The impact of interactions before, during and after meetings on meeting effectiveness: a coordination theory perspective

Abstract

Purpose

Most studies on workplace meetings have examined them as physical gatherings but have not linked them to interactions before and after meetings. Drawing upon coordination theory, this study aims to examine the impact of interactions before, during and after meetings on meeting effectiveness.

Design/methodology/approach

A survey design was used, and regular workplace meeting attendees were recruited. A mediation model was developed to test the effect of interactions on perceived meeting effectiveness.

Findings

Interactions before meetings positively influenced attendee involvement during the meeting, and attendee involvement mediated the positive relationship between attendee interactions during the meeting and perceived meeting effectiveness. A novel finding of this study is that incorporating meeting outcomes in subsequent work positively influenced perceived meeting effectiveness because it fostered common understanding of the meeting agenda.

Originality/value

The present results link prior empirical findings on interactions before and during meetings to new predictions regarding the effect of interactions after meetings. Coordination theory expands current conceptualizations of workplace meetings by broadening the notion of meetings to cover a more extended period of interdependent interactions.

Citation

Král, P., Králová, V. and Šimáček, P. (2023), "The impact of interactions before, during and after meetings on meeting effectiveness: a coordination theory perspective", Measuring Business Excellence, Vol. 27 No. 3, pp. 403-420. https://doi.org/10.1108/MBE-08-2021-0108

Introduction

Workplace meetings are ubiquitous in organizational life. Estimates of the number of meetings reach 11 million per day in the USA alone (Allen and Rogelberg, 2013) and 55 million globally (Lehmann-Willenbrock et al., 2018). A great deal of time and organizational resources would be wasted if meetings are ineffective. Effective meetings not only enhance business performance (Kauffeld and Lehmann-Willenbrock, 2012; Rogelberg et al., 2012) but also promote employee engagement (Lehmann-Willenbrock et al., 2016; Yoerger et al., 2015), empowerment (Allen et al., 2016), well-being (Luong and Rogelberg, 2005; Rogelberg et al., 2006) and job satisfaction (Rogelberg et al., 2010). As a result, running effective meetings has attracted the attention of researchers. The literature on meeting effectiveness is centered around the process of physical gathering. Researchers have scrutinized various aspects leading to meeting effectiveness, such as meeting design characteristics (Allen et al., 2012, 2014a; Cohen et al., 2011; Leach et al., 2009), meeting leadership (Baran et al., 2012; Mroz et al., 2020; Shanock et al., 2013), communication during meetings (Allen and Rogelberg, 2013; Kauffeld and Lehmann-Willenbrock, 2012) and found numerous determinants of meeting effectiveness.

However, meetings are a complex form of workplace interactions that begin before a physical gathering and continue after it ends (Asmuß and Svennevig, 2009). The interactions before meetings (in the subsequent text, we use the terms interactions before meetings and pre-meeting interactions interchangeably) and after meetings (in the subsequent text, we use the terms interactions after meetings and post-meeting interactions interchangeably) are scattered among daily routines, and thus participants or meeting managers do no value the interactions before and after the meetings (Allen et al., 2014b; Yoerger et al., 2018). Researchers find it challenging to capture the various interactions related to meetings in a particular organization. As a result, the interactions that occur before and after physical gatherings are often neglected. This raises two concerns. First, the value of interactions before and after meetings is recognized in business settings (MacLeod, 2011), but recommended or best practices offer only prescriptions rather the explanations of how these interactions influence perceived meeting effectiveness. So far, few studies have focused on specific interactions before meetings such as scheduling (Baran et al., 2012; Eisenbart et al., 2016) and pre-meeting talk (Allen et al., 2014b; Yoerger et al., 2018). More importantly, the effect of interactions after meetings has been completely neglected. Second, the sequence of interactions is interdependent, and the perception of effectiveness evolves through a series of interactions. For example, thorough preparation before a meeting yields more developed arguments for the discussion during the meeting, these arguments foster more productive discussion, which improves the outcomes of the meeting, and eventually leads to higher perceived meeting effectiveness. The interactions affect perceived meeting effectiveness individually but also jointly. Therefore, we need empirical evidence on the interdependence of interactions that occur before, during and after meetings.

To study the effect of interactions before, during and after meetings, we pose two research questions. First, we ask how interactions before and during meetings influence attendee involvement. Second, we reveal how interactions after meetings, jointly with attendee involvement, influence perceived meeting effectiveness. The primary purpose of a meeting is coordination (Geimer et al., 2015; Luong and Rogelberg, 2005; Okhuysen and Bechky, 2009; Pezzillo Iacono et al., 2016; Quinn and Dutton, 2005), and thus, coordination theory offers an overlooked but valuable perspective to study meetings. Coordination theory acknowledges the central role of meetings (Jarzabkowski et al., 2012; Okhuysen and Bechky, 2009) and delineates the mechanisms through which social interactions lead to desired outcomes (Jarzabkowski and Seidl, 2008). It enables capturing the social interactions related to meetings, and thus, provides a holistic and simple framework explaining causal mechanisms. The research framework synthesizes prior empirical findings on interactions before and during meetings and expands current knowledge by testing newly developed hypotheses on the effect of interactions after meetings.

The findings contribute to the knowledge of workplace meetings in three ways. First, it expands the temporal framework of meetings by linking the effect of interactions before, during and after meetings. Second, the study underscores the effect of interactions after meetings, which have received minimal attention so far but have a recognizable effect on perceived meeting effectiveness. Third, the findings contribute to the meeting sensemaking literature (Jarzabkowski and Seidl, 2008; Lehmann-Willenbrock et al., 2016; Stigliani and Ravasi, 2012) by showing a strong effect of social interactions on meeting attendee involvement and perceived meeting effectiveness. The results also offer practical implications for meeting leaders.

Literature review and hypotheses development

Temporally extended view on workplace meetings

Traditionally, a meeting has been referred to as a “gathering of people for the purposes of interaction and focused communication” (Volkema and Niederman, 1995, p. 5). This definition emphasizes a gathering, which explains why most meeting studies focus on the process of physical gathering, such as good meeting practices. Only a few studies have paid attention to what occurs before and after meetings, and how these interactions affect meeting effectiveness. Pre-meeting interactions have received some empirical attention. For example, two studies proved that pre-meeting talk improves perceived meeting effectiveness. Allen et al. (2014b) found that pre-meeting talk improved perceived meeting effectiveness even after controlling for good meeting practices. Yoerger et al. (2018) have explained the causal mechanism, as they found that pre-meeting talk improves both the social and problem-solving atmosphere and, thus, influences perceived group performance. However, post-meeting interactions remain neglected. Although some authors have acknowledged the existence of a post-meeting phase (Lehmann-Willenbrock et al., 2018; Mroz et al., 2018), the exploratory study of Geimer et al. (2015) was the first to point out the interdependence of the processes that occur before, during and after meetings. Geimer et al. (2015) provided some narrative evidence and suggested broadening the notion of meetings into an extended period of interrelated activities. This study used their findings to develop an extended view of meetings and test hypotheses concerning the effect of interactions before, during and after meetings.

Coordination theory as a perspective on meetings

The coordination theory states that managing interdependencies between activities constitutes a central purpose for organizations (Crowston, 1997; Crowston et al., 2006; Malone and Crowston, 1994; Okhuysen and Bechky, 2009). The primary purpose of a meeting is coordination (Geimer et al., 2015; Okhuysen and Bechky, 2009; Pezzillo Iacono et al., 2016; Quinn and Dutton, 2005), and thus, coordination theory is an appropriate perspective to study meetings. The coordination theory is particularly useful for two reasons.

First, coordination theory provides a framework for the temporal extension of the view of the meetings. Although modern definitions of meetings characterize meetings as a complex system of workplace interactions rather than mere physical gatherings (Asmuß and Svennevig, 2009), it is difficult to determine how to capture continuous social interactions at the workplace. Coordination theory suggests that these interactions can be assessed through the occurrence or frequency of tangible interactions requiring intentional effort (Espinosa et al., 2004; Jarzabkowski et al., 2012; Joseph and Gaba, 2020). Such interactions are, for example, programming tools (e.g. schedules, plans, procedures) or purposeful communications (e.g. oral, written, formal and informal) with any link to the meeting. These meeting-related interactions provide the constructs included in the research framework.

Second, coordination theory explains three underlying mechanisms through which meeting-related interactions impact meeting participants; predictability, accountability and common understanding. Applying these mechanisms enables us to classify meeting-related interactions into three types. The first type of meeting-related interactions (e.g. scheduling and sharing information about a meeting) enables employees to plan and predict their tasks. These interactions also inform the employees about the interdependencies among organizational activities, so they can anticipate other employees’ activities. Thus, these interactions improve predictability (Crowston, 1997; Okhuysen and Bechky, 2009). The second type of meeting-related interactions assigns the tasks and makes clear who is responsible for specific aspects of a task. Organizational members know their responsibilities, and individual contributions are made visible. These interactions form accountability (Crowston, 1997; Okhuysen and Bechky, 2009). The third type of meeting-related interactions helps meeting attendees recognize the value of the meetings by translating the social interactions into shared meanings (Geimer et al., 2015; Stensaker et al., 2008). Shared meanings provide the comprehension of the complex unit and how individual contributions fit within it (Joseph and Gaba, 2020). Thus, these interactions provide common understanding (Okhuysen and Bechky, 2009).

From attendee involvement to meeting effectiveness

Two distinct constructs represent the effects of meeting-related interactions: attendee involvement and meeting effectiveness. Attendee involvement is emotional evaluation, which reflects feelings about the meeting (Allen et al., 2012; Baran et al., 2012; Blanchard and McBride, 2020; Geimer et al., 2015; Leach et al., 2009). It is correlated with direct meeting satisfaction (Allen et al., 2016; Kauffeld and Lehmann-Willenbrock, 2012; Lehmann-Willenbrock et al., 2016; Rogelberg et al., 2010) and meeting quality (Cohen et al., 2011). Attendee involvement is a predictor of perceived meeting effectiveness and mediates the effect of some design characteristics on meeting effectiveness (Geimer et al., 2015; Leach et al., 2009).

Meeting effectiveness is the most used measure connected with meetings. In most cases, a questionnaire on perceived meeting effectiveness is used (Allen et al., 2014b; Leach et al., 2009; Shanock et al., 2013). Gathering feedback on meeting effectiveness through an anonymous paper-and-pencil or electronic survey after meetings is also a recommended tool for meeting leaders in practice (Lehmann-Willenbrock et al., 2018). Although meeting effectiveness should reflect employees’ work-related results, both researchers and practitioners rarely assess meeting effectiveness based on work results. The rationale for this lies in social construction through individual perception. The evaluation of the results (Geimer et al., 2015; Kauffeld and Lehmann-Willenbrock, 2012; Lehmann-Willenbrock et al., 2018), employee goal achievement (Allen et al., 2014b; Rogelberg et al., 2006; Shanock et al., 2013) and perceived performance (Yoerger et al., 2018) occurs through individual perception and shared meanings of the situation (Stensaker et al., 2008).

Interactions before meetings and attendee involvement

Meetings can be scheduled or unscheduled, but effective meetings are often planned in advance (Allen et al., 2014a; Baran et al., 2012; Volkema and Niederman, 1995). Scheduled meetings allow participants to prepare their arguments and strategies and enable them to explore other participants’ opinions about the agenda (Eisenbart et al., 2016). Appropriate meeting intervals also allow participants to plan their workload more effectively and come to a meeting feeling more motivated (Cohen et al., 2011; Geimer et al., 2015). Thus, planning works through the predictability coordination mechanism. Therefore, our first hypothesis posited the following:

H1.

Planning meetings positively influences perceptions of attendee involvement in meetings.

Providing a meeting agenda in advance is one of the prominent design characteristics of meetings (Baran et al., 2012; Leach et al., 2009). An agenda comprises three distinct pieces of information. First, information about the meeting venue, date, time and program provides participants with basic information about the meeting (Leach et al., 2009; Lehmann-Willenbrock et al., 2018). Second, the meeting goal frames participant expectations (Cohen et al., 2011; Leach et al., 2009). Third, distributing the documents needed for the meeting notifies attendees of the topics to be discussed (Allen et al., 2012; Cohen et al., 2011; Eisenbart et al., 2016; Leach et al., 2009). A meeting agenda that is distributed before a meeting increases motivation to participate in the meeting (Geimer et al., 2015). Information availability is the basis for an affective orientation to a meeting and works through the predictability coordination mechanism. Therefore, we formulated the following hypothesis:

H2.

Providing a meeting agenda before meetings positively influences perceptions of attendee involvement in meetings.

The intentional effort that participants invest in meeting preparation strengthens their motivation to actively participate in a meeting (Allen et al., 2016; Geimer et al., 2015; Rogelberg et al., 2010). Pre-meeting contributions to the meeting agenda require such intentional effort, and there are several reasons why working on the meeting agenda may improve attendee involvement. First, agenda-based preparation often requires a search for further information, which increases the probability that participants will be engaged in the discussion (Eisenbart et al., 2016). Second, the preparation of the agenda signals to attendees that they can influence the meeting and they understand how they can contribute to the meeting (Baran et al., 2012). Third, thorough preparation for the meeting enables deeper immersion into the agenda and thus effective participation (Geimer et al., 2015). In general, pre-meeting contributions to the meeting agenda make participants more aware of their responsibilities and tasks and create their accountability about the course of the meeting. Thus, we formulated the following hypothesis:

H3.

Pre-meeting contributions to the meeting agenda positively influence perceptions of attendee involvement in meetings.

The literature suggests examining attendee opinions before a meeting (Baran et al., 2012; MacLeod, 2011; Yoerger et al., 2018). Just the opportunity to provide an opinion (Allen et al., 2014b) or asking attendees for their views beforehand (Yoerger et al., 2018) engages participants in the subsequent meetings because just asking for an opinion makes participants prepare for the meeting. This idea is based on the general principle that the opportunity to provide an opinion increases feelings of personal importance, which in turn increases personal accountability (Meyer and Allen, 1991).

However, the effect does not have to be direct. This effect may be mediated by the action of providing the opinion because a shared opinion could sparkle involvement (Baran et al., 2012). The effect may also arise through informal talks, as attendees may share their opinion with their peers. Talking about the meeting agenda is a common part of pre-meeting talks, and these informal talks contribute to attendee involvement (Allen et al., 2014b). Therefore, we proposed the following hypothesis:

H4.

The opportunity to provide an opinion about the meeting agenda positively influences perceptions of attendee involvement in meetings.

Interactions during meetings and the mediating effect of attendee involvement

Although previous studies have examined many aspects of meeting design and leadership, one variable can represent effective interactions during meetings above all others. This variable is active participation in two-way communication during a meeting. Previous research has concluded that participation and feedback (Kauffeld and Lehmann-Willenbrock, 2012; Rogelberg et al., 2012; Yoerger et al., 2015), open communication (Allen et al., 2014b; Williams and La Brie, 2015), leader-attendee interactions (Mroz et al., 2018), or interactivity (Blanchard and McBride, 2020) are the most significant contributors to attendee involvement. Thus, it is possible to conceptualize interactions during meetings as a single variable without losing explanatory power.

Coordination theory proposes that interactions and conversations generate the energy that people invest in a meeting, and this energy, in turn, improves coordination (Quinn and Dutton, 2005). This confirms that a simple and an overarching construct for assessing the coordinating role during a meeting is evaluating attendee interactions during the meeting. The impact of attendee interactions during meetings on effectiveness is not direct. In the first phase, interactions during meetings strengthen affective orientation toward a meeting (Geimer et al., 2015; Jarzabkowski and Seidl, 2008), and, consequently, influence attendee involvement. In the second phase, attendee involvement helps them immerse themselves into the process (Beadle and Knight, 2012), and, thus, builds common understanding (Joseph and Gaba, 2020; Stensaker et al., 2008), thereby leading to perceived meeting effectiveness. To validate the proposed mediation relationship, we proposed the following hypotheses:

H5a.

Own interactions during the meeting positively influence the perceptions of attendee involvement in meetings.

H5b.

Perception of attendee involvement in meetings positively influences perceived meeting effectiveness.

Interactions after meetings and perceived meeting effectiveness

Although meeting minutes and other reports are often described as a critical meeting outcomes (Huang et al., 2018; Lehmann-Willenbrock et al., 2018; Mroz et al., 2018), there is no evidence on the effects of incorporating such outcomes in the subsequent work of attendees. Geimer et al. (2015) have quoted an attendee who had said, “too much time is spent addressing issues and no action steps ensue” (p. 2020). Generally, incorporating outcomes in work after a meeting makes sense of what was discussed because the outcomes materialize the verbal and cognitive processes that support common understanding (Stigliani and Ravasi, 2012). Accordingly, we proposed the following hypothesis:

H6.

Incorporating meeting outcomes in subsequent work positively influences perceived meeting effectiveness.

Although business practice is aware of post-meeting interactions (MacLeod, 2011), informal interactions after meetings have not received any scholarly attention. The analogy of pre-meeting talks (Allen et al., 2014b; Yoerger et al., 2018) can be used, but the mechanism for follow-up discussion differs. Not only do meetings have the potential to satisfy social needs (Geimer et al., 2015) but also follow-up discussions may create shared meanings (Stensaker et al., 2008), which helps build common understanding. Accordingly, the following hypothesis was formulated:

H7.

Follow-up discussions after meetings positively influence perceived meeting effectiveness.

The constructs included in the research framework in Figure 1 present the explicit meeting-related interactions and their effects in particular phases. The research framework links previous findings to new predictions on post-meeting interactions. H1–H4 were formulated to test previous findings on interactions before meetings. H5a and H5b include interactions during meetings and provide the central link to perceived meeting effectiveness. H6 and H7 were formulated to test new predictions related to interactions after meetings.

Methods

Sample and procedures

A survey research design was used to test the effect of interactions on perceived meeting effectiveness among regular workplace meeting attendees. The respondent-driven sampling was used (Heckathorn, 2011; Heckathorn and Cameron, 2017). This sampling method allows for making asymptotically unbiased estimates about populations, in which it is difficult to list population members. For organizational research, this selection process yields a sample that is more representative than larger samples recruited using convenience sampling or self-selection techniques (Heckathorn, 2011; James, 2006; Kalleberg et al., 1990).

The sample was selected from organizations providing internships to the students of a prominent business school. Representatives (CEOs in small organizations and HR managers in large organizations) from 18 organizations were asked to recruit 10 (if possible due to organizational size) participants each from their organization who would regularly attend workplace meetings as participants. The representatives were asked to recruit participants at different positions (specialists, lower and middle managers) with different levels of expertise and to include participants of both genders and different age groups. This selection method allows for the achievement of maximum variation of the sample and representativeness (Heckathorn and Cameron, 2017; Teddlie and Yu, 2007) on the one hand. On the other hand, we could not ask for identifying variables (e.g. sex, age, experience and position) in the questionnaire to ensure the anonymity and confidentiality of the respondents. In sum, the representatives recruited and sent the questionnaire link to 174 participants. From this sample, 128 participants completed the questionnaire, which yielded a response rate of 74%. Despite the relatively low size, the sample could yield satisfactory results for the proposed mediating effect (Fritz and MacKinnon, 2007).

The organizations represent both the secondary (manufacturing) and tertiary (services) sectors. According to the OECD enterprise size categories, three organizations are micro, four are small, four are medium and seven are large. A total of 25% of participants attend meetings daily, 31% several times a week, 19% weekly and 25% monthly. The duration of the meetings is normally distributed. The typical meeting takes between 30 and 60 min (53%), while meetings shorter than 15 min (1.6%) or longer than 90 min (6.2%) are rare. Only 14% of the participants perceived the number of meetings as excessive.

Pilot testing across two phases preceded data collection. The participants of the pilot testing were middle managers who regularly attended workplace meetings. In the first phase, five participants checked whether the questionnaire used terminology that would be meaningful to potential respondents (Gehlbach and Brinkworth, 2011). Several changes, most of which were related to phrasing, were made. In the second phase, five different participants completed the questionnaire, and their response time was monitored. After completing the questionnaire, they discussed their understanding of the individual items. The discussion made it apparent that they understood the questionnaire contents, and the validity of the measures was found to be promising. None of the participants of pilot testing was included in the consequent data collection.

Measures

To capture explicit interactions before, during and after meetings, we focused on concrete interactions. Measuring concrete constructs and actions permitted the use of single-item scales for most of the variables. Single-item scales yield predictive validity indices that are similar to those of multiple-item scales when measuring concrete constructs (Bergkvist and Rossiter, 2009, 2007). They have high reliability in measuring even psychological constructs, such as job satisfaction (Wanous et al., 1997; Wanous and Hudy, 2001) or satisfaction with an event (Ginns and Barrie, 2004). Additionally, single-item scales are more suitable for use with respondents from business settings (Bergkvist and Rossiter, 2009, 2007) and prevent common method bias (Bergkvist and Rossiter, 2007). The only variable assessed using a multiple-item assessment was “Provided agenda.” Three items were aggregated into a composite variable. Its Cronbach α (0.8560) was indicative of good composite reliability. To maximize reliability, a six-point frequency scale was used. Frequency scales are particularly useful for the occurrence of facts and behavior, when they yield more reliable results than agreement scales, such as Likert scale (Brown, 2004; Krosnick, 1999). All the response options were labelled, which improves reliability (Weng, 2004). Table 1 presents the variables and items and refers to the previous findings that were used to create the items.

