervention to reduce prolonged occupational sitting time. Qual Health Res. 2014;24(1):90–101. 46. Luo Y, Lee B, Wohn DY, Rebar AL, Conroy DE, Choe EK. Time for break: Understanding information workers’ sedentary behavior through a break prompting system. Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems. 2018;127. 47. Gorman E, Ashe MC, Dunstan DW, Hanson HM, Madden K, Winkler EAH, et al. Does an ‘activity-permissive’ workplace change office workers’ sitting and activity time? PLoS One. 2013;8(10):e76723.
The International Journal of Behavioral Nutrition and Physical Activity – Pubmed Central
Published: Oct 12, 2019
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera