Possible Selves: An Exploration of the Utility of a Narrative Approach

This study has two main aims: (a) to explore young men's and women's hopes and dreams for the future and (b) in so doing to explore the utility of a story-writing method. It is argued here that when identity is conceptualized as a life story, researchers need to consider story-writing methods in their investigations of identity formation. Three methods were used to compare findings from the story-telling method with more traditional methods, in this case an interview and a questionnaire. This research revealed that the story-writing method was useful for a variety of reasons. It provided rich data that generated themes that were taken up by the following methods. Moreover, it was more successful than the other methods at bringing to light participants' ideal selves.

Various methods have been used to investigate future aspirations. The most popular choice of method used to study people's hopes and dreams has been either a questionnaire (e.g., Kalakoski & Nurmi, 1995;Nurmi, Poole, & Katakoski, 1994;Nurmi, Seginer, & Poole, 1990) or an events listing approach (e.g., Poole & Cooney, 1987). Some researchers have elected to use interviews: for example, the intensive interview technique developed by Levinson (1978Levinson ( , 1996. However, in the main, most researchers have opted for the quick and cost-efficient questionnaire.
The research presented in this article draws from a much larger study that examined young people's hopes and dreams for the future. The focus here is on the effectiveness of a story-writing approach in this investigation. The utility of the story-writing approach is assessed by comparing results obtained from an interview method and a questionnaire. More specifically, I focus on age and sex differences in young people's dreams and compare the findings yielded from each method. If, as is argued here, identity is a life story, then using narrative techniques to investigate people's hopes and dreams could be of great benefit to researchers.

IDENTITY AS A LIFE STORY
Over the past 20 years, there has been a growing interest in examining people's stories of their own lives (e.g., Bruner, 1986;Cohler, 1982;Gergen & Gergen, 1987;Goodson, 2000;Mair, 1990;McAdams, 1989McAdams, , 1993McAdams, , 1996Ochberg, 1994;Polkinghorne, 1988;Sarbin, 1986). Such theorists have proposed that human beings are storytellers by nature and that stories are a natural way to report experience. For example, Sarbin suggested that almost everything of interest in psychology is narrative. In fact, he proposed that narrative may be a root metaphor for psychology as a whole. Sarbin argued that to know a person one must know the story in which that person is participating. Making sense of others requires knowledge of their subjective experience. Mair (1988) also suggested that psychology be approached as a storytelling discipline. He claimed that we live in and through our stories. Mair (1988) stated that the only world we can know is a story world. Moreover, Howard (1991) argued that: We are in the process of creating value in our lives-of finding the meaning of our lives. A life becomes meaningful when one sees himself or herself as an actor within the context of a story-be it a cultural tale, a religious narrative, a family saga, the march of science, a political movement, and so forth. (p. 196) Many narrative psychologists are more concerned with what people say than with what they do. Bruner (1987) proposed that "narrative imitates life, life imitates narrative. 'Life' in this sense is the same kind of construction of the human imagination as 'a narrative' is" (p. 13). In addition, Ricoeur (1984) argued that there seems to be no other way of describing lived time except in the form of narrative. Even if we were to give a brief account of events that have happened to us, these events would be chosen according to who we now see ourselves to be and thus would take their place in the form of narrative. He argued that we do not copy reality (i.e., recall the empirical facts) but rather give a new reading of events. Other theorists such as Neisser and Fivush (1994) have taken a similar line of argument, stating that it is important to recognize that narratives are not fixed or static but rather change with every retelling.
The stance taken in this article is that it is important to recognize, in line with the theorists cited previously, that the stories people construct about their lives are influenced by how they see themselves at a particular point in time. In particular, Dan McAdams's (1993) understanding of identity is drawn from in this current research. McAdams focused on the stories that individuals construct. He used the term personal myth to describe the stories people create when forming their identity. In essence, a personal myth is "an act of imagination that is a pat-terned integration of our remembered past, perceived present, and anticipated future" (McAdams, 1993, p. 12). The personal myth does not change dramatically from day to day but rather undergoes transformations over time (McAdams, 1996). Therefore, the early memories we recall can be considered an indication of our present understanding of the self. We choose in our present lives to remember particular events in our past in a certain way. Likewise, we select possible future scenarios depending on how we see ourselves in our current lives. This means that stories of the self can change with each retelling depending on who one understands oneself to be in the present. Our memories of the past and our projections of the future are not fixed. Identity is a life story. In accordance with this view, I explore the utility of a story-writing method in investigating young people's hopes and dreams for the future.