The regression models included four control variables. The first two control variables were included for testing the hypotheses with attendee involvement as the dependent variable (H1–H5a). Meanwhile, the third control variable was included for testing the hypotheses leading to perceived meeting effectiveness (H5a–H7). This is because meeting excess affects overall perceived meeting effectiveness rather than attendee involvement during the meeting (Rogelberg et al., 2006). Organization size was the first control variable because smaller organizations are less formal, and employees spend less time on meetings in smaller organizations (Scott et al., 2012). Frequent meetings can disrupt employee activity and increase fatigue and subjective workload (Luong and Rogelberg, 2005). Meeting frequency served as the second control variable because appropriate intervals between meetings enable employees to reserve time for meetings and immerse themselves in the meeting agenda at the time of the meeting (Allen et al., 2012, 2014a; Geimer et al., 2015). Individual perceptions of an excessive number of meetings (meeting excessiveness) was the third control variable because the perception of an appropriate interval between meetings is subjective (Luong and Rogelberg, 2005), and perceptions of unnecessary meetings decrease worker satisfaction with meetings and overall productivity (Elsayed-Elkhouly et al., 1997).

Finally, the study controlled for meeting purposes, because meetings are not homogenous activities and can be conducted for various purposes, which may affect its perceived effectiveness. We distinguished 10 typical meeting purposes following the coordination perspective (which is more suitable for the research framework than the traditional functional perspective on meeting types; cf. Allen et al., 2014; Schwartzman, 1989). The first group of meeting purposes generates accountability and contains the purposes of goal setting, task allocation and task integration. The second group of meeting purposes forms predictability and covers purposes of information sharing, approvals of reports and resolving problems. The third type of meeting purposes facilitates social interactions and enhances common understanding. These are meetings that reconcile conflicts, promote sharing ideas, build trust and facilitate communication. The overview of the meeting purposes, occurrence and further analysis are in Table 5.

Results

Ordinal logistic hierarchical regression analysis was conducted using Stata (ver. 17.0) to test the hypotheses (note: because of permanent disputes over continuous/ordinal character of the data produced by using Likert or frequency scales, we ran also a linear regression analysis. The direction and statistical significance of all tested variables in all models were identical). Multicollinearity did not affect the regressions because the correlation matrix revealed low correlations among the variables. In accordance with the research framework, we used two sets of hierarchical regression models. Table 2 presents the first set of five regression models representing the left side of the framework (H1–H5) with perceptions of attendee involvement as the result of interactions before and during meetings (Models 1–4). Model 1 tested only the control variables. Model 2 tested the control variables and interactions before meetings, whereas Model 3 yielded regression results for the control variables and interactions during meetings. Model 4 included all the variables from the left side of the framework. Model 5 added the purpose of the meetings as additional control variables.

Organization size did not have a significant effect in any of the models. Meeting frequency was significant only in Models 1 and 2; however, the effect diminished when other variables were added, with better predictability indicated by the pseudo R-squared value. The type of the meetings was insignificant for all purposes except for a small significant positive effect of task allocation. Models 2, 4 and 5 yielded nonsignificant results for the hypothesized effect of meeting planning on attendee involvement, and therefore H1 did not find support. H2 predicted that providing meeting agendas before meetings will positively influence perceptions of attendee involvement. Models 2 and 4 supported H2, and Model 5 is only slightly above the threshold p-value (β = 0.4203, p = 0.063). The results generated for H3 provide good support for H2; the effect of pre-meeting work on perceptions of attendee involvement in meetings was positive and significant. On the other hand, H4, which predicted that the opportunity to share opinions about the meeting agenda before meetings positively influences perceptions of attendee involvement, was not supported in either model. A possible explanation is that the opportunity does not have a direct effect, but the act of providing an opinion acts as the intermediate between the opportunity to share opinions and perceptions of attendee involvement. Models 3, 4 and 5 showed that there was a significant and positive relationship between participants’ own interactions during the meeting and attendee involvement, which meets the first criterion for testing the mediating effect predicted by H5a. In sum, the results supported H2 and H3 but not H1 and H4.

Table 3 Presents the second set of five regression models representing the right side of the framework (H5–H7) with perceived meeting effectiveness as the result of attendee involvement during meetings and interactions after meetings (Models 6–10). Model 6 included only the control variable. Model 7 included the control variable and perceptions of attendee involvement. Model 8 included the control variable and interactions after meetings. Model 9 included all the variables from the right side of the framework. Model 10 added the purpose of the meetings as additional control variables. All the models confirmed that meeting excess had a significant and negative effect on perceived meeting effectiveness. Two meeting purposes had a significant effect on perceived meeting effectiveness. Resolving problems had a negative effect (β = −0.711; p = 0.002) and common understanding had a positive effect (β = 1.314; p < 0.001). The effect of meeting purposes is subject to post hoc analysis in Table 5.

Incorporating meeting outcomes in subsequent work positively influenced perceived meeting effectiveness; thus, H6 was supported. The results did not support H7. Conversely, Models 9 and 10 indicated that discussions after meetings negatively influenced perceived meeting effectiveness, thereby providing partial support for an alternative explanation. Perceptions of attendee involvement had a significant and positive effect on perceived meeting effectiveness, which supports H5b. These results further supported the suitability of a mediation test.

H5, as a whole, pertained to the mediating effect of attendee involvement on the relationship between participant own interactions during meetings and their perceived meeting effectiveness. Table 4 presents the set of regressions that met all the conditions for a mediating effect to be detected (Aguinis et al., 2017; Frazier et al., 2004). In Step 1, we found a significant relationship between the predictor (own interactions during meetings) and outcome (perceived meeting effectiveness). In Step 2, the predictor was found to be related to the mediator (attendee involvement). In Step 3, the mediator was found to be associated with the outcome. Finally, in Step 4, the relationship between the predictor and outcome weakened when the mediator was included in the model. The relationship became nonsignificant; thus, the mediating effect was confirmed.

