Methodological Choice and Its Consequences for ... - CiteSeerX

3 downloads 99 Views 114KB Size Report
... to Becky Wai-Ling Packard, Department of Psychology and. Education, Mount Holyoke College, South Hadley, MA 01075. E-mail: [email protected] ...
IDENTITY: AN INTERNATIONAL JOURNAL OF THEORY AND RESEARCH, 6(3), 251–271 Copyright © 2006, Lawrence Erlbaum Associates, Inc.

Methodological Choice and Its Consequences for Possible Selves Research Becky Wai-Ling Packard Mount Holyoke College

Paul F. Conway National University of Ireland, Cork

The literature on possible selves has grown significantly since the development of the construct in the mid-1980s. Situating our discussion of possible selves in the long-standing and extensive literature on self and identity, our goal in this article is to examine methodological choice and its consequences within possible selves research for both the researcher and the participant. Toward this goal, we reviewed 141 empirical articles and highlight here 4 methodological clusters within them, including the original predominant cluster, involving structured survey and interview, and 3 more recent less frequently employed clusters: narrative, visual, and drama. In this article, we discuss various limitations and advantages of each cluster, focusing on researcher investment (e.g., time, resources), researcher role (e.g., information gathering, intervention), the nature of participant input, and what is learned by each approach. In our conclusion, we outline areas of future development and concern for the study of possible selves and the development of self-knowledge and identity more generally. We argue that for possible selves research to prosper conceptually, it is vital that researchers attend carefully to their methodological choices and the implications of those choices for what can be learned.

How people of all ages construct a positive sense of self is an important and burgeoning topic in research on human development. With it comes a renewed interest in how researchers construe and subsequently operationalize constructs related to the development of identity and self-knowledge (Ashmore & Jussim, 1997; Byrne, Correspondence should be addressed to Becky Wai-Ling Packard, Department of Psychology and Education, Mount Holyoke College, South Hadley, MA 01075. E-mail: [email protected]

252

PACKARD AND CONWAY

1996; Harter, 1999; Schwartz, 2001; Schwartz, Mullis, Waterman, & Dunham, 2000). This can be a difficult task because the “self” has been an important psychological concept since the early days of psychology—for example, James’s “empirical self” (1890) and Baldwin’s “socius,” combining the ego and alter (1897). Further, the importance of understanding identity exploration as being integral to an individual’s developing sense of self has been recognized for decades (Ashmore & Jussim, 1997; Erikson, 1968; Kroger, 2004; Nakkula & Ravitch, 1998; Tesser, Felson, & Suls, 2000). As a result, there is much literature with which to contend in this broad arena, including that which focuses on the development of selfknowledge and that which focuses on identity exploration or formation. Over the past two decades, research guided by future-oriented models of self-knowledge has proliferated, whether positioned within a broad future-time perspective (Husman & Lens, 1999), framed as working on personal projects (Little, 1983), articulated as possible selves (Markus & Nurius, 1986), or described as life goals (Nurmi, 1992). Indeed, to the extent that learning, motivation, and development are all directed toward some eventual end, one could argue that developing a sense of self is above all a future-oriented process. We argue that, as such, these literatures have relevance for researchers and practitioners concerned with identity formation. In particular, this article focuses on one generative and widely used futureoriented psychological construct for understanding the development of selfknowledge: possible selves (Markus & Nurius, 1984, 1986). In the mid-1980s, Markus and Nurius (1984, 1986), responding to what they viewed as the neglect of future-oriented aspects of self-definition in psychological research, developed the construct of possible selves. According to Markus and Nurius (1986), possible selves are images of what people hope to become, expect to become, or fear becoming in the future. Informed by one’s social context, including past and anticipated experiences, these personalized images motivate behavior by providing energy and guidance to pursue hoped-for selves and to avoid feared selves. We view identity as a dynamic process involving a search for self-definitions in the present. In other words, people seek some coherence in their present or immediate self-definitions by drawing on both their past, remembered selves (retrospective selves) as well as their future, imagined selves (prospective selves), which are structured around and through identity elements, including goals, values, and beliefs. This definition highlights the dynamic, temporal, and integrative aspects of identity, for which possible selves provide a generative framework. Several key identity researchers have emphasized a dynamic future-oriented process of identity development, where identities are explored and created as people gain information about themselves through life experiences (e.g., Berzonsky, 1989; Dunkel, 2000; Grotevant, 1987). As Dunkel (2000) argued, the development of possible selves can serve as an important mechanism in identity development, with the process of generating or “trying on” possible selves being viewed as an in-

POSSIBLE SELVES METHODOLOGY

253

tegral aspect of identity exploration. Our work builds on that of Dunkel (2000); we see possible selves as a useful framework for understanding how people develop self-knowledge and for contributing to identity formation research more broadly. Furthermore, we draw attention to the temporal and future-oriented nature of identity formation by situating our examination of possible selves in the wider literature on the future-oriented aspects of the self. The literature on possible selves has grown significantly since the mid-1980s, including hundreds of published articles in peer-reviewed journals, conference presentations, book chapters, and dissertations. A review of this literature is beyond our scope here. Instead, our goal in this article is to examine the methodological choices underlying possible selves research and, beyond this, to delve into the more practical implications of these choices. We look back over the past two decades to identify both the predominantly employed methods and the more recent methodological developments for studying possible selves. Of particular interest are the implications of these methodological choices for researchers studying possible selves and of identity formation more generally. As we begin, we are cognizant of Harter’s cogent admonition (1990) to researchers to not put “the methodological cart before the conceptual horse” (p. 292). Put another way, one must use caution in assuming similarity in terms and construct measurement when drawing from a complex, dense body of literature, whether the focus be self-concept (Byrne, 1996), personal identity (Schwartz, 2001), or, in this case, possible selves. The potential exists to assume that the possible selves constructs within the numerous studies published over the past two decades are synonymous when in fact researchers may be conceiving of possible selves in distinctly different terms. On this front, we acknowledge that literature on the development of selfknowledge has proliferated with some distinct differences in metatheories, or global perspectives, about the self, each with its own definition of terms. Two metatheories have been particularly prominent in research on the self: self as a collection of schemas (Markus & Wurf, 1987) and self as a story (Bruner, 1990). The self as a collection of schemas has its origins in research on self-perceptions undertaken by social and personality psychologists based on a cognitive view of the mind (e.g., Markus, 1977). The self as a story is rooted in the tradition of narrative psychology (e.g., Bruner, 1990; McAdams, 1996). Bruner in particular argued for narrative as a distinct way of knowing and viewed the “self as storyteller” (p. 111). Although both of these are perspectives on the self, they are very different. Each of these metatheories has implications for a researcher’s position on the emic–etic continuum, the extent to which the researcher plays, respectively, a insider or an outsider role in the research process as that relating to or understanding one’s participants (see Berry, 1969; Pike, 1967). Consequently, this position has implications for the role of the researcher and the participant. For example, central to the self-as-storyteller metatheory is the notion of the self as an unfinished narrative

254

PACKARD AND CONWAY

project that can be co-constructed with the researcher, whereas the self as a collection of schemas may be more aligned with a view of the researcher as a miner of information (Kvale, 1996). Thus, as a researcher strives to understand the development of self-knowledge in a particular study, the researcher may take on the role of a miner separate from the participants or as a collaborative storyteller in dialogue with the participants. In summary, our critical review of possible selves methodology can serve as a starting point to talk broadly about the implications of methodological choices for studying the development of self-knowledge and identity. We begin by reviewing the original, predominant cluster of methodologies before moving on to review other recent, less frequently used clusters. As part of our assessment of the consequences of these different methodologies, we discuss implications of these methods for researcher investment, researcher role and participant input, and, above all, what is learned about the development of self-knowledge and identity. We conclude our review with a discussion of future directions and considerations for researchers of possible selves and the development of self-knowledge and identity in general.

CHOICES AND DEVELOPMENTS WITHIN POSSIBLE SELVES METHODOLOGY To assess methodological choices and developments, we reviewed only published papers that incorporated a discussion of possible selves. From the approximately 200 articles that met these criteria, we analyzed 141 empirical articles published in the English language that used possible selves as a framework to guide their inquiries. We found, rather quickly, that the possible selves construct has gained wide appeal across many fields of research, extending beyond human development, identity development, or personality more broadly construed. Although the research guided by a possible selves framework has primarily been psychological, including research on depression (e.g., Penland, Masten, Zelhart, Fournet, & Callahan, 2000), delinquency (e.g., Oyserman, Gant, & Ager, 1995; Oyserman & Markus, 1990; Oyserman & Saltz, 1993), and psychotherapy (e.g., Buirs & Martin, 1997), numerous others have used the construct, including researchers studying physical health (e.g., Black, Stein, & Loveland-Cherry, 2001; Shadel & Mermelstein, 1996) and marketing (Bower & Sprott, 1995; Phillips, 1996). Our examination of the possible selves literature published over the past 20 years revealed a large number of methodological choices. For each article, we identified the type of method used or the types of methods used; we then assembled all of the methods identified into clusters. We chose to organize the methods into clusters given the extent of variability within the methods themselves. For example, interviewing as a method can be structured or semistructured, and so we chose to split interviewing into structured survey and interview methods or narra-

POSSIBLE SELVES METHODOLOGY

255

tive methods, depending on the type of interview used. Ultimately, we organized the methods underlying research into possible selves into four methodological clusters: structured survey and interview, narrative, visual, and drama. Although we do not claim this as the only organizational framework possible, it does account for all of the methods employed in the articles we reviewed, and we think it provides us with ample opportunity to view emerging trends over the past two decades and highlight the affordances and constraints of various methods, including the researcher’s role, the participants’ input, and what is learned about self-knowledge and identity (see Table 1). In looking at these broad clusters, we found that the original and most frequently used practices were structured surveys and interviews (64% of methods used in the articles that we reviewed)—specifically, survey measures developed by Markus and colleagues and structured interview methods developed by Oyserman and colleagues. The other methods have been used mainly since the mid-1990s, employed to a lesser extent than that of the structured survey and interview cluster of methods (see Table 2). With regard to the three less frequently used methodological clusters, we noted that narrative approaches were used three times more frequently than the visual and drama methods combined. In the next sections, we select particular studies that allow us to highlight the core measures and approaches within each methodological cluster and the consequences of each choice. Structured Survey and Interview Methods The structured survey was the original method and remains the primary method to assess possible selves. Specifically, the Possible Selves Questionnaire was the original survey constructed to assess possible selves (Markus & Nurius, 1986), and researchers continue to use the measure commonly (Dunkel, 2000; Knox, Funk, Elliott, & Bush, 1998; Stein & Hedger, 1997). In the Possible Selves Questionnaire, participants use a Likert-type scale to rate their consideration of various possible selves, which include personality characteristics and names of occupations. Another popular possible selves survey method involves having the participant generate a list of possible selves and then complete a Likert-type survey of how likely it is that their most salient possible selves will happen in the future (e.g., Aloise-Young, Hennigan, & Leong, 2001; Black et al., 2001; Cross & Markus, 1991; Dunkel & Anthis, 2001; Penland et al., 2000). Other methods in this cluster include structured interviews and card sorting. Specifically, the method of list generation followed by Likert survey can be completed orally as a structured interview, as per Oyserman and Markus (1990) and Hooker (1992). These researchers found oral, face-to-face administration of measures to be more effective in yielding complete data than self-completion surveys with adolescent and elderly participants. The card-sorting method entails having

256 TABLE 1 Methods in Possible Selves Research

Structured Survey and Interview Occurrence

Metatheory of self

Examples

Dominant method since Commencement of Possible Selves research in the 1980s. Self as schema

Survey (Markus & Nurius, 1986) Structured interview (Oyserman & Markus, 1990)

Narrative

Visual (Graphical and Image Based

Drama (Role-Play, Visualization, and Observational)

Second most frequently used

Seldom used to date

Seldom used to date

Self as story

Self as story Self as visual schema

Written narratives (King, 2001; Whitty, 2002) Oral narratives (Krieshok et al., 1999)

Photo interviewing Packard et al., 2004) Drawing analysis (Gibbons et al., 1993) Pie chart (Strauss & Goldberg, 1999)

Self as story Self as action schema Dramaturgical self Role-play (Buirs & Martin, 1997) Daily diary (Zirkel, 2002)

POSSIBLE SELVES METHODOLOGY

257

TABLE 2 Frequencies and Percentages of Instances Within Methodological Clusters Over Time (Papers = 141; Instances = 152) Structured Survey and Interview 1986–1989 1990–1995 1996–1999 2000–2004 Total (%)

Note.

2 19 37 40 98 (64%)

Narrative

Visual

Drama

Total

2 7 14 18 41 (27%)

0 2 2 2 6 (4%)

0 0 5 2 7 (5%)

4 (3%) 28 (18%) 58 (38%) 62 (41%) 152

Ten papers used methods from two clusters and one paper used methods from three clusters.

participants sort cards labeled with particular possible selves descriptors generated by the researcher into a hierarchical organization that reveals participants’ relative rankings of the possible selves. This method can be advantageous when the researcher wants to gain a sense of the relative importance of a variety of possible selves and force a hierarchical structure, which eliminates the possibility that participants will endorse all possible selves to the same degree (Kerpelman, Shoffner, & Ross-Griffin, 2002). In terms of researcher investment, one major advantage is that structured methods—and surveys in particular—are efficient to administer across large numbers of participants. As with other surveys, participants tend to be familiar with them. Furthermore, researcher control is an important advantage because the researcher can decide which particular items will be rated or ranked and can then compare respondents with relative ease. The structured interview approach is preferable to pencil-and-paper surveys because the researchers can explain procedures more clearly as the questions progress. Also, the face-to-face, one-on-one format is more likely to yield complete responses from certain participants, including adolescents and the elderly. However, this method requires trained interviewers who can spend the time to administer the questions in a one-on-one, face-to-face manner. Similar to when using structured surveys, however, the researcher has control over the uniformity of the questions, and again the responses are provided in a form that efficiently allows for comparison analyses, such as when comparing an elderly sample and a college sample (Cross & Markus, 1991). Card sorting requires a similar investment on the part of the researcher because of the face-to-face time required for explaining procedures. It is also possible for the researcher to provide the directions in a group format, allowing for greater efficiency. As mentioned, card sorting can provide information about the relative importance to the participant of various possible selves. In addition, card sorting generally requires the participant to interact with the questions to a higher degree than does a survey, due to the physical act of sorting cards.

258

PACKARD AND CONWAY

Although structured methods do allow for greater researcher control, they do not encourage input from the participants themselves apart from the answers to the predefined questions. Thus, a disadvantage is that the researcher may not learn about aspects of possible selves not originally conceived by the researcher. A good example raised by Hooker (1992) is when a set of items is developed with one particular group—in this case, a college student sample—and is later used with an elderly sample. A researcher may omit potentially salient items, such as health-related questions, from the larger pool given that the items were not originally conceived of by the researcher or instrument developer. Although structured research methods are efficient, researchers are generally discouraged from extended dialogue with the participants, dialogue that, if pursued, might reveal an alternative perspective or critique of the procedure and measures. Thus, this cluster of methodologies is anchored closer to the etic, or outside, end of an emic–etic continuum, with little collaboration or input from the participant and greater researcher control. Researchers need to be mindful that what they are learning regarding the development of self-knowledge and identity could be influenced by the method they are choosing. Dunkel and Anthis (2001) cautioned that surveys that involve item endorsement could be problematic to administer to college students. Because the college students are in a stage of identity exploration, they may already be likely to endorse a large number of items, and so a survey format that allows for a lack of deliberate agency, such as endorsing items already present on a researchergenerated list, may yield an artificially large number of items being endorsed. In fact, Dunkel and Anthis reported that they found a greater range of possible selves in college students when using an item endorsement survey (Dunkel, 2000) than when using a participant-generated list of possible selves followed by a Likerttype survey (Dunkel & Anthis, 2001). As a result, multiple methods may be useful to address this issue. Furthermore, what is learned about possible selves in this cluster is typically constrained by numerical scales provided to participants about a list of items that are generated by the researcher. With regard to what is learned about the development of self-knowledge, the survey and structured interviews generally assume a set of self-schemas that can be elicited or gathered from the participant. The Likert-type survey or structured-interview approach provides a sense of the range of possible selves considered (e.g., positive, negative, career related, health related), the degree to which these have been considered, the participant’s perception of confidence in achieving these possible selves, and their relative importance. This cluster of methods most closely aligns with conceiving of possible selves as a set of self-schemas that the researcher aims to gather and analyze. Typically, the extent of the researcher–participant relationship ends when the data are collected.

POSSIBLE SELVES METHODOLOGY

259

Narrative Methods: Greater Participant Input and Potential for Dialogue Narrative methods for studying possible selves represent the second most frequently employed cluster. In them, researchers ask participants open-ended questions about their possible selves and then analyze these descriptions. For example, Ruvolo and Markus (1992) asked participants to describe their visions of their future lives where everything has gone as well as, or where everything has gone as poorly as it could, providing written narrative descriptions of their hoped-for and feared possible selves. Similarly, Lips (2000) asked college students to describe possible selves and then content-analyzed these written narratives for gender differences. Gonzales, Burgess, and Mobilio (2001) asked students to write out their plans for the future, after which the authors analyzed the quality of these plans for comprehensiveness. In a slight variant of this method, Gibbons et al. (1993) asked young people to tell stories about their possible selves, using their written stories and illustrations to conduct cross-cultural comparisons. Semistructured or openended interviews can also be useful for gaining access to narratives about possible selves (Whitty, 2002) because these transcripts form narratives that can be analyzed in a fashion similar to that of written narratives. In addition to analyzing individual narratives about possible selves, participants have shared their possible selves in a focus-group context, where the product is a group narrative (e.g., Burack et al., 1997; Curry, Trew, Turner, & Hunter, 1994; Kao, 2000). The focus-group method can be effective for soliciting responses from numerous participants to gain a sense of general trends, such as the specialties within medicine that students were hoping for or planning to avoid (Burack et al., 1997), what types of jobs were seen as being possible for persons of various ethnicities (Kao, 2000), and the realm of possibilities for girls (Curry et al., 1994). Given that it is a group narrative, the results are less useful for detecting individual differences than are one-to-one interviews, but it can be more efficient than conducting individual interviews or analyzing numerous individual written narratives. In terms of researcher investment and participant input, narrative methods can allow for more input on the part of the participant, given the dialogue that may occur and the open-ended nature of the format. However, the results can be more cumbersome to analyze and may not have the same uniformity in response and analysis. Coding rubrics need to be developed by the researcher, and some disagreement is possible among co-researchers that would then need to be resolved, making the method less efficient. The act of research in this area can become a form of intervention. With regard to the researcher’s role, data elicitation can be viewed as a preliminary phase, providing context for subsequent dialogue about the participants’ possible selves. For example, Greenberg, Wortman, and Stone (1996) used narrative methods to begin

260

PACKARD AND CONWAY

discussions of resilience, whereas Krieshok, Hastings, Ebberwein, Wettersten, and Owen (1999) used participants’ narratives to facilitate vocational rehabilitation for veterans. King (2001) used story writing as a way of learning about participants’ possible selves and also studied the health benefits, given that the writing process can be cathartic (Pennebaker & Seagal, 1999). Thus, open-ended, narrative methods allow participants greater control and input into the process. Further, the researcher can go beyond collecting information about the development of selfknowledge to working with participants on expanding or refining their possible selves. Conversely, the participants can provide insights into the research process, allowing the researcher to revise the approach, analysis, and conclusions. This places narrative methods closer to the emic, or insider, end of the emic–etic continuum. As discussed in the previous section on structured surveys and interviews (Dunkel & Anthis, 2001), what is learned about the development of selfknowledge or identity formation may also be influenced by the method selected. Whitty (2002) combined multiple methods—including story writing, a semistructured interview, and a questionnaire—and found that the story-writing method generated a larger number of possible selves than the other methods, possibly because of its open-ended nature. As mentioned, it may be advantageous, where possible, to triangulate multiple methods. Finally, when a researcher assumes that the self is a narrative or story, she or he may be more inclined to choose a narrative method that allows for the participant’s story to unfold. Rather than elicit a set of self-schemas as data, the self-as-story metaphor positions the researcher and participant as potential co-constructors of a story that is continuing to unfold. Visual Methods: Focusing on Representation of Possible Selves Visual methods form a third cluster of possible selves methods, although only scarcely used thus far. Because possible selves are rooted in personalized imagery (Ruvolo & Markus, 1992), some possible selves researchers have begun to use research methods that allow participants to go beyond the written word to visual methods. Visual methods can include image-based methods and graphical methods. Image-based research uses images to inform the research questions, not merely for decorative, illustrative purposes (Prosser & Schwartz, 1998). Typically, participants construct their own images of possible selves, and the researcher analyzes these images, either alone or in concert with the input of participants. Another method within this cluster is a pie graph, to chart depictions of the relative importance of possible selves (Strauss & Goldberg, 1999). Visual methods have been used infrequently and typically as a complementary method to structured survey or interview or narrative methods.

POSSIBLE SELVES METHODOLOGY

261

For example, Packard, Ellison, and Sequenzia (2004) used photo interviews in addition to survey and semistructured interviews to examine possible selves of urban adolescent girls. After completing list-generation surveys that elicited their hoped-for and feared possible selves and after engaging in semistructured interviews that targeted explanations for these hopes and fears, the participants were asked to think about relevant imagery that would capture their hopes and fears. The participants took their own photos with disposable cameras, annotated the pictures with captions and titles, and then engaged in interviews where the photos were used as a stimulus to spark conversation, as that consistent with visual ethnography protocols (Collier & Collier, 1986). In this study, the photo interviews provided additional insights. The participants took photos of places in their neighborhoods where hopes and fears originated, illustrating the importance of the community context in the development of possible selves. In addition, the participants took photos that reminded them of significant past events or aspects of their current selves, which illustrated links between past selves in forming their future selves. In another study, Gibbons et al. (1993) asked participants from three cultural groups to create drawings of women at work and to annotate them with stories. The images of the women communicated subtle nuances that might go undetected if participants had only written about their images of women at work. Indeed, the images provided important information to be analyzed in combination with the written texts. Another approach involves charts or graphs rather than drawings or photos; participants use a pie graph to portray the relative salience among possible selves, such as the percentage of the self that is concerned with home versus career (Strauss & Goldberg, 1999). Graphs provide another visual format that can provide a quantitative, proportional representation of salience rather than rely on ratings or narrative descriptions. One major advantage of visual methods is that they are sensitive to nontextual, spatial forms of representation and participants are likely to find them a novel and creative form of expression. In addition, it allows researchers to provide participants with a choice when the visual approach may be preferable to written or interview methods for participants with limited literacy skills, and it can expand both the contribution and the diversity of participants. This cluster of methods may therefore more closely parallel that which is being studied, images of possible selves under consideration, providing a more direct means of communication on the part of the participant than that of the written word. At a more basic level, images may provide additional “texts,” or narratives, to be analyzed. In discussing story lines—another graphical or visual mode of representation where participants chart out their stories along a timeline—Gergen and Gergen (1988) summarized their affordances, which include ease of construction, vividness of representation, and the participants’ ability to pictorially represent their experiences and help the researcher to understand experiences in a manner difficult to accomplish via other means.