MEASURING HOPES AND DREAMS
This research investigates one aspect of the personal myth-individuals' narratives of their future lives. Past research on hopes and dreams has focused primarily on individuals in their midteens. Given that identity is neither determined nor fixed by the end of adolescence, there is a need to examine individuals' dreams beyond midadolescence. Although adolescence is an appropriate developmental stage to examine in terms of life choices, there is also a need to focus research beyond adolescence, in light of the substantial changes that occur in the lives of young adults (Clausen, 1995). Hence, in this study I elected to examine young adults' hopes and dreams for the future and the impediments they foresee to these dreams.
No standardized narrative method exists for conducting narrative research. For example, some researchers prefer to use biography, whereas others prefer a story-writing technique. Although some narrative techniques, such as the sentence completion test (e.g., Agarwal & Tiwari, 1988) and Levinson's (1996) intensive interview technique, have been used by researchers when investigating people's hopes for the future, these techniques are not common. I propose here that although narrative techniques are more time consuming, they can present a richer picture and give greater insights into the construction of identity than other, more traditional methods.
In examining the usefulness of a story-writing methodology in the investigation of young people's hopes and dreams for the future, I compared three methods in this article: (a) a questionnaire method, (b) an interview method, and (c) a story-writing method. Usefulness was determined by how adequately each method was able to evaluate how young people would like to see their future lives.
In our quest to determine the most appropriate techniques to examine the life story, we also need to consider the different types of selves that various methods might be measuring. Nurius (1986, 1987) developed a theory about possible selves. They argued that the three possible selves include the expected self, the hoped-for self, and the feared self. The expected self is the person one believes one can realistically become. According to these theorists, the expected self acts as the focal point for one's energies in striving for the future. In contrast, the hoped-for self is farther away from the present self. The hoped-for self may or may not be realistic. Finally, the feared self is a possible self that one does not desire to become. The feared self serves as a motivator, so that the individual takes action to avoid that possible self. In this article, I investigate the hoped-for self.
In addition, I take account of Higgins and his colleagues' (Alexander & Higgins, 1993;Higgins, 1987;Strauman & Higgins, 1987) distinction between the different aspects of the self. They proposed that we need to consider three selves: an actual self, an ideal self, and an ought-to self. The actual self is the representation of how you or another actually believes you are; the ideal self is the representation of how you or another would like to see yourself, including hopes and wishes for you; and the ought-to self represents the attributes that you believe you should possess. It is not always clear whether researchers are tapping into actual selves, ideal selves, or ought-to selves, when they are investigating people's future lives. For example, a teenage girl who is asked to list her hopes and fears for the future might be writing the hopes she thinks that an adolescent female ought to have. I am interested in how effective the three different methods are at elucidating the ideal self in respect to young people's hopes for the future.

Aims of This Research
This study contributes to the research that has investigated people's hopes and dreams for the future by examining age and sex differences in young people's dreams.
Past studies of people's hopes, dreams, and concerns for the future have focused mainly on adolescence (e.g., Gillies, 1989;Greene, 1986;Greene & Wheatley, 1992;Nurmi, Poole, & Seginer, 1995;Seginer, 1992;Trommsdorff, Lamm, & Schmidt, 1979) because it is at this age the transition from childhood to adulthood occurs. The onset and completion of adolescence typically are defined by biological and social changes, which makes the actual age range for this period somewhat fuzzy; however, most research concentrates on the secondary school years, from ages 13 years to 18 years. In respect to hopes and dreams studies, few studies have moved beyond midadolescence to examine the transition period from late adolescence into adulthood. I attempt to redress this omission by examining the hopes and dreams of older adolescents (17 to 22 years) and young adults (28 to 33 years).
In line with past research, sex differences are also considered. For example, Roberts and Newton (1987) reviewed four unpublished dissertations based on Levinson's (1978) theory and found that, unlike those of men, women's dreams usually were split between occupational and interpersonal goals. Other studies of adolescents' dreams have found that boys tend to be more interested in material dreams, whereas girls are more interested in dreams relating to future family (Nurmi, 1991). Accordingly, I explicitly examine sex differences in young adults' dreams and obstacles to these dreams.
Toward this end, several methods were used so that findings from the story-telling method could be compared with those from an interview and questionnaire method. The story-writing method was the first method used; next, an interview was constructed in light of the themes generated from the story-writing method; and finally, a questionnaire was constructed out of the themes generated from the first two methods. Although the sequence of methods used could indeed influence the final assessment of the benefit of each method, it was the preferred means to generate the maximum amount of themes, without limiting the scope of the research too early in the study. Therefore, the main aim of this article is to explore the value of a story-writing method in the investigation of young people's hopes and dreams. Because of the breadth of data collected from the three methods reported here, it is impossible to report all of the results. Hitherto, the results that relate to answering the main research question of the utility of a story-writing method are included in this article.