Post hoc analysis on purpose of meetings

An important concern in this study is whether the results are consistent regardless of the purpose of meetings, or whether they differ with changing the purpose of meetings. As each meeting can have multiple purposes, which are not mutually exclusive, the questionnaire asked for the occurrence of every purpose individually. Table 5 presents the occurrence of meeting purposes, in which 10 different meeting purposes form three types. A mean greater than 4 suggests that a particular purpose is prevalent in most meetings. In contrast, a mean smaller than 3 implies that such a purpose does not occur during most meetings. While individual purposes of Accountability (a–c) and Predictability (d–e) types are individually distinguishable, variables in the Common understanding group yield strong internal consistency (α = 0.813). Therefore, variables g–j were treated as a composite variable Common understanding in the analyses.

To test the stability of the main results across different purposes of the meetings, we ran over 50 additional regression analyses. For Models 1–4 and Models 6–9, individual regression analysis for each purpose of the meeting was performed. The results are stable and confirm both the magnitude and significance of the regression coefficients from Tables 2 and 3 for all types of meetings with only two exceptions. First, in Model 4, if the purpose of a meeting is task allocation, meeting planning (H1) is a significant predictor of attendee involvement with a small and positive effect (β = 0.285; p = 0.031). Second, if the purpose of a meeting was common understanding, meeting excess was not a significant predictor of perceived meeting effectiveness anymore (β = −14.458; p = 0.994). The latter exception may be explained either by the lower number of meetings with the purpose of common understanding (n = 60) or alternatively, by employee desire for the social aspect of meetings in coherence with Model 10 in Table 3. In sum, these exceptions do not provide alternative or conflicting explanations and the results may be generalized regardless of the type of meetings.

Discussion

The use of coordination theory enabled us to expand the temporal framework of meetings with a focus on the interdependence of interactions before and after meetings, which have been neglected in previous research on workplace meetings. The findings extend existing theory in several ways. First, only a few studies have acknowledged the existence of interactions after meetings (Geimer et al., 2015; MacLeod, 2011; Mroz et al., 2018), but research has not yet investigated how and to what degree they affect meeting effectiveness. This study underscores the significant effect of incorporating meeting outcomes in subsequent work on perceived meeting effectiveness because interactions after meetings improve perceived meeting effectiveness. The evaluation of meeting effectiveness occurs not only during the meetings but also especially after the meetings with the incorporation of the meeting outcomes into work.

Second, the literature on meetings largely offers suggestions on how meetings should be designed and run. The literature often treats meetings as temporary and locally delimited gatherings, but this study expanded this conceptualization of meetings into one characterized by a longer period and explains how interactions before and after meetings contribute to meeting effectiveness. The mediating effect of attendee involvement demonstrates the interdependence of interactions before, during and after meetings. The results leverage previous conceptual developments (Geimer et al., 2015; Lehmann-Willenbrock et al., 2018; Mroz et al., 2018) by providing empirical evidence.

Third, our findings support the notion that explicit interactions ascribe meaning to meetings (Espinosa et al., 2004; Jarzabkowski and Seidl, 2008). Interactions that are tangible and require intentional effort, make it easier to notice individual contributions as a result of such interactions. This may explain why interactions that require greater intentional effort (i.e. pre-meeting work, own interactions during meetings and incorporating meeting outcomes in subsequent work) had a strong and significant effect, whereas interactions requiring less or no intentional effort (i.e. meeting planning, opinion opportunity and discussions after meetings) were not significantly related to attendee involvement or perceived meeting effectiveness.

Finally, the results did not show that informal discussions after meetings are a significant contributor to perceived meeting effectiveness. This finding does not imply that informal interactions are irrelevant. Instead, they may affect different aspects of organizational life, such as organizational culture (Geimer et al., 2015), well-being (Luong and Rogelberg, 2005) and job satisfaction (Rogelberg et al., 2010). Alternatively, employees may separate their work conditions from their personal life or find it difficult to perceive interactions after meetings as being connected to prior meetings (Geimer et al., 2015). This finding provides grounds for further inquiry into formal and informal interactions during and after meetings.

Practical implications

The present findings offer several implications with the potential to improve perceived meeting effectiveness in organizations. First, perceived meeting effectiveness depends on the pre-meeting work of participants and using meeting outcomes in their consequent work. Pre-meeting work, including preparing presentations, analyses, summaries or comments, significantly improves attendee involvement. Using outputs from meeting in participants’ subsequent work further improves perceived meeting effectiveness. Thus, we recommend considering roles and tasks for each participant, assigning them tangible tasks before the meeting and presenting the conclusions of meetings in a way that exposes their usefulness in participants’ subsequent work. Considering this recommendation, organizers can assess which participants to invite to a particular meeting and limit the number of unnecessary attendees.

The second implication relates to meeting frequency. Our results support the notion that meeting frequency does not affect perceived meeting effectiveness per se. Even frequent meetings may be perceived as effective if interactions before and during meetings foster involvement, and the meeting outcomes serve as inputs for subsequent work. This finding explains why meeting quality is more important than quantity (Allen et al., 2012) and meeting length is unrelated to effectiveness (Leach et al., 2009).

The third implication considers the mode of meetings. While we studied workplace on-site meetings, virtual meetings have become a new norm. It would be courageous to draw bold conclusions about online meetings, but our results strongly support that participants’ own interactions during meetings are an important factor in perceived meeting effectiveness. Virtual meetings are characterized by fewer interactions (Blanchard and McBride, 2020) and offer fewer opportunities to engage with all participants. Hence, effective involvement during online meetings may be important for their effectiveness. In contrast, informal interactions after meetings were not significantly related to perceived meeting effectiveness; thus, the typical lack of these interactions during online meetings may not negatively impact perceived meeting effectiveness.