262

PACKARD AND CONWAY

However, the approach overall may be more cumbersome and time-consuming for the researcher. For example, the researcher requires a scheme to analyze the imagery developed by participants. One may not know at the study’s inception what types of images will be constructed, and a likely strategy is to work from the data to inform one’s analytical categories via grounded theory (Glaser & Strauss, 1967). Further, an important caveat in using visual methods is that they might miss important meanings while, of course, allowing others—particularly in the absence of textual or verbal annotation. Visual methods require careful structuring using trained researchers who can spend the time required to administer the visual task and analyze the results. It also requires a researcher who is comfortable with the one-on-one, face-to-face interview required to elicit oral or written information to accompany the visual images. Participant input, as previously discussed, is relatively high when participants are able to construct images and annotate their meaning. Thus, the researcher is likely to gain insights that will serve to answer gaps left by other methods or spark new questions for later study. The researcher can gain both an insider and an outsider perspective rather than only the latter (Prosser & Schwartz, 1998). In addition, the visual images can be obtained through interactive projects that engage the participants (Ewald & Lightfoot, 2001), suggesting the possibility for dialogue and intervention beyond the data collection phase. Because a disadvantage with reference to researcher interpretation is in the absence of either verbal or written annotation, the richness and nuanced nature of visual data may be difficult to interpret, suggesting a need for dialogue between participant and researcher. Thus, the use of visual or graphical methods is anchored closer to the emic (insider) end of an emic–etic continuum, with potential for considerable collaboration or input from the participant. With regard to what is learned about the development of self-knowledge, visual methods are not necessarily rooted in either the self-as-schema or the self-as-story metatheory. For example, the pie graphical approach provides a sense of the salience of possible selves and can be viewed as being similar to the card-sorting method but involving an alternative form of representation. Similarly, an image-based method could be embedded in a context where there will be ongoing dialogue or where it is simply a complementary form of data that is collected at one point in time and not discussed or built on. Thus, the manner in which the method is used, the data analyzed, and the findings represented will convey the root metatheory. Drama Methods: Observing and Choreographing Possible Selves in Action Drama methods represent a fourth cluster, also less frequently employed thus far. In seeking to categorize this emerging cluster, we debated whether to title it psy-

POSSIBLE SELVES METHODOLOGY

263

chodrama or to use the more general term drama. We opted for drama because it denotes a somewhat broader framework, one that reflects observing and choreographing possible selves in action. For example, rather than ask participants to describe their possible selves through words or images, researchers can ask participants to act out possible selves. Buirs and Martin (1997) used role play in therapy to ask participants to try on various possible selves, such as a possible self that did not involve addiction. In this way, the researcher learns more about the malleability of the participants’ possible selves and the effectiveness of various interventions in changing the nature of possible selves. Pham and Taylor (1999) suggested that as an alternative to actually acting out possible selves, mental simulation or visualization techniques could be effectively used to help participants try on desired possible selves or to enact the steps that might lead to the desired possible selves. In their study, they found that college students who visualized the steps required in preparing for a midterm examination were more successful than those who concentrated their visualizations on the successful outcomes of excelling on the examination. Other researchers have opted to observe participants in natural settings in order to identify what these activities or actions reveal about participants’ possible selves. For example, Kyratzis (1999) observed children during play, whereas Cotte (1997) observed gamblers in a casino. Rather than actually observe the participants, other researchers have asked participants about their actions in order to link the nature of their activities and their descriptions of their possible selves. For example, Stein and Hedger (1997) used activity logs in addition to the Possible Selves Questionnaire, whereas Zirkel (2002) used diary entries in addition to a life task survey to learn about possible selves. In these latter examples, the records of activity were used as an ancillary method for understanding possible selves. The major advantages and disadvantages to drama methods are similar to those of visual methods. The researcher is not constrained by written or spoken words to capture possible selves but can use actions and behaviors to provide insight into possible selves. Possible selves are important because they provide a link between cognition and motivation, guiding people into action (Markus & Nurius, 1986). Thus, behavior of participants provides an important expression of possible selves: behavior that can be analyzed. However, the researcher’s investment in time and energy can be great, especially in therapeutic or educational settings where role plays or visualizations are used with participants. Researchers function as therapists and educators, training participants in these techniques and possibly working together over a period of time. The analysis of outcomes can also be more cumbersome, given that the “texts” are behavioral and may not be conducive to simple ratings or frequency counts. With regard to participant input, drama methods allow participants to be active in the research process because dialogue is important in the role plays and it may provide a more natural setting in which to dialogue. In the case of activity logs or observations of naturalistic behavior, for example, participants are encouraged to

264

PACKARD AND CONWAY

behave as they would normally. As previously mentioned, the researcher’s role can shift dramatically, requiring some degree of flexibility. The researcher may double as a therapist or educator, or the researcher may observe from afar or analyze activity logs without interacting with the participant. When the goal of the research is to learn more about the potential malleability of possible selves, intervention becomes an objective, and as a result the researcher’s role changes. Rather than collect data and analyze the meanings of behaviors or actions, the researcher may prolong engagement with the participant’s attempts to change, whether reducing or expanding possible selves. Finally, much like that with the visual methods, the root metaphor of self (e.g., a set of schemas or a narrative story) is independent of this method. The self could be viewed as a collection of action schemas to be collected and analyzed, or it could be viewed as a story in action to be co-constructed and changed over time.

AREAS FOR FUTURE GROWTH WITHIN POSSIBLE SELVES All four methodological clusters that we identify in this review are likely to remain appealing to possible selves researchers in the future. Nevertheless, although no single method is likely to monopolize the field, various survey and interview methods, given their dominance in the wider social science world (Kvale, 1996), are likely to remain to the forefront. Having said this, however, we believe that it is critically important that researchers be aware of the affordances and constraints of their methodological choices when seeking to advance understanding of the development of self-knowledge. For instance, in seeking to develop an understanding of possible selves, both in general and for specific populations, how might researchers who adopt different approaches (e.g., structured interview versus narrative interview) converse and engage, if at all, with each other’s literature? We have identified a number of methodological considerations for researchers in the area of possible selves. First, we note with interest the small but important number of studies that incorporate multiple methods to assess possible selves. Good examples of places for researchers to build on for future research include Whitty’s thoughtful discussion (2002) of the complementary nature of multiple methods to assess possible selves, including limits of each type of method, and Dunkel and Anthis’s provocative consideration (2001) of the potential method by stage-of-life interactions. This will not only help to increase understandings of a more diverse range of participants but will also help to create methods that are normed with diverse participants. Second, we noticed the limited number of longitudinal studies (e.g., Dunkel & Anthis, 2001; Frazier, Hooker, Johnson, & Kraus, 2000) and studies that employed data collection with dyads instead of individual participants (Kerpelman et al.,

POSSIBLE SELVES METHODOLOGY

265

2002). There is a need to consider investing in these types of studies, especially to learn about the malleability of possible selves over time and about the complex constructions of possible selves, respectively. Third, we noticed that most research on possible selves has been undertaken with young adults who are college students, which is not surprising given the prominent issues of identity at this stage of life. However, of studies with college students, many are samples of convenience, recruited from introductory psychology pools, where students tend to be similar not only in their developmental stage but in their social and economic background. Few studies of possible selves focus on young children, although a growing number of studies include adults (e.g., Frazier et al., 2000) and the elderly (e.g., Chasteen, 2000). Focusing on participants who vary in their stages of life is clearly an area for future development.

