Picturing Uncertainty
1
Picturing Uncertainty:
Frames, Schemata, and Social Capital Activation in Organizations1
Sameer B. Srivastava
Haas School of Business, University of California – Berkeley
September, 2012
1Direct all correspondence to Sameer B. Srivastava, Haas School of Business, University of California - Berkeley,
545 Student Services, #1900, Berkeley, CA 94720-1900; [email protected]; 510-643-5922. I thank
Max Bazerman, Frank Dobbin, Roberto Fernandez, Richard Hackman, Laura Kray, Omar Lizardo, Peter Marsden,
Don Moore, Chris Muller, Sucheta Nadkarni, Erin Reid, Ned Smith, Toby Stuart, Andras Tilcsik, Cat Turco and
participants of the MIT Economic Sociology Working Group, the Harvard Business School “Non-Lab,” the
Academy of Management’s Cognition in the Rough Workshop, and Harvard’s Work, Organizations, and Markets
Seminar for helpful comments and suggestions on prior drafts. Jim Dowd, Amy Edmondson, Pam Hallagan, Maura
McCurdy, and Ingrid Peters helped me gain access to the research site and supported my data collection efforts. Jim
Cutler supplied his voice for the audio recordings used in the experiment. The usual disclaimer applies.
© Copyright 2012, Sameer B. Srivastava. All rights reserved. This paper is for the reader's personal use only. This
paper may not be quoted, reproduced, distributed, transmitted or retransmitted, performed, displayed, downloaded,
or adapted in any medium for any purpose, including, without limitations, teaching purposes, without the Author's
express written permission. Permission requests should be directed to [email protected].
Picturing Uncertainty
2
Picturing Uncertainty:
Frames, Schemata, and Social Capital Activation in Organizations
Abstract: This article examines how the frames and schemata through which
organizational actors interpret and respond to uncertain events influence social capital
activation – the conversion of latent social ties into active relationships. Bringing together
insights about social resource mobilization and managerial cognition during times of
organizational change, the author derives propositions about the size and composition of
networks that actors activate when operating under schemata associated with a commonly
used uncertainty frame – threat / opportunity. These propositions are tested in a field-
based experiment involving 158 individuals in a non-profit health care organization.
Results indicate that a loss (rather than gain) schema led actors to activate more ties and a
greater proportion of bridging ties. A limited control (rather than control) schema led
both males and females to activate a greater proportion of ties to male colleagues.
Moreover, responses to the control / limited control schemata were moderated by an
individual difference construct: the locus of control. These findings contribute to our
understanding of the cognitive mechanisms that underpin how organizational actors
mobilize social resources during times of tumult.
September, 2012
Picturing Uncertainty
3
INTRODUCTION
Organizational theorists have long studied events – such as restructuring, mergers, and senior leadership
transitions – that transform internal social structure (Burkhardt and Brass 1990; Gulati and Puranam
2009; Shah 2000). Large-scale organizational change often produces high levels of uncertainty for
organizational actors, for example, about how their status, resources, or structural position will change.
Uncertainty can, in turn, trigger the mobilization of social resources – such as information, influence, and
social support – that are accessible through interpersonal networks (McDonald and Westphal 2003;
Mizruchi and Stearns 2001; Pescosolido 1992). Just because valuable resources are available through
social relations does not, however, always mean they will be accessed. Trust-based barriers (Smith 2005),
interpersonal affect (Casciaro and Lobo 2008), and the cognitive recall of ties (Smith et al. 2012) all can
constrain the set of contacts people turn to and the nature of resources they obtain. That is, in many
situations, people activate only a subset of the relations to which they have access.
A growing body of research at the intersection of cognition and social networks has examined the
role of frames, which enable actors to “locate, perceive, identify, and label” events in their own terms
(Goffman 1974: 21), in shaping network dynamics (Hurlbert et al. 2000; McLean 1998; Stevenson and
Greenburg 2000). When operative, these frames make salient certain schemata, which “allow the brain to
exclude the specific details of a new experience and retain only the generalities that liken the event to
other experiences in one’s past” (Cerulo 2002: 8). Schemata, in turn, provide templates for action – the
“guided doings” an actor follows in a given situation (Goffman 1974: 22).
Drawing on the extensive literature about how managers identify, interpret, and respond to
strategic issues (Billings et al. 1980; Chattopadhyay et al. 2001; Dutton and Jackson 1987; George et al.
2006; Gladstein and Reilly 1985; Jackson and Dutton 1988; Ocasio 1995), I focus on a frame that is
widely used to make sense of uncertain organizational events: threat / opportunity. This frame
encompasses two distinct pairs of schemata: loss / gain and control / limited control (George et al. 2006;
Ocasio 1995). That is, people can see threat (opportunity) in a situation in two distinct ways: if they
perceive the potential to lose (gain) resources or the potential to lose (maintain) control. Some situations
Picturing Uncertainty
4
can, of course, make both pairs of schemata salient simultaneously. Responses to these schemata are
thought to be governed by different cognitive mechanisms: loss aversion (Kahneman and Tversky 1979),
threat rigidity (Staw et al. 1981), and reactance (Brehm 1966; Wortman and Brehm 1975). Drawing on
these theories, I derive predictions about the size and composition of networks that organizational actors
will activate when placed into situations that make salient the schemata of loss / gain and of control /
limited control. Next I report on a field experiment, involving 158 executives in a non-profit health care
organization, which was designed to test these predictions. The experimental set-up, which builds on but
importantly modifies an existing protocol (Smith et al. 2012), allows for causal identification of the
effects of schemata not only on the recall of ties but also on the purposive choice to activate ties – that is,
the choice to convert latent ties into active relationships. Moreover, the use of samples of working
professionals, rather than undergraduate subjects, enhances the external validity of the findings. I
conclude by discussing how this work contributes to our understanding of: (1) the cognitive
underpinnings of social capital activation; (2) how stable individual differences can influence network
action; (3) how formal and informal organizational structure can diverge during times of change; and (4)
sex differences in workplace networks.
SOCIAL CAPITAL ACTIVATION
Following Lin (2001: 29), I conceive of social capital as “resources embedded in a social structure that
are accessed and/or mobilized in purposive actions.” At any given time, many network ties that are
potential sources of valuable resources are latent – that is, people have pre-existing relationships but no
current interaction with a set of individuals. When faced with situations that require gaining access to
social resources – such as information, influence, or social support – actors convert some latent ties into
active ones. I follow Pescosolido (1992: 1105) in conceiving of networks as “antecedent to an event” and
in considering “how individuals, in response to a particular event, choose to activate particular sectors of
the multiple networks in which they are embedded.” Thus, consistent with Hurlbert, Haines, and Beggs
Picturing Uncertainty
5
(2000: 599), I define social capital activation as the choice to initiate contact with certain individuals
among the set of actors in one’s pre-existing network.2
FRAMES, SCHEMATA, AND SOCIAL CAPITAL ACTIVATION
The Threat / Opportunity Frame
Of the frames through which organizational actors view uncertain events, one of the most commonly used
is: threat / opportunity (Jackson and Dutton 1988). This frame operates across a wide variety of situations
– for example, the entry of a new competitor, the imposition of new legal or regulatory requirements, and
impending restructuring or merger – and can influence cognition and social action. For example, a recent
laboratory experiment reports that, when facing situations of threat, high status actors cognitively recalled
larger and less dense networks than did low status actors (Smith et al. 2012). The present study builds on
and extends this work in four key ways. First, consistent with the view that the practical evaluation of
uncertain future states is an inherently deliberative act (Emirbayer and Mische 1998; Hitlin and Elder
2007; Mische 2009), I consider not the cognitive recall of contacts but rather the purposive choice to
initiate interaction with those contacts. Second, I examine these choices not in a laboratory but rather in a
real organization, with subjects who have pre-existing organizational networks. Third, I consider not only
threat but also its counterpart – opportunity. Moreover, I conceptually unpack the threat / opportunity
frame into two constituent pairs of schemata: gain / loss and control / limited control (George et al. 2006;
Ocasio 1995). Existing theory makes disparate, sometimes contradictory, predictions about the kinds of
networks actors will activate when operating under these schemata. The research design used in this study
helps adjudicate among these perspectives. Finally, I consider how gender dynamics in the workplace can
shape network action when these schemata are operative.
2Smith (2005) defines social capital activation to include both an individual’s choice to seek resources from a
contact and the contact’s choice to provide aid to the help seeker. Because my arguments pertain only to the choices
of the focal actor, I focus on the former choice.
Picturing Uncertainty
6
The Gain / Loss Schemata
The theory of loss aversion supplies predictions about network activation when actors are operating under
the gain / loss schemata. It suggests that risk attitudes and behavior are based not only on the expected
returns of a decision but also on where the decision outcome stands relative to a predetermined reference
point in the mind of the decision maker (Kahneman and Tversky 1979; Tversky and Kahneman 1990). A
core prediction is that decision makers often prefer to avoid losses rather than achieve gains. This
empirical pattern does not only pertain to economic decisions; it also applies in situations of potential
status gain or loss. Recent experimental evidence suggests, for example, that people assign greater value
to status when recalling the risk of status loss than when recalling the risk of status gain and put forth
greater effort to avoid status losses than to achieve status gains (Pettit et al. 2010). To put it differently,
organizational actors can be expected to work harder to maintain the standing they have already achieved,
but which is at risk, than to strive for an uncertain improvement in their standing. Because social capital
activation requires effort, and because social connections are potential conduits to resources, the threat of
losing tangible resources or status will lead to the activation of a larger network than will the opportunity
to gain tangible resources or status. Thus, loss aversion predicts: Hypothesis 1: Organizational actors
will activate more ties in situations of loss than in situations of gain.
Potential losses can also accelerate information processing and trigger problemistic search
(Kahneman 1973; March and Simon 1958). In particular, losses shift attention toward novel or risky
solutions and away from well-learned routines. In deciding which network ties to activate, organizational
actors facing a potential loss can be expected to turn their attention toward the distal environment.
Moreover, bridging ties – for example, to colleagues in other departments or functions – are likely to be
sources of novel and valuable information (Friedkin 1982; McEvily and Zaheer 1999; but see Tortoriello
and Krackhardt [2010] on the role of Simmelian ties). At the same time, a variety of factors – for
example, divergent interpretive schemes (Dougherty 1992), inter-unit competition (Tsai 2002),
incompatible language systems (Bechky 2003), and knowledge differences and dependencies (Carlile
2004) – make it especially challenging to activate bridging ties. If prospective loss leads people to expend
Picturing Uncertainty
7
greater effort, then they are more likely to incur the costs of activating bridging ties. Thus, I expect:
Hypothesis 2: Organizational actors will activate a greater proportion of bridging ties in situations
of loss than in situations of gain.
The Control / Limited Control Schemata
Two competing theories supply predictions about network activation when actors face situations that
make salient the control / limited control schemata: threat-rigidity and reactance. Threat-rigidity theory
suggests that, relative to situations in which actors perceive a high level of control, situations of limited
control lead actors to constrict information processing and focus their attention on the proximate
environment (Chattopadhyay et al. 2001; George et al. 2006; Staw et al. 1981). These responses have
implications for social capital activation. If actors seek less information in situations of limited control,
they will activate fewer ties – at least for purposes of information gathering. Similarly, if actors are more
likely to search for information in the proximate, rather than distal, environment in situations of limited
control, they will activate a lower proportion of bridging ties. Thus, threat-rigidity theory predicts:
Hypothesis 3a: Organizational actors will activate fewer ties in situations of limited control than in
situations of control. Hypothesis 4: Organizational actors will activate a lower proportion of
bridging ties in situations of limited control than in situations of control.
By contrast, reactance theory suggests that, when faced with the loss of control, people respond –
at least initially – by seeking to gain personal mastery over the situation (Brehm 1966; Elster 2007;
Taylor 1983; Wicklund 1974). As Elster (2007: 41) explains: “Consider the…case of a person who faces
a barrier or impediment to her goal. This threat to her freedom of action may induce what psychologists
call ‘reactance’ – a motivation to recover or reestablish freedom….As an illustration, think of the effect of
hiding from a small boy a drum his parents do not want him to play with.” Such reactions to a loss of a
control are thought to occur in stages, with people first attempting to regain a sense of mastery and then
moving to a state of “learned helplessness” if these efforts fail (Wortman and Brehm 1975). People seek
mastery over a situation through a variety of means – for example, seeking information (Fiske and Dépret
1996) or interpersonal influence from powerful colleagues (Pfeffer 1992). The search for these resources
Picturing Uncertainty
8
will lead people to activate more social ties – at least in the initial stages after they experience a loss of
control. Thus, reactance theory suggests: Hypothesis 3b: Organizational actors will activate more ties
in situations of limited control than in situations of control.
Reactance theory does not directly address the question of what kinds of information or influence
people will seek as they gain mastery over a situation. It therefore does not offer predictions about the
proportion of ties activated that bridge organizational boundaries. It does, however, suggest that
individuals can vary in their responses to a loss of control. In particular, a well-established individual
difference construct – internal / external locus of control (Rotter 1954, 1966) – is thought to moderate
responses to control deprivation.3 People with an internal (rather than external) locus of control believe
they have mastery over the events they experience. They think that their own actions – rather than the
actions of others, fate, or chance – primarily determine their outcomes – including their outcomes in the
workplace (Spector 1988). When confronted with a situation of limited control, such individuals can be
expected to expend even greater effort to regain a sense of mastery – in particular, by initiating social
contact with others (Ng et al. 2006). These insights suggest the following: Hypothesis 3c: Internal,
rather than external, locus of control will amplify the tendency for organizational actors to activate
more ties in situations of limited control than in situations of control.
Finally, reactance theory has implications for the activation of same- or opposite-sex ties when
the control / limited control schemata are operative. As noted above, the theory predicts that a loss of
control prompts people to seek to regain control. In organizational settings, an important means for doing
so is by seeking out more powerful others who can exert influence on an individual’s behalf. In many
organizational settings, females are less likely than males to occupy positions of power (Reskin et al.
1999; Ridgeway 1997; Ridgeway and Smith-Lovin 1999). Moreover, females are often less central than
males in informal communication, advice, and influence networks in the workplace (Ibarra 1992a;
3There are a number of constructs closely related to locus of control – e.g., perceived powerlessness (Ross and Mirowsky
1989), perceived helplessness (Elder and Liker 1982), fatalism (Wheaton 1980), and personal autonomy (Seeman and Seeman
1983). In different ways, they represent the belief that events and outcomes are within or outside an individual’s control (for a
review, see Ross, Mirowsky, and Pribesh [2001]).
Picturing Uncertainty
9
Lincoln and Miller 1979; Ridgeway and Smith-Lovin 2006). Conditions of uncertainty also lead people to
rely more heavily on social categories, such as gender, when forming judgments about others (Gorman
2006; Pfeffer et al. 1976). Thus, in assessing the value of social contacts during periods of uncertainty-
producing organizational change, actors are more likely to rely on stereotyped gender beliefs about power
– regardless of the actual distribution of power in the organization (for a review, see Ridgeway and
Smith-Lovin 1999). Finally, experimental evidence suggests that males are more likely to be perceived to
occupy central positions associated with power – even after controlling for the actual network position of
males and females (Brands and Kilduff 2012). Thus, in situations of limited control, males will be even
more inclined to initiate contact with other male colleagues who wield more real or perceived power. At
the same time, situations of limited control will also increase the tendency for females to seek contact
with male colleagues whom they perceive to have more power. These arguments suggest: Hypothesis 5:
Males and females will both activate a greater proportion of ties to male colleagues in situations of
limited control than in situations of control.
DATA AND METHODS
Research Setting and Study Participants
The study sample comprised employees of a leading non-profit health care company who participated in a
custom (i.e., company-specific) executive education program at an East Coast business school. The
organization employed over 50,000 individuals. All current program participants (63 individuals) and
alumni of the program (188 individuals) were invited to participate. The program had been running for
four years, with approximately 60 people per cohort. I introduced the research study to one cohort of
participants during one of their on-campus sessions and invited past cohorts to participate through an
email communication.
The response rate for current participants was 82.5%, while that for alumni was 56.4%. The total
response rate was 62.9% (N=158). The resulting sample had the following characteristics: mean age –
50.4 years (standard deviation, 6.17); mean years of work experience – 25.4 years (standard deviation,
Picturing Uncertainty
10
7.59); proportion female – .468; proportion White – .722; proportion Black / African American – .089;
proportion Asian – .120; proportion Hispanic / Latino – .038; proportion born outside the US – .158;
proportion now married – .785; proportion never married – .095; and proportion working in a small office
(i.e., less than 500 employees) – .392.
This sample was well-suited to the objectives of this study. It included professionals with pre-
existing workplace networks and therefore afforded a more realistic picture of network activation than
could be achieved in a typical laboratory study. Because all respondents worked in the same company,
this design implicitly controlled for organization-level sources of variation, such as norms governing
workplace interaction. Finally, because the sample included long-tenured employees who had
experienced a great deal of organizational change, it was possible to construct experimental scenarios to
which they could easily relate and about which they could respond based on past experience.
Experimental Procedure
Participants were randomly assigned to one of four conditions: (1) loss – limited control; (2) loss –
control; (3) gain – limited control; and (4) gain – control.4 I sent them an email with a link to one of four
on-line surveys (depending on the condition to which they were assigned). I instructed them to click on
the link when they were alone, free from distractions, and had at least 10-15 minutes to devote to the
exercise. I also indicated that the exercise involved listening to an audio clip and that they should turn the
sound up on their speakers before beginning.
The first section of the survey included questions about their employment history (e.g., years of
work experience), role within the organization (e.g., individual contributor, manager, middle manager,
executive), and size of the local office in which they work. The manipulation came next. I asked subjects
to imagine a hypothetical situation playing out in their organization. They first listened to a voicemail
recording of an actor playing the part of the company’s CEO and describing an impending organization-
wide change. They had the option to rewind or replay the recording as many times as they wished. In
4In supplemental analysis (reported below), I also tested for potential interaction effects – e.g., how situations of loss
and limited control can affect network activation choices.
Picturing Uncertainty
11
addition, I provided them with a transcript of the recording. Next they read details about how the
organizational change could potentially affect them personally. In particular, they read a summary of a
hypothetical conversation they had with a trusted colleague who was well placed in the organization and
who had been a reliable source of information in the past. See full details of the manipulation below and
see Table 1 for a summary of the manipulation. After learning about this hypothetical situation, subjects
were asked how they interpreted the situation (see manipulation checks).
Manipulation5
CEO Voicemail: Good morning. I would like to share some important news with you. In light of
changing business conditions, we have decided today to implement a [Loss: restructuring / Gain: new
growth plan], which will result in some changes in organizational structure and reporting lines.
Later today, you will receive a memo that outlines these changes and explains why they are necessary to
ensure the long-term health and competitiveness of our enterprise. As these changes play out, you can
expect to receive regular updates from me and others in your management team. Thank you for your
attention and support.
Follow-up Communication: After listening to the voicemail from the CEO, you had a private meeting
with a trusted colleague who works elsewhere in the organization. This colleague has heard through the
grapevine (i.e., through unofficial channels) some additional details about the situation and what it might
mean for you. The colleague is well placed in the organization and has been a reliable source for you in
the past.
[Loss: Your colleague informed you that – as part of the reorganization – the organizational unit you are
in will be merged with another unit. Your manager, who heads your unit, will be moving to a different
part of the organization. The head of the other unit will run the combined group.
Several different options for how to structure the combined entity are being considered. One option
would involve inserting a management layer between you and the new unit head (i.e., you would report
to someone else, who would report to the unit head). The person they are considering to be your new
manager is someone from the other unit whom you do not know well but have generally considered a
peer.]
[Gain – Limited Control: Your colleague informed you that – as part of the reorganization – a new
position is opening up to lead a new unit that will pursue exciting new growth opportunities for the
organization. There are several candidates for this position, and you are among those being considered.]
[Gain – Control: Your colleague informed you that – as part of the reorganization – several positions are
opening up to lead new units that will pursue exciting growth opportunities for the organization. There
5Because this study focused on how the schemata associated with a given frame influence network activation (rather
than on how different ways of framing the same situation might affect network activation), the manipulation varied
substantive features of the situation across conditions (e.g., the resources or status an actor could potentially gain or
lose).
Picturing Uncertainty
12
are several candidates for these positions, and you are among those being considered for one of these
positions.]
Such a change would represent a significant step [Loss: back / Gain: forward] for you in your career and
[Loss: hurt / Gain: build] your status in the organization. In addition, this change in job role would likely
result in [Loss: a reduction / Gain: an increase] in your total financial rewards. [Gain: You would not
have to relocate to take on this new role, and the workload and travel requirements would be no worse
than what they currently are.]
[Loss – Limited Control: Given the current business climate and mix of available skills, you are fairly
confident that – if this change were considered necessary – you would have little choice in the decision
or in the design of your new job role. There would be limited room to maneuver.]
[Loss – Control: The person to whom you could potentially report is, however, known for being a hands-
off manager, who would likely give you a great deal of freedom to shape the job role and work
autonomously. The person has a well-deserved reputation for creating space for subordinates to operate
independently, and with the combination of the two units, you would have considerable room to
maneuver.]
[Gain – Limited Control: Given the organization’s ambitious growth plans and the mix of available
skills, you are fairly certain that – if you were offered this position – you would have little choice in the
decision. You would be asked to take on this role in the best interest of the organization, and it would be
very hard to turn down the offer.]
[Gain – Control: Given that several new positions are opening up, you would likely have considerable
freedom to choose among other comparable positions – or to stay in your current position – if you were
made an offer and decided to turn it down.]
Your colleague concluded the conversation by emphasizing that no decisions have yet been made and
that various organizational and staffing options are still being considered. [Limited Control: You are,
however, unlikely to have much influence on the decision outcome. / Control: You might still be able to
influence the decision outcome.
- Insert Table 1 about here -
Network Activation Questions
The network activation questions came next in the survey. I gave them the following instructions: “Most
people discuss important matters, such as the situation just described, with others within and outside their
organization. In the boxes below, please list the initials of the people with whom you would discuss this
situation.” Then I asked two standard name generators (Burt 1984): “Who are the people within
[Company] with whom you would discuss this situation?” and “Who are the people outside [Company]
with whom you would discuss this situation?” That is, rather than just asking people to recall ties in
Picturing Uncertainty
13
general, I asked them to make a purposive choice about which actual network ties to activate in response
to a hypothetical situation.
Because prior research has shown that – in self-administered web surveys – the number of names
a respondent provides can be especially sensitive to question wording (e.g., “list up to ten contacts”) and
even display format (i.e., the number of text boxes shown) (Vehovar et al. 2008), I did not prime subjects
with a particular number of names to provide. Instead, I programmed the survey to dynamically adjust the
number of boxes displayed (i.e., one box was initially displayed per question; once subjects began typing
in that box, another box appeared below). Although subjects were not told of the limit, they could in
practice enter up to thirteen initials per question (or 26 initials in total). Only four respondents (2.5% of
subjects) reached either of the two name limits. Following the name generators were a series of name
interpreters about each contact listed, including one about organizational distance (five point scale,
ranging from “same department, function, or operating unit” to “unrelated department, function, or
operating unit”). Finally, the last section of the survey included questions about the respondent’s own
background and personal characteristics (described in greater detail below).
Manipulation Checks
I included three manipulation checks: (1) perceived uncertainty about the situation; (2) perceived control
or lack of control; and (3) perceived gain or loss. For perceived uncertainty, I adapted four items used in
prior research (Caplan et al. 1975): “Based on what you have learned so far about this situation, how
certain are you about…” (1) “…what your specific job responsibilities will be six months from now?” (2)
“…what your future career picture in this organization looks like?” (3) “…how much the financial
rewards you can expect to receive will change?” and (4) “…how much your status in the organization will
change?” Responses could range from 1 (“Not at all certain”) to 4 (“Very certain”). I combined these
items into a composite measure of perceived uncertainty (alpha = 0.84), which could range from four to
sixteen. The mean of this composite measure was 6.60 (standard deviation of 2.63), with 89.2% of
subjects perceiving the situation as more uncertain than certain (i.e., mean of the composite measure less
than or equal to 10). There were no significant differences in perceptions of uncertainty across the four
Picturing Uncertainty
14
conditions (mean difference between loss and gain conditions of .537 (t statistic=1.28, p=.202); mean
difference between the control and limited control conditions of .094 (t statistic=.223, p=.824).
For perceived control or limited control, I adapted four items from prior research (Pearlin and
Schooler 1978): “Based on what you have learned so far about this situation and your expectations about
how it might play out, how likely is it that…” (1) “…you will be able to control what happens to you next
in the organization?” (2) “…what happens to you next in the organization depends mostly on what you
do?” (3) “…you will not be able to influence organizational decisions that relate to you?” and (4) “…you
will have the freedom to choose or design the job role you want?” Responses could range from 1 (“Very
unlikely”) to 4 (“Very likely”). Reverse coding the third item, I constructed a composite measure (alpha =
.78), which could range from 4 to 16 (mean of 10.41; standard deviation of 2.49). Subjects in the control
condition perceived having significantly more influence over the situation than those in the limited
control condition; the mean difference was 1.25 (t statistic=3.24; p=.002). There were no significant
differences in perceived control or lack of control between respondents in the gain and loss conditions.
Finally, for perceived gain or loss, I adapted four items used in prior research (Highhouse et al.
2002): “Based on what you have learned so far about this situation and your expectations about how it
might play out, how likely is it that…” (1) “…the situation will result in a successful outcome for you?”
(2) “…you may lose from this situation and are unlikely to gain?” (3) “…you may gain in this situation
and are unlikely to lose?” and (4) there will be personal loss for you in this situation?” Responses could
range from 1 (“Very unlikely”) to 4 (“Very likely”). Reverse coding the second and fourth items, I
constructed a composite measure (alpha = 0.86), which could range from 4 to 16 (mean of 10.84; standard
deviation of 2.41). Respondents in the gain condition perceived significantly more potential gain in the
situation than those in the loss condition; the mean difference was 2.98 (t statistic=9.85; p=.000).
Respondents in the limited control condition perceived somewhat less gain than those in the control
condition; however, this difference was not statistically significant. Overall, the manipulation checks
indicated that participants’ perceptions of the hypothetical situations were consistent with those intended
in the study design. That is, the experimental conditions made salient each of the intended schemata.
Picturing Uncertainty
15
Response Variables
For Hypotheses 1, 3a, 3b, and 3c, the response variable was the number of ties a person activated. I
considered the total number of ties, as well as the number activated within the organization and outside
the organization. For Hypotheses 2 and 4, the response variable was the proportion of within-organization
ties activated that were bridging. I considered a tie to be bridging if a respondent identified a contact as
being at least a three on the five point organizational distance scale (i.e., “different but somewhat
related…”, “different but loosely related…”, or “different and unrelated department, function, or
operating unit”). I discuss below the robustness of reported findings to alternative measures of
organizational distance. Finally, for Hypothesis 5, the response variable was the proportion of same-sex
ties activated.
Explanatory Variables
To assess the main effects of the treatments, I used two indicator variables – Loss and Limited Control.
For Hypothesis 3c, which involved internal versus external locus of control, I used twelve items from a
validated and widely used scale that is adapted to workplace settings (Spector 1988): (1) “A job is what
you make of it;” (2) “On most jobs, people can pretty much accomplish whatever they set out to
accomplish;” (3) “If employees are unhappy with a decision made by their boss, they should do
something about it;” (4) “Making money is primarily a matter of good fortune;” (5) “Most people are
capable of doing their jobs well if they make the effort;” (6) “Promotions are usually a matter of good
fortune;” (7) “Promotions are given to employees who perform well on the job;” (8) “To make a lot of
money, you have to know the right people;” (9) “It takes a lot of luck to be an outstanding employee on
most jobs;” (10) “People who perform their jobs well generally get rewarded for it;” (11) “Most
employees have more influence on their supervisors than they think they do;” and (12) “The main
difference between people who make a lot of money and people who make a little money is luck.”
Reverse coding items 4, 6, 8, 9, and 12, I constructed a composite work locus of control measure (alpha =
.716). After centering this measure, I interacted it with one of the treatment indicators: Limited Control x
Work Locus of Control.
Picturing Uncertainty
16
Finally, I included various control variables: age, age-squared, sex (indicator set to 1 for female),
ethnicity (indicators for White, Black, and Asian, with Hispanic / Latino and Other serving as the
reference category), country of origin – ex-US (indicator set to 1 for subjects born outside the US),
marital status (indicators for now married and never married, with widowed, divorced, or separated
serving as the reference category), past uncertainty experience (indicator set to 1 for subjects who
reported “having experienced a period of organizational change that was similar to the scenario presented
in this exercise in terms of its potential implications for you”), and small office (to account for the
availability of contacts in the subject’s work setting; this was an indicator set to 1 for subjects whose
primary work location included less than 500 employees).
Estimation
For Hypotheses 1, 3a, 3b, and 3c, for which the response variable is a count of the number of ties
activated, I used the Poisson Quasi-Maximum Likelihood (PQML) estimator. This estimator is consistent
so long as the conditional mean is correctly specified; it makes no assumptions about the conditional
variance or distribution of the data (Wooldridge 1997). For Hypotheses 2, 4, and 5, for which the
response variable is a proportion, I employed the fractional logit estimator (Papke and Wooldridge 1996).
In all cases, inferences were based on robust standard errors.
Results
Table 2 reports descriptive statistics and a correlation matrix. Although I used block random assignment
to the experimental conditions, the number of subjects per cell varied based on response rates to the
survey. 54% of respondents were assigned to the loss (rather than gain) condition, while 51% of
respondents were assigned to limited control (rather than control) condition. It is worth noting that 75% of
respondents reported having experienced a situation like the one described in the manipulation at least
once in the past. Thus, they could presumably relate to the hypothetical scenario and draw on their actual
experience when deciding which actual network ties to activate. Across all four conditions, subjects listed
a mean of 7.06 contacts, 29% of which were bridging and 51% of which were of the same sex. Responses
to the name interpreters – in particular, the frequency of contact – indicate that all contacts listed were
Picturing Uncertainty
17
part of subjects’ pre-existing networks (given that the reported frequency of contact was at least once per
year for all contacts listed). That is, reported ties represented the activation of pre-existing ties, rather than
attempts to form new ties.
- Table 2 about here -
Table 3 reports results pertaining to Hypothesis 1 – that organizational actors will activate more
ties in situations of loss than in situations of gain. The coefficient for Loss was of the expected (positive)
sign and significant in Model 1 (all activated ties), Model 2 (all activated ties, with controls), Model 3
(ties activated within the organization), and Model 4 (ties activated within the organization, with
controls). Loss was not a significant covariate in Models 5 and 6, which considered ties activated outside
the organization. In the models that included control variables, several of these variables were
significantly associated with the response variable. Age was significantly associated with ties activated
outside the organization, with a positive coefficient for the quadratic term and a negative coefficient for
the linear term. White was marginally significant and had a positive coefficient in Model 2 and was
positive and significant in Model 4. Country of Origin – Ex-US was marginally significant and positively
associated with ties activated within the organization. Never Married was marginally significant and
positively associated with ties activated and significant and positively associated with ties activated
within the organization. Finally, perhaps reflecting the availability of contacts in smaller offices, Small
Office was negatively associated with ties activated and ties activated outside the organization. The
inclusion of these control variables did not, however, materially change the significance or magnitude of
Loss. Figure 1 depicts the size of this effect. Subjects in the gain condition activated a mean of 6.23 ties,
3.60 of which were within the organization and 2.63 of which were outside the organization. Subjects in
the loss condition activated 7.78 ties, 4.71 of which were within the organization and 3.07 of which were
outside the organization. Overall, subjects in the loss condition activated nearly 25% more ties than those
in the gain condition. These results provide support for Hypothesis 1.
- Table 3 about here –
- Figure 1 about here -
Picturing Uncertainty
18
Results pertaining to Hypothesis 2 – that organizational actors will activate a greater proportion
of bridging ties in situations of loss than in situations of gain – are reported in Table 4. Loss was positive
and significant in both Models 7 and 8. In the model with control variables, only one – Never Married –
was statistically significant; it was positively associated with the proportion of ties activated that were
bridging. The inclusion of control variables did not change the significance, but did result in a slight
decline in the magnitude, of Loss. The effect size is depicted in Figure 2: 22% of ties activated by subjects
in the gain condition bridged internal organizational boundaries, while 35% of ties activated by subjects
in the loss condition were bridging. Because the designation of a tie as bridging was somewhat arbitrary
(i.e., at least a three on a five point scale of organizational distance), I implemented an alternative
analytical approach as a robustness check. Using ordinary least squares, I regressed the mean
organizational distance of ties activated on Loss. The results were comparable – i.e., under conditions of
loss, the mean organizational distance of ties activated was significantly greater than under condition of
gain.6 These results lend support for Hypothesis 2.
- Table 4 about here –
- Figure 2 about here -
Results concerning Hypotheses 3a (that organizational actors will activate fewer ties in situations
of limited control than in situations of control), 3b (that organizational actors will activate more ties in
situations of limited control than in situations of control), and 3c (that internal, rather than external, locus
of control will amplify the tendency to activate more ties in situations of limited control than in situations
of control) are reported in Table 5. Limited Control was not significantly associated with the number of
ties activated in Model 9 (without control variables), nor was it significant in Model 10 (with control
variables). Thus, neither Hypothesis 3a nor Hypothesis 3b was supported. In Models 11 (without control
variables) and 12 (with control variables), the interaction of limited control with internal locus of control
was, however, significant and had a positive coefficient. That is, when confronted with a situation of
limited control, those with a high internal locus of control were apt to activate a larger number of ties than
6Results of all supplemental analyses are available upon request from the author.
Picturing Uncertainty
19
those with a low internal locus of control. Figure 3 depicts the size of this effect using a median split of
the sample on the composite internal locus of control measure. Those in the lower half of the distribution
activated 7.82 ties under conditions of control and 6.42 ties under conditions of limited control, while
those in the upper half of the distribution activated 6.50 ties under conditions of control and 7.84 ties
under conditions of limited control. That is, those with a more internal locus of control activated 22%
more ties under conditions of limited control than in conditions of control, whereas those with a more
external locus of control activated 17% fewer ties under conditions of limited control than in conditions of
control. Thus, there is support for Hypothesis 3c.
- Table 5 about here –
- Figure 3 about here -
Table 6 reports results that speak to Hypothesis 4, which suggests that organizational actors will
activate a lower proportion of organizationally distant ties in situations of limited control than in
situations of control. Limited Control was not statistically significant in Models 13-16, including
specifications that added control variables, internal locus of control, and the interaction of limited control
and locus of control. Thus, there is no support for Hypothesis 4.
- Table 6 about here –
Finally, Table 7 addresses Hypothesis 5 – that both males and females will activate a greater
proportion of ties to male colleagues in situations of limited control relative to situations of control. In
Model 17, the main effect of limited control was not statistically significant. However, the interaction
term – Limited Control x Female – was marginally significant in Model 18 (without control variables)
and significant in Model 19 (with control variables). In both cases, it was of the expected (i.e., negative)
sign. Figure 4 depicts this effect. For males, 51% of ties activated were of the same sex under conditions
of control, while 56% of ties activated were of the same sex under conditions of limited control. This
pattern was reversed for females: 54% of ties activated were of the same sex under conditions of control,
while 45% of ties activated were of the same sex under conditions of limited control. Thus, Hypothesis 5
was supported – though the interaction term was marginally significant in the absence of control variables
Picturing Uncertainty
20
(which presumably helped reduce error variance). As a supplemental analysis, I also considered whether
work locus of control moderated this effect. As Table 7 further indicates, there was weak support for this
moderating effect. In Model 20, the three-way interaction term – Limited Control x Female x Work Locus
of Control – was marginally significant (p<.10) and of the expected (i.e., negative) sign. That is, the
tendency for females to activate a greater proportion of ties to males was more pronounced among those
females with a high internal locus of control; however, this effect was not robust to the inclusion of
control variables in Model 21.
- Table 7 about here –
- Figure 4 about here –
DISCUSSION AND CONCLUSION
This article has sought to clarify how the frames and schemata through which organizational actors
interpret and respond to uncertain events can shape the mobilization of social resources through
workplace networks. It examines, in particular, network activation – or the conversion of latent social ties
into active relationships – when a widely used uncertainty frame – threat / opportunity – is operative in
the minds of managers (Dutton and Jackson 1987; Jackson and Dutton 1988). The threat / opportunity
frame can make salient two distinct pairs of schemata: gain / loss and control / limited control (George et
al. 2006; Ocasio 1995). Three theories – loss aversion (Kahneman and Tversky 1979), threat-rigidity
(Staw et al. 1981), and reactance (Brehm 1966; Wortman and Brehm 1975) – inform expectations about
the size and composition of network ties that actors will activate when facing situations that invoke these
schemata. Results from an experimental study involving 158 people working in a non-profit health
company indicate that a loss (rather than gain) schema led participants to activate more ties and a greater
proportion of bridging ties. That is, the predictions informed by loss aversion were supported. By
contrast, there was no support for the predictions of threat-rigidity theory: a limited control (rather than
control) schema did not lead subjects to activate fewer ties or a lower proportion of bridging ties. There
was conditional support for the predictions of reactance theory. Although the limited control (rather than
Picturing Uncertainty
21
control) schema did not lead actors to activate more ties in general, responses to this schema were
moderated by an individual difference construct – the locus of control. Subjects with an internal locus of
control activated more ties in situations of limited control relative to situations of control, whereas those
with an external locus of control had the opposite response. Finally, a limited control (rather than control)
schema led both females and males to activate a greater proportion ties to male colleagues.
I turn next to addressing two outstanding questions raised by this investigation. First, when the
gain / loss and control / limited control schemata are salient at the same time, do they have a
multiplicative effect on activation choices? Supplemental analyses suggest a positive interaction between
loss and limited control on the number of ties activated but not on the proportion of organizationally
distant ties or the proportion of same-sex ties activated. Second, given prior research suggesting value of
assessing the content, not just the structure, of network ties (Podolny and Baron 1997), I also considered
the resources (i.e., information, influence, career advice, and social / emotional support) that people
sought in the different uncertainty conditions. There were no discernible differences across treatment
conditions in the size or composition of networks activated to obtain these different resources (perhaps
because respondents could report that they would seek multiple resources from a given contact and most
ties reported were multiplex).
This study makes four noteworthy contributions. First, it contributes to the growing body of
research on social capital activation (Hurlbert et al. 2000; Renzulli and Aldrich 2005; Smith 2000, 2005;
Srivastava 2012). In particular, it builds on and extends prior work suggesting that different situations can
lead people to cognitively recall different subsections of their networks. As Smith, Menon, and Thompson
(2012: 78-79) argue: “[The cognitive recall] of ties is a precondition to mobilization, whereby people who
activate different microstructures might therefore access different resources.” Yet their experimental
protocol considered only the recall of ties – not the purposive choice to activate ties. Findings from the
present study provide the empirical bridge between, on one hand, the frames and schemata made salient
by situations of threat and opportunity and, on the other hand, the mobilization of resources through
social networks. That is, the different situations people encounter can not only influence which ties actors
Picturing Uncertainty
22
recall but also their purposive choice to activate network ties in pursuit of valuable social resources. In
addition, the study considers not only situations of threat but also of opportunity. It conceptually unpacks
threat and opportunity into two distinct pairs of schemata – gain / loss and control / limited control – and
shows that network actions taken in response to threat or opportunity depend on which of these pairs is
operative in a given situation. Thus, these results provide a more complete picture of how situations of
threat or opportunity influence both cognitive recall and purposive choices of network activation.
Next, these findings expand our understanding of how individual differences – such as self-
monitoring orientation (Mehra et al. 2001; Mehra and Schenkel 2008; Sasovova et al. 2010), relational
self-construal (O'Connor and Sauer 2006), need for cognition (Anderson 2008), and network entrepreneur
personality (Burt et al. 1998) – can influence interpersonal network dynamics. This study represents, to
the best of my knowledge, the first to investigate the role of another well-established construct – locus of
control – in shaping network action. People with an internal (rather than external) locus of control
responded to situations of limited control by more vigorously activating social ties. This result
underscores the need for more nuanced theoretical accounts of cognition and networks, which take into
consideration the role of these various stable individual differences (cf. Zerubavel 1997).
At a more macro level, the study contributes to the growing body of research that examines the
divergence between formal and informal organization during periods of restructuring (Gulati and
Puranam 2009; Hannan et al. 2003a, b; Nickerson and Zenger 2002) and its implications for performance
(Krackhardt and Stern 1988). The present study identifies a novel mechanism – social capital activation
when the schemata of gain / loss are operative in people’s minds – that can give rise to divergence
between formal and informal structure. The loss of tangible resources associated with restructuring led
people to activate a greater proportion of ties outside of their subunit. That is, whereas prior research has
shown that interpersonal networks in workplace settings tend to hew to the formal organizational structure
(Han 1996; Hinds and Kiesler 1995; Ibarra 1992b), this study suggests the formal structure imposes a
relatively weak constraint on interpersonal networks during events that entail potential gain or loss for
employees. Thus, network activation may importantly contribute to the previously theorized and
Picturing Uncertainty
23
documented lag in adjustment between formal and informal structure following a restructuring event. This
divergence may also have implications for organizational performance. A prior simulation study, for
example, indicates that organizations endowed with a large proportion of ties that bridge formal subunits
perform better during tumultuous times (Krackhardt and Stern 1988). The present study provides
suggestive evidence that the proportion of activated bridging ties in an organization can endogenously
change during uncertain events. It remains to be explored whether individual-level choices to activate
bridging ties when facing uncertain losses can cumulate to affect organizational resilience.
Finally, this study informs our understanding of sex differences in workplace networks (Brass
1985; Ibarra 1992a, 1997; Kleinbaum, Stuart, and Tushman 2012; Miller 1986; Moore 1990; Olson and
Miller 1983; Ridgeway 1997; Ridgeway and Correll 2003). Whereas this literature has tended to take a
static view of structural differences in the workplace networks of males versus females, this study draws
attention to differences in the network actions that males and females take during periods of uncertainty,
when power and organizational resources are often contested (Pfeffer 1992). For example, based on a
cross-sectional network survey, Ibarra (1992a) concludes that males tend to build homophilous networks
for both instrumental and expressive purposes, while females tend to form ties to female colleagues for
expressive purposes and to male colleagues for instrumental purposes. The present study suggests the
need to complicate this account by considering the role of situations that organizational actors face and
their choices about how to use the network resources available to them. Under conditions of limited
control (relative to control), females tended to reach out to males, while males appeared to close ranks
with other males. Because the research design entailed random assignment and therefore effectively
controlled for the opportunity structure for interaction, this effect likely represents situational differences
in preferences for contact. To the extent that these patterns generalize across organizational settings, they
may help account for the fact that times of organizational change – when people may feel they have
limited control over their careers and organizational lives – can often produce increases in sex-based
inequality, even when the opportunities for discrimination are attenuated (Acker 1992, 2006; Dencker
2008; DiPrete and Nonnemaker 1997). Sex-differences in network activation during periods of
Picturing Uncertainty
24
organizational change therefore represent a plausible mechanism that can partially account for the
persistence of gender inequality in the workplace (Reskin 2003).
The study is not without limitations, which also point to directions for future research. First, prior
research suggests that pre-existing network structure can enable and constrain network change (Uehara
1990). In this study, individual differences in pre-existing network structure were accounted for through
random assignment. Still, a fruitful avenue for future research would entail first mapping subjects’ pre-
existing networks and then assessing how network structure conditions network activation choices.
Second, this study design did not allow one to distinguish whether the activation choices people made
were functional (e.g., seeking career advice from trusted mentors) or dysfunctional (e.g., spreading false
rumors) for the individual or the organization. Future designs should consider such consequences of
network activation. Third, this study took the perspective of the help seeker but not of the help provider.
Prior research has importantly highlighted that the choices of help providers – i.e., the targets of network
activation – can importantly determine the nature and quality of resources that accrue to the help seeker
(Smith 2005). How do the frames and schemata used by the prospective help provider influence his or her
willingness to share resources with the help seeker? Finally, uncertain events sometimes do not fall neatly
into buckets of threat or opportunity; rather, some situations are ambiguous (Cacioppo and Berntson
1994; Plambeck and Weber 2009). Moreover, interpretations of threat and opportunity vary across
cultural settings (Barr and Glynn 2004). Future research can profitably examine how people activate
networks when operating under ambiguous frames and in different cultural contexts.
In sum, this article demonstrates the value of deepening the exploration of managerial cognition
during times of organizational change and clarifying its link to social network dynamics in the workplace.
It also highlights the promise of using field-based experimental methods in unearthing the
interrelationships between cognition and social structure.
Picturing Uncertainty
25
References
Acker, J. 1992. The Future of Women and Work: Ending the Twentieth Century Sociological
Perspectives 35(1) 53-68.
Acker, J. 2006. Inequality Regimes: Gender, Class, and Race in Organizations. Gender and
Society 20(4) 441-464.
Anderson, M.H. 2008. Social Networks and the Cognitive Motivation to Realize Network
Opportunities: A Study of Managers' Information Gathering Behaviors. J. of Organ.
Behavior 29(1) 51-78.
Barr, P.S., M.A. Glynn. 2004. Cultural Variations in Strategic Issue Interpretation: Relating
Cultural Uncertainty Avoidance to Controllability in Discriminating Threat and
Opportunity. Strategic Management J. 25(1) 59-67.
Bechky, B.A. 2003. Sharing meaning across occupational communities: the transformation of
understanding on a production floor. Organ. Sci. 14(3) 312-330.
Billings, R.S., T.W. Milburn, M.L. Schaalman. 1980. A Model of Crisis Perception: A
Theoretical and Empirical Analysis. Admin. Sci. Quart. 25(2) 300-316.
Brands, R., M. Kilduff. 2012. Gender Stereotyping in the Perception of Social Network Roles:
Consequences for Individuals and Groups Acad. of Management Conference Presentation.
Brass, D.J. 1985. Men's and Women's Networks: A Study of Interaction Patterns and Influence
in an Organization. Acad. of Management J. 28(2) 327-343.
Brehm, J.W. 1966. Response to Loss of Freedom: A Theory of Psychological Reactance.
Academic Press, New York.
Burkhardt, M.E., D.J. Brass. 1990. Changing Patterns or Patterns of Change: The Effects of a
Change in Technology on Social Network Structure and Power. Admin. Sci. Quart. 35(1)
104-127.
Burt, R.S. 1984. Network Items and the General Social Survey. Soc. Networks 6(4) 293-339.
Burt, R.S., J.E. Jannotta, Jr., J.T. Mahoney. 1998. Personality Correlates of Structural Holes.
Soc. Networks 20(1) 63-87.
Cacioppo, J.T., G.G. Berntson. 1994. Relationship between Attitudes and Evaluative Space: A
critical review, with Emphasis on the Separability of Positive and Negative Substrates.
Psychological Bulletin 115(3) 401-423.
Caplan, R.D., S. Cobb, J.R.P. French, Jr., R.V. Van Harrison, S.R. Pinneau, Jr. 1975. Job
Demands and Worker Health: Main Effects and Occupational Differences U.S. Department
of Health, Education, and Welfare., Washington, DC.
Carlile, P.R. 2004. Transferring, Translating, and Transforming: An Integrative Framework for
Managing Knowledge Across Boundaries. Organ. Sci. 15(5) 555.
Casciaro, T., M.S. Lobo. 2008. When Competence Is Irrelevant: The Role of Interpersonal
Affect in Task-Related Ties. Admin. Sci. Quart. 53(4) 655-684.
Cerulo, K.A. 2002. Culture in Mind: Toward a Sociology of Culture and Cognition Routledge,
New York.
Chattopadhyay, P., W.H. Glick, G.P. Huber. 2001. Organizational Actions in Response to
Threats and Opportunities. Acad. of Management J. 44(5) 937-955.
Child, J. 1997. Strategic Choice in the Analysis of Action, Structure, Organizations, and
Environment: Retrospect and Prospect. Organ. Studies 18(1) 43-76.
Dencker, J.C. 2008. Corporate Restructuring and Sex Differences in Managerial Promotion.
Amer. Sociological Rev. 73 455-476.
Picturing Uncertainty
26
DiPrete, T.A., K.L. Nonnemaker. 1997. Structural Change, Labor Market Turbulence, and
Labor Market Outcomes. Amer. Sociological Rev. 62 386-404.
Dougherty, D. 1992. Interpretive Barriers to Successful Product Innovation in Large Firms.
Organ. Sci. 3(2) 179-202.
Dutton, J.E., S.E. Jackson. 1987. Categorizing Strategic Issues: Links to Organizational
Action. Acad. of Management Rev. 12(1) 76-90.
Elder, G.H., J.K. Liker. 1982. Hard Times in Women's Lives: Historical Influences across
Forty Years. Amer. J. Sociology 88 241-266.
Elster, J. 2007. Explaining Social Behavior: More Nuts and Bolts for the Social Sciences.
Cambridge University Press, New York.
Emirbayer, M., A. Mische. 1998. What is Agency? Amer. J. Sociology 103(4) 962-1023.
Fiske, S.T., E. Dépret. 1996. Control, Interdependence and Power: Understanding Social
Cognition in Its Social Context. Eur. Rev. of Soc. Psychology 7(1) 31-61.
Friedkin, N.E. 1982. Information Flow Through Strong and Weak Ties in lntraorganizational
Social Networks. Soc. Networks 3 273-285.
George, E., P. Chattopadhyay, S.B. Sitkin, J. Barden. 2006. Cognitive Underpinnings of
Institutional Persistence and Change: A Framing Perspective. Acad. of Management Rev.
31(2) 347-365.
Gladstein, D.L., N.P. Reilly. 1985. Group Decision Making Under Threat: The Tycoon Game.
Acad. of Management J. 28(3) 613-627.
Goffman, E. 1974. Frame Analysis: An Essay on the Organization of the Experience. Harper &
Row, New York.
Gorman, E.H. 2006. Work Uncertainty and the Promotion of Professional Women: The Case of
Law Firm Partnership. Soc. Forces 85(2) 865-890.
Gulati, R., P. Puranam. 2009. Renewal through Reorganization: The Value of Inconsistencies
between Formal and Informal Organization. Organ. Sci. 20(2) 422-440.
Han, S.-K. 1996. Structuring Relations in On-the-Job Networks. Soc. Networks 18 47-67.
Hannan, M.T., L. Polos, G.R. Carroll. 2003a. Cascading Organizational Change. Organ. Sci.
14(5) 463-482.
Hannan, M.T., L. Polos, G.R. Carroll. 2003b. The Fog of Change: Opacity and Asperity in
Organizations. Admin. Sci. Quart. 48(3) 399-432.
Highhouse, S., S. Mohammed, J.R. Hoffman. 2002. Temporal Discounting of Strategic Issues:
Bold Forecasts for Opportunities and Threats. J. of Basic and Applied Psychology 24(1) 43-
56.
Hinds, P., S. Kiesler. 1995. Communication across Boundaries: Work, Structure, and Use of
Communication Technologies in a Large Organization. Organ. Sci. 6(4) 373-393.
Hitlin, S., G.H. Elder, Jr. 2007. Time, Self, and the Curiously Abstract Concept of Agency.
Sociological Theory 25(2) 170-191.
Hurlbert, J.S., V.A. Haines, J.J. Beggs. 2000. Core Networks and Tie Activation: What Kinds
of Routine Networks Allocate Resources in Nonroutine Situations? Amer. Sociological Rev.
65(4) 598-618.
Ibarra, H. 1992a. Homophily and Differential Returns: Sex-Differences in Network Structure
and Access in an Advertising Firm. Admin. Sci. Quart. 37(3) 422-447.
Ibarra, H. 1992b. Structural Alignments, Individual Strategies, and Managerial Action. N.
Nohria, R.G. Eccles, eds. Networks and Organizations: Structure, Form, and Action.
Harvard Business School Press, Boston, MA, 165-188.
Picturing Uncertainty
27
Ibarra, H. 1997. Paving an Alternative Route: Gender Differences in Managerial Networks.
Soc. Psychology Quart. 60(1) 91-102.
Jackson, S.E., J.E. Dutton. 1988. Discerning Threats and Opportunities. Admin. Sci. Quart.
33(3) 370-387.
Kahneman, D. 1973. Attention and Effort. Prentice-Hall, Englewood Cliffs, NJ.
Kahneman, D., A. Tversky. 1979. Prospect Theory: An Analysis of Decision under Risk.
Econometrica 47(2) 263-292.
Kleinbaum, A., T.E. Stuart, M. Tushman. 2012. Discretion within the Constraints of
Opportunity: Gender Homophily and Structure in a Formal Organization. Tuck School of
Business Working Paper 2011-90.
Krackhardt, D., R.N. Stern. 1988. Informal Networks and Organizational Crises: An
Experimental Simulation. Soc. Psychology Quart. 51(2) 123-140.
Lincoln, J.R., J. Miller. 1979. Work and Friendship Ties in Organizations: A Comparative
Analysis of Relational Networks. Admin. Sci. Quart. 24(2) 181-199.
March, J., H. Simon. 1958. Organizations. Wiley, New York.
Marsden, P.V., K.E. Campbell. 1984. Measuring Tie Strength. Soc. Forces 63(2) 482-501.
McDonald, M.L., J.D. Westphal. 2003. Getting By With the Advice of Their Friends: CEOs'
Advice Networks and Firms' Strategic Responses to Poor Performance. Admin. Sci. Quart.
48(1) 1-32.
McEvily, B., A. Zaheer. 1999. Bridging Ties: A Source of Firm Heterogeneity in Competitive
Capabilities. Strategic Management J. 20(12) 1133-1156.
McLean, P.D. 1998. A Frame Analysis of Favor Seeking in the Renaissance: Agency,
Networks, and Political Culture. Amer. J. Sociology 104(1) 51-91.
Mehra, A., M. Kilduff, D.J. Brass. 2001. The Social Networks of High and Low Self-Monitors:
Implications for Workplace Performance. Admin. Sci. Quart. 46(1) 121-146.
Mehra, A., M.T. Schenkel. 2008. The Price Chameleons Pay: Self-Monitoring, Boundary
Spanning and Role Conflict in the Workplace. British J. of Management 19(2) 138-144.
Miller, J. 1986. Pathways in the Workplace: The Effects of Gender and Race on Access to
Organizational Resources. Cambridge University Press, London.
Mische, A. 2009. Projects and Possibilities: Researching Futures in Action. Sociological
Forum 24(3) 694-704.
Mizruchi, M.S., L.B. Stearns. 2001. Getting Deals Done: The Use of Social Networks in Bank
Decision-Making. Amer. Sociological Rev. 66(5) 647-671.
Moore, G. 1990. Structural Determinants of Men's and Women's Personal Networks. Amer.
Sociological Rev. 55(5) 726-735.
Ng, T.W.H., K.L. Sorensen, L.T. Eby. 2006. Locus of Control at Work: A Meta-Analysis. J. of
Organ. Behavior 27 1057–1087.
Nickerson, J.A., T.R. Zenger. 2002. Being Efficiently Fickle: A Dynamic Theory of
Organizational Choice. Organ. Sci. 13(5) 547-566.
O'Connor, K.M., S.J. Sauer. 2006. Recognizing Social Capital in Social Networks:
Experimental Results. Johnson School Research Paper Series No. 18-06
Ocasio, W. 1995. The Enactment of Economic Diversity: A Reconciliation of Theories of
Failure-Induced Change and Threat-Rigidity. L.L. Cummings, B.M. Staw, eds. Research in
Organizational Behavior. JAI Press, Greenwich, CT, 287-331.
Picturing Uncertainty
28
Olson, J., J. Miller. 1983. Gender and Interaction in the Workplace. H.Z. Lopata, J.H. Pleck,
eds. Research in the Interweave of Social Roles: Families and Jobs. JAI Press, Greenwich,
CT, 35-58.
Papke, L.E., J.M. Wooldridge. 1996. Econometric Methods for Fractional Response Variables
with an Application to 401(K) Plan Participation Rates. J. of Applied Econometrics 11(6)
619-632.
Pearlin, L.I., C. Schooler. 1978. The Structure of Coping. J. of Health and Soc. Behavior 19(1)
2-21.
Pescosolido, B.A. 1992. Beyond Rational Choice: The Social Dynamics of How People Seek
Help. Amer. J. Sociology 97(4) 1096-1138.
Pettit, N.C., K. Yong, S.E. Spataro. 2010. Holding Your Place: Reactions to the Prospect of
Status Gains and Losses. J. of Experimental Soc. Psychology 46(2) 395-401.
Pfeffer, J. 1992. Managing with Power: Politics and Influence in Organizations. Harvard
Business School Press, Boston, MA.
Pfeffer, J., G.R. Salancik, H. Leblebici. 1976. The Effect of Uncertainty on the Use of Social
Influence in Organizational Decision Making. Admin. Sci. Quart. 21(2) 227-245.
Plambeck, N., K. Weber. 2009. CEO Ambivalence and Responses to Strategic Issues. Organ.
Sci. 20(6) 993-1010.
Podolny, J.M., J.N. Baron. 1997. Resources and Relationships: Social Networks and Mobility
in the Workplace. Amer. Sociological Rev. 62 673-693.
Renzulli, L.A., H. Aldrich. 2005. Who Can You Turn to? Tie Activation within Core Business
Discussion Networks. Soc. Forces 84(1) 323-341.
Reskin, B.F. 2003. Including Mechanisms in our Models of Ascriptive Inequality. Amer.
Sociological Rev. 68 1-21.
Reskin, B.F., D.B. McBrier, J.A. Kmec. 1999. The Determinants and Consequences of
Workplace Sex and Race Composition. Annual Rev. of Sociology 25 335-361.
Ridgeway, C.L. 1997. Interaction and the Conservation of Gender Inequality: Considering
Employment. Amer. Sociological Rev. 62(2) 218-235.
Ridgeway, C.L., S.J. Correll. 2003. Expectation states theory. J. Delamater, ed. The Handbook
of Social Psychology. Kluwer Academic/Plenum, New York, 29-52.
Ridgeway, C.L., L. Smith-Lovin. 1999. The Gender System and Interaction. Annual Rev. of
Sociology 25(1) 191-216.
Ridgeway, C.L., L. Smith-Lovin. 2006. Gender and Interaction. J.S. Chafetz, ed. Handbook of
the Sociology of Gender. Springer, New York, 247-274.
Ross, C.E., J. Mirowsky, S. Pribesh. 2001. Powerlessness and the Amplification of Threat:
Neighborhood Disadvantage, Disorder, and Mistrust. Amer. Sociological Rev. 66(4) 568-
591.
Rotter, J.B. 1954. Social Learning and Clinical Psychology. Prentice Hall, Englewood Cliffs,
NJ.
Rotter, J.B. 1966. Generalized Expectancies for Internal versus External Control of
Reinforcement. Psychological Monographs 80(Whole No. 609).
Sasovova, Z., A. Mehra, S.P. Borgatti, M.C. Schippers. 2010. Network Churn: The Effects of
Self-Monitoring Personality on Brokerage Dynamics. Admin. Sci. Quart. 55 639-670.
Seeman, M., T.E. Seeman. 1983. Health Behavior and Personal Autonomy: A Longitudinal
Study of the Sense of Control in Illness. Journal of Health and Soc. Behavior 24 44-59.
Picturing Uncertainty
29
Shah, P.P. 2000. Network Destruction: The Structural Implications of Downsizing. Acad. of
Management J. 43(1) 101-112.
Smith, E.B., T. Menon, L. Thompson. 2012. Status Differences in the Cognitive Activation of
Social Networks. Organ. Sci. 23(1) 67-82.
Smith, S.S. 2000. Mobilizing Social Resources: Race, Ethnic, and Gender Differences in
Social Capital and Persisting Wage Inequalities. The Sociological Quarterly 41(4) 509-537.
Smith, S.S. 2005. 'Don't Put My Name On It': Social Capital Activation and Job-Finding
Assistance among the Black Urban Poor. Amer. J. Sociology 111(1) 1-57.
Spector, P.E. 1988. Development of the Work Locus of Control Scale. J. of Occupational
Psychology 61(4) 335-340.
Srivastava, S.B. 2012. Social Capital Activation during the Uncertainty of Organizational
Restructuring. University of California - Berkeley Working Paper.
Staw, B.M., L.E. Sandelands, J.E. Dutton. 1981. Threat-Rigidity Effects in Organizational
Behavior: A Multilevel Analysis. Admin. Sci. Quart. 26 501-524.
Stevenson, W.B., D. Greenburg. 2000. Agency and Social Networks: Strategies of Action in a
Social Structure of Position, Opposition, and Opportunity. Admin. Sci. Quart. 45(4) 651-
678.
Taylor, S.E. 1983. Adjustment to Threatening Events: A Theory of Cognitive Adaptation.
Amer. Psychologist 38(11) 1161-1173.
Tortoriello, M., D. Krackhardt. 2010. Activating Cross-Boundary Knowledge: The Role of
Simmelian Ties in the Generation of Innovations. Acad. of Management J. 53(1) 167-181.
Tsai, W. 2002. Social Structure of "Coopetition" within a Multiunit Organization:
Coordination, Competition, and Intraorganizational Knowledge Sharing. Organ. Sci. 13(2)
179-190.
Tversky, A., D. Kahneman. 1990. Rational Choice and the Framing of Decisions. K.S. Cook,
M. Levi, eds. The Limits of Rationality. The University of Chicago Press, Chicago, IL.
Uehara, E. 1990. Dual Exchange Theory, Social Networks, and Informal Social Support. Amer.
J. Sociology 96(3) 521-557.
Vehovar, V., K.L. Manfreda, G. Koren, V. Hlebec. 2008. Measuring Ego-Centered Social
Networks on the Web: Questionnaire Design Issues Soc. Networks 30(3) 213-222.
Wheaton, B. 1980. The Sociogenesis of Psychological Disorder: An Attributional Theory.
Journal of Health and Soc. Behavior 21 100-124.
Wicklund, R.A. 1974. Freedom and Reactance. Lawrence Erlbaum, Oxford, England.
Wooldridge, J.M. 1997. Quasi-Likelihood Methods for Count Data. M.H. Pesaran, P. Schmidt,
eds. Handbook of Applied Econometrics. Blackwell, Oxford, UK.
Wortman, C.B., J.W. Brehm. 1975. Responses to Uncontrollable Outcomes: An Integration of
Reactance Theory and the Learned Helplessness Model. L. Berkowitz, ed. Advances in
Experimental Social Psychology. Academic Press, New York, 278-336.
Zerubavel, E. 1997. Social Mindscapes: An Invitation to Cognitive Sociology. Harvard
University Press, Cambridge, MA.
Picturing Uncertainty
30
Figures and Tables
Figure 1
Figure 2
6.23
3.60
2.63
7.78
4.71
3.07
02
46
8
Num
ber
of T
ies
Condition - Gain Condition - Loss
N=158
Ties Activated - by Gain / Loss
Ties Activated Ties Activated - Within Org.
Ties Activated - Outside Org.
0.22
0.35
0.1
.2.3
.4
Pro
po
rtio
n
Condition - Gain Condition - LossN=148
Proportion of Ties that are Bridging - by Gain / Loss
Picturing Uncertainty
31
Figure 3
Figure 4
7.82
6.42 6.50
7.84
02
46
8
Num
ber
of T
ies
Lower Int. Locus of Control Higher Int. Locus of Control
Control Limited Control Control Limited Control
N=158
Ties Activated - Control / Limited Control x Locus of Control
0.51
0.560.54
0.45
0.2
.4.6
Pro
po
rtio
n
Male Female
Control Limited Control Control Limited Control
N=157
Proportion of Same Gender Ties - Control x Gender
Picturing Uncertainty
32
Table 1: Overview of Manipulation
Loss Gain
Control Uncertain Threat of Downward
Mobility
Freedom to Shape Job Role and
Potential to Influence Decision
Outcome
Uncertain Opportunity for
Upward Mobility
Multiple Available Job Roles
and Considerable Freedom to
Choose Among Them
Limited Control Uncertain Threat of Downward
Mobility
Limited Influence over Job Role
or Decision Outcome
Uncertain Opportunity for
Upward Mobility
One Available Job Role and
Limited Influence over Decision
Outcome
Picturing Uncertainty
33
Table 2: Descriptive Statistics and Correlation Matrix (N=158)
Mean S.D. (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16)
(1) Ties
Activated
7.06 4.67 1.0
(2) Prop.
Bridging
0.29 0.33 0.2 1.0
(3) Prop.
Same Gen.
0.51 0.25 0.3 0.0 1.0
(4) Loss 0.54 0.50 0.2 0.2 0.0 1.0
(5) Lim.
Control
0.51 0.50 -0.0 0.1 -0.0 0.0 1.0
(6) Age 50.38 6.17 -0.0 -0.2 0.1 0.1 -0.0 1.0
(7) Female 0.47 0.50 0.0 0.1 -0.1 -0.0 0.0 0.1 1.0
(8) White 0.72 0.45 0.1 -0.1 0.2 0.1 -0.0 0.2 -0.1 1.0
(9) Black /
Af. Amer.
0.09 0.29 -0.0 0.1 0.0 0.0 0.1 -0.0 -0.0 -0.5 1.0
(10) Asian 0.12 0.33 -0.1 0.0 -0.2 -0.1 0.0 -0.2 0.1 -0.6 -0.1 1.0
(11) Born
– Ex. US
0.16 0.37 0.1 -0.0 0.0 -0.1 0.0 -0.2 0.0 -0.3 0.0 0.3 1.0
(12) Now
Married
0.78 0.41 -0.1 -0.2 -0.1 -0.1 0.0 -0.1 -0.1 -0.0 0.0 0.1 0.1 1.0
(13) Never
Married
0.09 0.29 0.2 0.3 0.1 0.1 -0.0 0.0 0.0 0.1 -0.0 -0.1 -0.0 -0.6 1.0
(14) Past
Experience
0.75 0.43 0.1 -0.1 0.0 -0.2 0.1 0.1 0.2 0.0 0.1 -0.1 -0.1 -0.1 -0.0 1.0
(15) Small
Office
0.39 0.49 -0.2 0.0 -0.1 0.0 0.0 -0.0 -0.1 -0.1 -0.0 0.1 -0.0 -0.1 0.0 -0.1 1.0
(16) Work
Locus Of
Control
0.00 6.00 0.1 0.0 0.0 0.1 -0.1 0.1 -0.1 0.1 -0.1 -0.2 -0.1 -0.1 0.0 -0.1 -0.2 1.0
Picturing Uncertainty
Table 3: PQML Regression of Ties Activated on Treatment (Loss) and Controls
Model 1:
All Ties
Activated
Model 2:
All Ties
Activated
Model 3:
Ties
Activated
Within Org.
Model 4:
Ties
Activated
Within Org.
Model 5:
Ties
Activated
Outside
Org.
Model 6:
Ties
Activated
Outside
Org.
Loss 0.221* 0.227* 0.267** 0.275* 0.155 0.158
(0.098) (0.101) (0.102) (0.108) (0.120) (0.124)
Age -0.106 -0.012 -0.235*
(0.095) (0.107) (0.103)
Age - Sq. 0.001 0.000 0.002*
(0.001) (0.001) (0.001)
Female -0.012 0.022 -0.061
(0.101) (0.104) (0.118)
White 0.349+ 0.435* 0.234
(0.204) (0.220) (0.235)
Black 0.204 0.100 0.333
(0.249) (0.271) (0.283)
Asian 0.189 0.203 0.173
(0.237) (0.252) (0.273)
Born – Ex. US 0.207 0.292+ 0.084
(0.151) (0.153) (0.186)
Now Married -0.007 0.048 -0.090
(0.168) (0.177) (0.189)
Never Married 0.341+ 0.416* 0.233
(0.200) (0.202) (0.237)
Past Exp. 0.189 0.160 0.231
(0.123) (0.127) (0.149)
Small Office -0.232* -0.157 -0.345**
(0.100) (0.100) (0.128)
Constant 1.830*** 4.106+ 1.282*** 0.968 0.967*** 6.800**
(0.062) (2.384) (0.070) (2.695) (0.077) (2.591)
Chi2 5.046 25.534 6.921 28.877 1.658 21.600
prob>Chi2 .025 .012 .009 .004 .198 .042
AIC 1010 981 773 763 677 670
N 158 158 158 158 158 158
+ p<0.10, * p<0.05, ** p<0.01, *** p<.001; two-tailed tests; robust std. errors in parentheses.
Picturing Uncertainty
Table 4: Fractional Logit Regression of
Proportion Bridging Ties on Treatment (Loss) and Controls
Model 7 Model 8
Loss 0.653* 0.614*
(0.259) (0.275)
Age 0.192
(0.263)
Age - Sq. -0.002
(0.003)
Female 0.253
(0.273)
White -0.201
(0.609)
Black 0.424
(0.708)
Asian 0.028
(0.671)
Born – Ex. US -0.396
(0.394)
Now Married -0.248
(0.406)
Never Married 1.083*
(0.495)
Past Exp. -0.175
(0.323)
Small Office -0.032
(0.275)
Constant -1.272*** -4.265
(0.192) (6.378)
Chi2 6.359 26.046
prob>Chi2 .012 .011
AIC 150 162
N 148 148
+ p<0.10, * p<0.05, ** p<0.01, *** p<.001; two-tailed tests; robust std. errors in parentheses.
Picturing Uncertainty
Table 5: PQML Regression of Ties Activated on Treatment (Limited Control), Work Locus of
Control, and Control Variables
Model 9 Model 10 Model 11 Model 12
Limited Control -0.029 -0.019 -0.036 -0.034
(0.105) (0.102) (0.103) (0.097)
Work Locus of
Control
-0.016
(0.011)
-0.022*
(0.011)
Limited Control x
Work Locus of
Control
0.045**
(0.016)
0.051***
(0.015)
Age -0.139 -0.100
(0.096) (0.086)
Age - Sq. 0.001 0.001
(0.001) (0.001)
Female -0.014 -0.009
(0.102) (0.097)
White 0.359+ 0.345+
(0.190) (0.205)
Black / African
American
0.225 0.242
(0.239) (0.256)
Asian 0.153 0.159
(0.227) (0.235)
Born – Ex. US 0.198 0.306*
(0.151) (0.150)
Now Married -0.050 -0.019
(0.156) (0.150)
Never Married 0.330 0.335+
(0.201) (0.197)
Past Experience 0.135 0.179
(0.118) (0.114)
Small Office -0.230* -0.222*
(0.101) (0.100)
Constant 1.970*** 5.117* 1.971*** 4.035+
(0.079) (2.386) (0.078) (2.145)
Chi2 0.076 21.582 8.752 39.037
prob>Chi2 .782 .043 .033 .000
AIC 1023 994 1006 975
N 158 158 158 158
+ p<0.10, * p<0.05, ** p<0.01, *** p<.001; two-tailed tests; robust std. errors in parentheses.
Picturing Uncertainty
Table 6: Fractional Logit Regression of Proportion Bridging Ties on
Treatment (Limited Control), Work Locus of Control, and Control Variables
Model 13 Model 14 Model 15 Model 16
Limited Control 0.384 0.389 0.387 0.384
(0.261) (0.268) (0.261) (0.270)
Work Locus of
Control
-0.007
(0.029)
0.007
(0.032)
Limited Control x
Work Locus of
Control
0.022
(0.041)
0.005
(0.045)
Age 0.059 0.066
(0.265) (0.270)
Age - Sq. -0.001 -0.001
(0.003) (0.003)
Female 0.218 0.221
(0.269) (0.270)
White -0.216 -0.190
(0.597) (0.607)
Black / African
American
0.398 0.461
(0.695) (0.724)
Asian -0.149 -0.087
(0.673) (0.714)
Born - Ex US -0.393 -0.383
(0.410) (0.432)
Now Married -0.359 -0.350
(0.402) (0.407)
Never Married 1.069* 1.074*
(0.479) (0.481)
Past Experience -0.364 -0.347
(0.320) (0.322)
Small Office -0.032 -0.015
(0.271) (0.272)
Constant -1.099*** -0.666 -1.098*** -0.894
(0.193) (6.418) (0.192) (6.572)
Chi2 2.167 27.594 2.480 27.478
prob>Chi2 .1410452 .0063397 .4789681 .0166754
AIC 152 164 156 168
N 148 148 148 148
+ p<0.10, * p<0.05, ** p<0.01, *** p<.001; two-tailed tests; robust std. errors in parentheses.
Picturing Uncertainty
Table 7: Fractional Logit Regression of Prop. Same-Sex Ties on
Treatment (Limited Control), Female, Work Locus of Control, and Control Variables
Model 17 Model 18 Model 19 Model 20 Model 21
Limited
Control
-0.069
(0.159)
0.186
(0.207)
0.256
(0.212)
0.180
(0.202)
0.265
(0.209)
Female 0.093 0.121 0.087 0.140
(0.245) (0.228) (0.241) (0.224)
Limited
Control x
Female
-0.534+
(0.319)
-0.658*
(0.296)
-0.545+
(0.316)
-0.698*
(0.297)
Work Locus of
Control (LOC)
-0.013
(0.020)
-0.007
(0.021)
Limited
Control x Work
LOC
0.032
(0.034)
0.020
(0.034)
Female x Work
LOC
0.048
(0.033)
0.028
(0.039)
Limited
Control x
Female x Work
LOC
-0.090+
(0.048)
-0.074
(0.053)
Control
Variables
No No Yes No Yes
Constant 0.092 0.050 6.563 0.057 7.213+
(0.121) (0.152) (4.460) (0.146) (4.278)
Chi2 0.187 4.989 33.160 9.304 35.282
prob>Chi2 .665 .173 .002 .232 .006
AIC 164 166 181 174 189
N 157 157 157 157 157
+ p<0.10, * p<0.05, ** p<0.01, *** p<.001; two-tailed tests; robust std. errors in parentheses.