Limitations and suggestions for future research

Our study has several limitations. The first limitation pertains to the use of a cross-sectional design of the study. We measured all the variables at a single point in time, which may produce artifactual covariance of the dependent and independent variables (Podsakoff et al., 2003). We paid particular attention to this issue in both phases of pilot testing of the questionnaire, where participants’ responses seemed promising in the differentiation of the content of the variables. The second limitation lies in the number of control variables included in the model. Meeting effectiveness can be influenced by a great number of other meeting design characteristics (Cohen et al., 2011 provide 18 characteristics affecting meeting effectiveness). This study controlled for four of them (frequency, excess, organizational size and modality), which were previously tested as the most important. The third limitation pertains to the use of single-item measures. We used these with caution (Diamantopoulos et al., 2012) and followed recommendations for the use of single-item measures (Bergkvist and Rossiter, 2009, 2007; Ginns and Barrie, 2004). This yielded a satisfactory response rate and allowed us to assess interactions before, during and after meetings. Moreover, the pilot tests convinced us that the respondents had understood the questionnaire contents accurately. The fourth limitation is connected with the sampling frame. Respondent-driven sampling allowed us to achieve a high response rate and representativeness of the data, but the control over the quality of the sample was delegated to the hands of representatives of the sample organizations, and we did not collect demographic characteristics to ensure anonymity and confidentiality of the respondents.

Eventually, we suggest some directions for future research. We revealed that interactions after meetings have been completely neglected, and our study provides a foundation for further studies on interactions after meetings. Future studies can explore how different types of interactions after meetings influence work directly or indirectly. This study found that discussions after meetings had a small but negative effect on the perceived meeting effectiveness. However, discussions after meetings may affect other aspects of organizations, such as interpersonal relationships or organizational culture, and thus, are worth exploring.

Regarding the modes of meetings, we currently witnessed diversification, as some organizations returned to on-site meetings, some switched to online meetings only and some used various types of hybrid meetings. Exploring the contingencies of various modes or comparing the effect of meeting characteristics on meeting effectiveness in different modes of meetings could help us make the most out of technological advancements.

From a methodological point of view, future studies can exploit current developments in technology to collect data concerning all possible interactions in online workplace meetings (Williams and La Brie, 2015). For example, software solutions will soon produce high-quality meeting reports based on speech recognition and automatized data analysis (Huang et al., 2018). Such data collection procedures will facilitate large longitudinal studies once their technological and ethical considerations are clarified.

Conclusion

This study links previous findings to new predictions on interactions after meetings and reveals that interactions before, during and after meetings jointly influence perceived meeting effectiveness. Before meetings, sharing the agenda and pre-meeting work increases attendee involvement. Own interactions during the meeting positively influence attendee involvement, which, in turn, positively influences perceived meeting effectiveness. Incorporating meeting outcomes in post-meeting work has a large and significant positive effect on perceived meeting effectiveness. In a nutshell, an active role of meeting participants not only during but also before and especially after the meeting, is crucial for perceived meeting effectiveness, regardless of the type of meeting. Our study provides ground for future studies, particularly in post-meeting interactions, and we encourage meeting leaders to consider the active role of individual attendees in workplace meetings.

Figures

Figure 1

Research framework and hypotheses

Table 1

Measures

Variables (adapted from…)	Items How often …
Explanatory variables
Meeting planning (Eisenbart et al., 2016; Geimer et al., 2015)	… are the meetings in your organization planned?
Provided agenda (Cohen et al., 2011; Leach et al., 2009; Lehmann-Willenbrock et al., 2018)	… do you receive the program of meetings in advance? … do you know the purpose/goal of meetings? … do you receive documents needed for meetings in advance?
Pre-meeting work (Geimer et al., 2015; Rogelberg et al., 2010)	… are you expected to prepare materials (presentations, analysis, comments, summaries, etc.) for meetings?
Opinion opportunity (Baran et al., 2012)	… do you have an opportunity to give your opinion on meeting agenda before meetings?
Own interaction during meetings (Allen et al., 2014b; Geimer et al., 2015; Rogelberg et al., 2012)	… are you involved in two-sided communication during meetings (excluding side communication)?
Meeting outcomes used for work (Geimer et al., 2015)	… do you use outputs from meetings in your subsequent work?
Discussions after meetings	… do you informally discuss the meeting agenda with other meeting participants after meetings?
Dependent and mediating variables
Attendee involvement (Geimer et al., 2015; Leach et al., 2009)	… do you perceive yourself as an involved person at meetings?
Meeting effectiveness (Geimer et al., 2015; Leach et al., 2009)	… do you perceive meetings effective?