DISCUSSION Future developments in research on possible selves will be, we think, most fruitful when researchers attend to the interdependence between conceptual and methodological issues not only within research on possible selves but in the wider field of research on the development of self-knowledge and identity formation more broadly (e.g., Schwartz, 2001). Thus, in this section, we discuss two emerging themes in research on possible selves and identity more generally: the interrelationships among methodology, epistemology, and ontology; and power and perspective. We think that any discussion of research methodology ought to recognize the interrelationships among three levels of questions: ontological, epistemological, and methodological (Mertens, 1998). At the ontological level, researchers implicitly or explicitly ask themselves about the nature of reality—specifically, what is it? At the epistemological level, researchers implicitly or explicitly take a stance on how they regard the nature of knowing and the relationship between the knower and the known. Finally, at the level of methodology researchers seek an answer to the question, how can the knower go about obtaining the desired knowledge and understandings? Clearly, in answering this last question, possible selves researchers have answered it in diverse ways over the last 20 years, as evidenced by the methodologies they have employed. We also want to draw attention to how possible selves researchers’ ontological and epistemological assumptions may reflect different understandings of the self, identity, and possible selves. For example, the mere use of a particular method does not in of itself convey a particular ontological and epistemological stance. As Kvale (1996) argued in relation to interviews, two different underlying epistemologies prevail, and he offered two contrasting metaphors with which to convey these different positions: the interviewer as miner and the interviewer as

266

PACKARD AND CONWAY

traveler. In the case of the interviewer as miner, “knowledge is understood as buried metal, and the interviewer is the miner who unearths the valuable metal. Some miners seek objective facts to be quantified; others seek nuggets of essential meaning” (p. 3). The contrasting, traveler metaphor, on the other hand, understands the “interviewer as a traveler on a journey that leads to a tale to be told upon returning home. … The interviewer wanders along with local inhabitants and asks questions that lead subjects to tell their own stories. The interviewer converses with them in the original Latin sense of the term: ‘wandering together with’” (p. 4). These contrasting metaphors of miner and traveler align with the notion of the self as a collection of schemas to be mined or as a story to be co-constructed via travels through another’s narrative world. We see this as an important consideration when conducting research on the development of self-knowledge, because the perspective of the researcher does interact with the nature of the self under study. As we have noted, the use of various methods in possible selves research has developed to include structured survey and interview, narrative, visual, and drama methods. Of critical importance in understanding the methods used is one’s paying close attention to researchers’ explicit or implicit epistemology of the self—that is, how the researcher assumes that one comes to know the self in the first place. The choice of methodology reveals how the researcher has operationalized possible selves and his or her stance toward the research. The methods can be considered on multiple dimensions, from researcher guided to participant guided, from pencil-and-paper to open-ended observations, from objective researcher to activist educator. The research methods adopted by possible selves researchers encompass both epistemologies of miner and traveler in Kvale’s explication (1996) of these contrasting metaphors. As such, researchers’ explicit or implicit assumptions about the nature of self-knowledge and identity formation and the manner in which one can come to understand the development of self-knowledge and identity formation shape methodological choices, which can in turn constrain or expand understandings of possible selves and phenomenological understandings of identity processes. Research on possible selves, particularly the rationale underpinning methodological choices, has been influenced quite considerably by the wider debates in social science regarding issues of power and perspective, specifically as these relate to the making of methodological choices. In the wider context, over the last two decades there has been greater attention paid by researchers to the manner in which the methods they employ are of themselves implicated in dominant or minority discourses in society. For example, feminist developmental psychologists have drawn attention to the now well-documented manner in which landmark studies of human social–emotional development sampled males only, resulting in gender-biased views of social–emotional development (Crick & Rose, 2000). Regarding possible selves, we observed that when research instruments are normed with a

POSSIBLE SELVES METHODOLOGY

267

particular type of population, the possibility exists to overlook categories that may be more salient to participants who are drawn from another type of population (Hooker, 1992). Beyond the sampling criticism, more recent critiques have drawn attention to the diverse experiences of people along lines of gender, race, and social class and how chosen methods may favor one group over another. This recognition of the validity of diverse experiences and its possible interaction with the chosen research methods has led some researchers to argue for methods that facilitate participants’ perspectives—that is, insider, or emic, views of social and emotional development. Moreover, the context of the study is important when conceptualizing the goals of the study and the power dynamics of the research. Studying possible selves of beginning teachers and their hopes and fears about learning to teach (Conway & Clark, 2003), when compared to the hopes and fears of minority adolescents (Oyserman et al., 1995; Packard et al., 2004) or the hopes and fears of older adults (Chasteen, 2000), provides very different research settings in terms of the actual data-collection context and the research study objectives. We think that both the research study context and the objectives are critically important in determining the methodological choices that researchers make in their study of possible selves. For example, as noted earlier, some uses of the narrative, visual, and drama methods have explored or addressed the malleability of possible selves to a greater degree than structured survey and interview methods—for example, role play as a potential intervention in terms of trying on an alternative or possible self—possibly leading the researcher to take a more deliberate role toward intervening or changing participants’ possible selves (see Oyserman, Terry, & Bybee, 2002). The implications of this stance, given power dynamics in particular research settings, is incredibly important for researchers to consider. In conclusion, the proliferation of methods on possible selves over the last two decades can be read as an extension of approaches to understanding and researching the development of self-knowledge and identity formation in the wider psychological literature. “Reflexive consciousness” (Baumeister, 2000, p. 10)—that is, a focus on self-knowledge, self-schemas, self-awareness, and self-perception— has been actively studied in psychological research over the last 50 years. To date, possible selves has had wide appeal within psychology and research on the future-oriented self in different domains. The future of possible selves as a topic of research looks promising given its appeal across a range of research disciplines and fields. This appeal, we think, reflects its richness in bringing forth the temporal and future-oriented dimensions of self-knowledge and meaning making (Kroger, 2004) as critical to identity exploration and formation throughout the life span. For possible selves research to prosper conceptually, as we have argued in this article, it is vital that researchers attend carefully to their methodological choices and the consequences thereof.