Notes:

All items are six-point frequency scales: 1 – never; 2 – rarely; 3 – occasionally; 4 – often; 5 – very often; 6 – always

Table 2

Ordinal logistic regression model for attendee involvement

	Model 1	Model 2	Model 3	Model 4	Model 5
Hypothesized relationships
Meeting planning (H1)	–	0.0177 (0.945)	–	0.3583 (0.181)	0.2369 (0.452)
Provided agenda (H2)	–	0.3716* (0.040)	–	0.4019* (0.047)	0.4203 (0.063)
Pre-meeting work (H3)	–	0.7935** (0.000)	–	0.4395** (0.008)	0.4171* (0.037)
Opinion opportunity(H4)	–	0.0771 (0.555)	–	−0.1887 (0.206)	−0.1116 (0.450)
Own interaction during meetings (H5a)	–	–	1.7858** (0.000)	1.7264** (0.000)	1.591** (0.000)
Control variables
Organizational size	0.2955 (0.109)	−0.3305 (0.101)	−0.1118 (0.558)	−0.3252 (0.122)	−0.3729 (0.126)
Meeting frequency	0.3352** (0.001)	0.2580* (0.018)	0.0037 (0.974)	0.0687 (0.556)	0.1015 (0.404)
Meeting purpose
Goal setting	–	–	–	–	0.1553 (0.444)
Task allocation	–	–	–	–	0.3971* (0.040)
Task integration	–	–	–	–	0.1244 (0.586)
Resolving problems	–	–	–	–	−0.0378 (0.865)
Report approvals	–	–	–	–	−0.0515 (0.746)
Information sharing	–	–	–	–	−0.2926 (0.107)
Common understanding	–	–	–	–	0.5714 (0.053)
Pseudo R2	0.0315	0.1591	0.2253	0.2831	0.3172

Notes:

Dependent variable is Attendee involvement in all models above. Meeting purpose is subject to further analysis later in the article. Common understanding is a composite variable for communication facilitation, conflict reconciliation, creating shared ideas, and building trust (α = 0.813; see post hoc analysis and Table 5 for details). Regression coefficients are the ordered logit coefficients; A Pseudo R² has meaning only in this particular model and indicates which model better predicts the outcome. p-values in parenthesis; *p < 0.05; **p < 0.01

Table 3

Ordinal logistic regression model for perceived meeting effectiveness

	Model 6	Model 7	Model 8	Model 9	Model 10
Hypothesized relationships
Attendee involvement (H5b)	–	1.0090** (0.000)	–	1.0393** (0.000)	1.1660** (0.000)
Meeting outcomes used for work (H6)	–	–	1.4651** (0.000)	1.3982** (0.000)	1.3650** (0.000)
Discussions after meetings (H7)	–	–	−0.0381 (0.849)	−0.5464* (0.022)	−0.7001* (0.012)
Control variables
Meeting excess	−1.743** (0.001)	−1.850** (0.001)	−1.6455** (0.004)	−1.8292** (0.003)	−2.3093** (0.001)
Meeting purpose
Goal setting	–	–	–	–	−0.1300 (0.568)
Task allocation	–	–	–	–	−0.1064 (0.681)
Task integration	–	–	–	–	0.0631 (0.765)
Resolving problems	–	–	–	–	−0.7109** (0.002)
Report approvals	–	–	–	–	−0.1789 (0.295)
Information sharing	–	–	–	–	−0.3234 (0.148)
Common understanding	–	–	–	–	1.3135** (0.000)
Pseudo R2	0.0384	0.1572	0.2163	0.2991	0.3835

Notes:

Dependent variable is Perceived meting effectiveness in all models above. Meeting purpose is subject to ad hoc analysis later in the article. Common understanding is a composite variable for communication facilitation, conflict reconciliation, creating shared ideas, and building trust (α = 0.813; see post hoc analysis and Table 5 for details); Regression coefficients are the ordered logit coefficients; A Pseudo R² has meaning only in this particular model and indicates which model better predicts the outcome. p-values in parenthesis; *p < 0.05; **p < 0.01

Table 4

Set of ordinal logistic regressions to demonstrate the mediating effect of perceived attendee involvement on the relationship between participants’ interaction during meetings and their perception of meeting effectiveness (H5)

	Log-odds coef.	p-value
Step 1 (Dependent variable: Perceived meeting effectiveness)
Own interaction during meetings	0.8573**	0.000
Step 2 (Dependent variable: Attendee involvement)
Own interaction during meetings	1.8718**	0.000
Step 3 (Dependent variable Perceived meeting effectiveness)
Attendee involvement	1.4030**	0.000
Step 4 (Dependent variable: Perceived meeting effectiveness)
Attendee involvement	0.7897**	0.000
Own interaction during meetings	0.3832	0.080

Notes:

Dependent variable is stated for each step. Regression coefficients are the ordered logit coefficients. *p < 0.05; **p < 0.01

Table 5

Meeting purposes: means, standard deviations, reliability coefficients, and correlation matrix

Meeting purpose	Mean	SD	a	b	c	d	e	g	CU
Accountability (α = 0.385)
a) Goal setting	4.67	1.00	1.00
b) Task allocation	4.61	0.98	0.08	1.00
c) Task integration	4.47	1.08	0.22*	0.22*	1.00
Predictability (α = 0.185)
d) Resolving problems	4.75	1.06	0.14	0.09	0.34**	1.00
e) Report approvals	2.64	1.42	0.10	−0.11	0.03	0.13	1.00
f) Information sharing	4.69	1.15	0.09	0.18	0.09	−0.01	0.08	1.00
Common understanding (CU) (α = 0.813)	3.30	0.99	−0.055	0.20*	0.02	0.30**	0.35**	0.36*	1.00
g) Communication facilitation	3.80	1.23
h) Conflict reconciliation	2.32	1.21
i) Creating shared ideas	3.80	1.08
j) Building trust	3.31	1.40

Notes:

All items are 6-point frequency scales: 1-never; 2-rarely; 3-in majority of meetings not; 4- in majority of meetings not; 5-almost always; 6-always. As a result of internal consistency (α = 0.813), Common understanding is treated as a composite variable and correlation coefficient is calculated as mean of (g + h+i + j)/4; Thus, correlation coefficients for individual variables g) – j) are omitted. *p < 0.05; **p < 0.01