268

PACKARD AND CONWAY

ACKNOWLEDGMENTS We would like to thank Joan Hughes and Mark Gover for their insightful comments on a draft of this paper.

REFERENCES Aloise-Young, P. A., Hennigan, K. M., & Leong, C. W. (2001). Possible selves and negative health behaviors during early adolescence. Journal of Early Adolescence, 21, 158–181. Ashmore, R. D., & Jussim, L. (Eds.). (1997). Self and identity: Fundamental issues. New York: Oxford University Press. Baldwin, J. M. (1897). Social and ethical interpretations in mental development: A study in social psychology. New York: Macmillan. Baumeister, R. F. (2000). Ego depletion and the self’s executive function. In A. Tesser, R. B. Felson, & J. M. Suls (Eds.), Psychological perspectives on self and identity. Washington, DC: American Psychological Association. Berry, J. W. (1969). On cross-cultural comparability. International Journal of Psychology, 4, 119–128. Berzonsky, M. D. (1989). Identity style: Conceptualization and measurement. Journal of Adolescent Research, 4, 268–282. Black, M. E. A., Stein, K. F., & Loveland-Cherry, C. J. (2001). Older women and mammography screening behavior: Do possible selves contribute? Health Education & Behavior, 28, 200–216. Bower, A. B., & Sprott, D. E. (1995). The case of the dusty stair climber: A taxonomy and exploratory study of product nonuse. Advances in Consumer Research, 22, 582–586. Bruner, J. (1990). Acts of meaning. Cambridge, MA: Harvard University Press. Buirs, R. S., & Martin, J. (1997). The therapeutic construction of possible selves: Imagination and its constraints. Journal of Constructivist Psychology, 10, 153–166. Burack, J. H., Irby, D. M., Carline, J. D., Ambrozy, D. M., Ellsbury, K. E., & Stritter, F. T. (1997). A study of medical students’ specialty-choice pathways: Trying on possible selves. Academic Medicine, 72, 534–541. Byrne, B. (1996). Measuring self-concept across the life span: Issues and instrumentation. Washington, DC: American Psychological Association Chasteen, A. L. (2000). The role of age and age-related attitudes in perceptions of elderly individuals. Basic and Applied Social Psychology, 22, 147–156. Collier, J., & Collier, M. (1986). Visual anthropology: Photography as a research method. Albuquerque, NM: University of New Mexico Press. Conway, P. F., & Clark, C. M. (2003). The journey inward and outward: A re-examination of Fuller’s concern-based model of teacher development. Teaching and Teacher Education, 19, 465–482. Cotte, J. (1997). Chances, trances, and lots of slots: Gambling motives and consumption experiences. Journal of Leisure Research, 29, 380–406. Crick, N., & Rose, A. (2000). Toward a gender-balanced approach to the study of social-emotional development: A look at relational aggression. In P. H. Miller & E. Kofsky Scholnick (Eds), Towards a feminist developmental psychology (pp. 153–168). New York: Routledge. Cross, S. E., & Markus, H. R. (1991). Possible selves across the life span. Human Development, 34, 230–255. Curry, C., Trew, K., Turner, I., & Hunter, J. (1994). The effect of life domains on girls’ possible selves. Adolescence, 29, 132–149. Dunkel, C. S. (2000). Possible selves as a mechanism for identity exploration. Journal of Adolescence, 23, 519–529.

POSSIBLE SELVES METHODOLOGY

269

Dunkel, C. S., & Anthis, K. S. (2001). The role of possible selves in identity formation: A short-term longitudinal study. Journal of Adolescence, 24, 765–776. Erickson, E. H. (1968). Identity: Youth and crisis. New York: Norton. Ewald, W., & Lightfoot, A. (2001). I wanna take me a picture: Teaching photography and writing to children. Boston: Beacon Press. Frazier, L. D., Hooker, K., Johnson, P. M., & Kraus, C. R. (2000). Continuity and change in possible selves in later life: A 5-year longitudinal study. Basic & Applied Social Psychology, 22, 237–243. Gergen, K. J., & Gergen, M. M. (1988). Narrative and the self as relationship. In L. Berkowitz (Ed.), Advances in experimental social psychology (pp. 17–56). New York: Academic Press. Gibbons, J. L., Lynn, M., Stiles, D. A., Jerez de Berducido, E., Richter, R., Walker, K., et al. (1993). Guatemalan, Filipino, and United-States adolescents’ images of women as office workers and homemakers. Psychology of Women Quarterly, 17, 373–388. Glaser, B. G., & Strauss, A. L. (1967). The discovery of grounded theory: Strategies for qualitative research. New York: Aldine. Gonzales, M. H., Burgess, D. J., & Mobilio, L. J. (2001). The allure of bad plans: Implications of plan quality for progress toward possible selves and postplanning energization. Basic and Applied Social Psychology, 23, 87–108. Greenberg, M. A., Wortman, C. B., & Stone, A. A. (1996). Emotional expression and physical health: Revising traumatic memories or fostering self-regulation? Journal of Personality and Social Psychology, 71, 588–602. Grotevant, H. D. (1987). Toward a process model of identity formation. Journal of Adolescent Research, 2, 419–438. Harter, S. (1990). Self and identity development. In S. Feldman & G. Elliott (Eds.), At the threshold: The developing adolescent. Cambridge, MA: Harvard University Press. Harter, S. (1999). The construction of the self: A developmental perspective. New York: Guilford. Hooker, K. (1992). Possible selves and perceived health in older adults and college students. Journals of Gerontology, 43, 85–95. Husman, J., & Lens, W. (1999). The role of the future in student motivation. Educational Psychologist, 34, 113–125. James, W. (1890). The principles of psychology. New York: Henry Holt. Kao, G. (2000). Group images and possible selves among adolescents: Linking stereotypes to expectations by race and ethnicity. Sociological Forum, 15, 407–430. Kerpelman, J. L., Shoffner, M. F., & Ross-Griffin, S. (2002). African American mothers’and daughters’ beliefs about possible selves and their strategies for reaching the adolescents’ future academic and career goals. Journal of Youth and Adolescence, 31, 289–302. King, L. A. (2001). The health benefits of writing about life goals. Personality and Social Psychology Bulletin, 27, 798–807. Knox, M., Funk, J., Elliott, R., & Bush, E. G. (1998). Adolescents’ possible selves and their relationship to global self-esteem. Sex Roles, 39, 61–80. Krieshok, T. S., Hastings, S., Ebberwein, C., Wettersten, K. B., & Owen, A. (1999). Telling a good story: Using narratives in vocational rehabilitation with veterans. Career Development Quarterly, 47, 204–214. Kroger, J. (2004). Identity in adolescence: The balance between self and other (3rd ed.). London: Routledge. Kvale, S. (1996). InterViews: An introduction to qualitative research interviewing. Thousand Oaks, CA: Sage. Kyratzis, A. (1999). Narrative identity: Preschoolers’ self-construction through narrative in same-sex friendship group dramatic play. Narrative Inquiry, 9, 427–455. Lips, H. M. (2000). College students’ visions of power and possibility as moderated by gender. Psychology of Women Quarterly, 24, 39–43.

270

PACKARD AND CONWAY

Little, B. R. (1983) Personal projects: A rationale and method for investigation. Environment and Behavior, 15, 273–277. Markus, H. (1977). Self-schemata and information about the self. Journal of Personality and Social Psychology, 35, 63–78. Markus, H., & Nurius, P. (1984). Possible selves. Bulletin of the British Psychological Society, 37, A116. Markus, H., & Nurius, P. (1986). Possible selves. American Psychologist, 41, 954–969. Markus, H., & Wurf, E. (1987). The dynamic self-concept. Annual Review of Psychology, 38, 299–337. McAdams, D. P. (1996). Personality, modernity, and the storied self: A contemporary framework for studying persons. Psychological Inquiry, 7, 295–321. Mertens, D. M. (1998). Research methods in education and psychology: Integrating diversity with quantitative and qualitative approaches. Thousand Oaks, CA: Sage. Nakkula, M. J., & Ravitch, S. M. (1998). Matters of interpretation: Reciprocal transformation in therapeutic and developmental relationships with youth. San Francisco: Jossey-Bass. Nurmi, J. E. (1992). Age-differences in adult life goals, concerns, and their temporal extension: A life course approach to future-oriented motivation. International Journal of Behavioral Development, 15, 487–508. Oyserman, D., Gant, L., & Ager, J. (1995). A socially contextualized model of African-American identity: Possible selves and school persistence. Journal of Personality and Social Psychology, 69, 1216–1232. Oyserman, D., & Markus, H. R. (1990). Possible selves and delinquency. Journal of Personality and Social Psychology, 59, 112–125. Oyserman, D. S., & Saltz, E. (1993). Competence, delinquency, and attempts to attain possible selves. Journal of Personality and Social Psychology, 65, 360–374. Oyserman, D., Terry, K., & Bybee, D. (2002). A possible selves intervention to enhance school involvement. Journal of Adolescence, 25(3), 313–326. Packard, B. W., Ellison, K. L., & Sequenzia, M. R. (2004). Show and tell: Photo-interviews with urban adolescent girls. International Journal of Education & the Arts, 5(3). Retrieved January 26, 2005, from http://ijea.asu.edu/v5n3 Penland, E. A. M., Masten, W. G., Zelhart, P., Fournet, G. P., & Callahan, T. A. (2000). Possible selves, depression and coping skills in university students. Personality & Individual Differences, 29, 963–969. Pennebaker, J. W., & Seagal, J. W. (1999). Forming a story: The health benefits of narrative. Journal of Clinical Psychology, 55, 1243–1254. Pham, L. B., & Taylor, S. E. (1999). From thought to action: Effects of process- versus outcome-based mental simulations on performance. Personality and Social Psychology Bulletin, 25, 250–260. Phillips, D. M. (1996). Anticipating the future: The role of consumption visions in consumer behavior. Advances in Consumer Research, 23, 70–75. Pike, K. L. (1967). Language in relation to a unified theory of the structure of human behavior (2nd ed.). The Hague, Netherlands: Mouton. Prosser, J., & Schwartz, D. (1998). Photographs within the sociological research process. In J. Prosser (Ed.), Image-based research: A sourcebook for qualitative researchers (pp. 115–130). London: Routledge Falmer. Ruvolo, A. P., & Markus, H. R. (1992). Possible selves and performance: The power of self-relevant imagery. Social Cognition, 10, 95–124. Schwartz, S. J. (2001). The evolution of Eriksonian and neo-Eriksonian identity theory and research: A review and integration. Identity: An International Journal of Theory and Research, 1, 7–58. Schwartz, S. J., Mullis, R. L., Waterman, A. S., & Dunham, R. M. (2000). Ego identity status, identity style, and personal expressiveness: An empirical investigation of three convergent constructs. Journal of Adolescent Research, 15, 504–521.

POSSIBLE SELVES METHODOLOGY

271

Shadel, W. G., & Mermelstein, R. (1996). Individual differences in self-concept among smokers attempting to quit: Validation and predictive utility of measures of the smoker self-concept and abstainer self-concept. Annals of Behavioral Medicine, 18, 151–156. Stein, K. F., & Hedger, K. M. (1997). Body weight and shape self-cognitions, emotional distress, and disordered eating in middle adolescent girls. Archives of Psychiatric Nursing, 11, 264–275. Strauss, R., & Goldberg, W. A. (1999). Self and possible selves during the transition to fatherhood. Journal of Family Psychology, 13, 244–259. Tesser, A., Felson, R. B., & Suls, J. M. (Eds.). (2000). Psychological perspectives on self and identity. Washington, DC: American Psychological Association. Whitty, M. (2002). Possible selves: An exploration of the utility of a narrative approach. Identity: An International Journal of Theory and Research, 2, 211–228. Zirkel, S. (2002). Is there a place for me? Role models and academic identity among White students and students of color. Teachers College Record, 104, 357–376.