METHODS
A multiple-method approach was used to research young people's hopes and dreams for the future. For the first study a story-writing method was devised, in the second study interviews were conducted, and in the third study a survey was administered.

Study 1: Story Writing
Participants first completed a short questionnaire asking for their demographic details. Then they were asked to write a story about how they would like to see themselves in 10 years. They were given an hour to complete the task, although most participants took about 40 min to write their stories. The instructions were as follows: Write a story about how you would like to see yourself in the next 10 years. Write the story in the third person. Include as much detail as possible and particularly avoid censoring details that seem irrelevant. Avoid psychological interpretation. Include any obstacles that get in the way of achieving your dreams/goals. Produce the most vivid story about how you will achieve your dreams/goals. These instructions were drawn from past studies that have asked adolescents to write essays about their future. For example, Gillespie and Allport (1955) and Mönks (1968) had their participants write essays anticipating their futures entitled "From now to 2000 A.D." Poole (1983) had respondents write down how their lives would be in 10 years. They were asked to write an essay on the following topic: Myself in ten years' time: e.g. what kind of job you'll have, whether or not you'll be married, the kind of person you'd like to marry, where you'll live, what kind of person you'll be, what kinds of things you'll be doing, etc. (p. 238) This seemed an appropriate focus for the narrative approach. However, I decided to exclude from the instructions suggestions of themes to focus on, so as not to close off themes that might emerge in the stories. Furthermore, I was interested in the obstacles that participants foresaw in relation to their dreams; thus, the directions requested participants to write the obstacles to their dreams. In addition, some concepts from memory work (Crawford, Kippax, Onyx, Gault, & Benton, 1992;Haug, 1987) were also considered in developing the instructions. A content analysis was carried out to clarify the number and types of dreams and impediments included in their stories.
First-year psychology students from two universities participated in this study. Other groups who participated included students' friends; social groups; and people attending two of Sydney's markets (one in the inner city and the other in the outskirts of Sydney).
In this sample, 3% had achieved a high school education, 9% had a school certificate, 5% had technical qualifications, 54% had completed the higher school certificate (Year 12), 23% had a degree or diploma, and 7% had achieved postgraduate qualifications. Although this sample is a little skewed toward the better educated, it is a reasonably good representation of the Australian population. In May 1997 (close to the time this sample was obtained), the Australian Bureau of Statistics (ABS, 1997a) reported that of 15 to 34 year-olds, 46.0% had not completed the higher school certificate, 32.3% had completed the higher school certificate, and 21.7% had postschool qualifications. It should be noted that 24.9% of the ABS population (considering it begins at 15 years) were still at school. When we consider that the age groups are somewhat different, with more people expected to be at school in the ABS population, the sample appears to be representative of the population. Table 1 indicates that this particular sample was more skewed toward professionals and intermediate clerical workers, salespeople, and personal service workers (ABS, 1997b). Those not included in the table were either students or unemployed.
Finally, 89% of the sample were born in Australia, another 2.5% came from English-speaking countries, and 9% came from non-English-speaking countries.
Instrument. In the second study, a semistructured interview schedule was used. Participants were questioned about the specific dreams that emerged in Study 1; however, the interview also allowed for flexibility to concentrate on dreams that participants considered important.
As in Study 1, the interview schedule asked participants to focus on how they would like to see themselves in 10 years. Demographic details were collected again. Questions were included about dreams that participants had already formed and tried to live out. Participants first were requested to describe the dreams they had for the next 10 years, in as much detail as possible. Then they were asked to explain how they might go about achieving these dreams and to consider any obstacles they foresaw to achieving these dreams. These interviews were audiotaped and transcribed for analysis.
Participants. The sample for the interview method was recruited from similar groups as used in Study 1, including university students, people attending markets, and social groups. It consisted of 24 men and 24 women from Sydney, Australia, of whom 24 were 17 to 22 year olds (M = 20.5, SD = 1.5) and 24 were 28 to 33 year olds (M = 30.6, SD = 1.8).
Of these, 4% had achieved a school certificate (Year 10), 52% had achieved their higher school certificate, 8% had achieved technical qualifications, 13% had achieved a diploma, 20% had achieved a degree, and 3% had achieved postgraduate qualifications.
The participants' occupations included 8.3% managers, 16.7% professionals, 4.2% tradespeople, 18.8% clerical workers, 22.9% salespeople or personal ser-POSSIBLE SELVES 217 Ninety-two percent of the sample were Australian born, 2.1% were born in English-speaking countries, and 6.3% were born in non-English-speaking countries.

Study 3: Questionnaire
Instrument. The main themes and subthemes that emerged from the previous studies were used in the questionnaire. 1 The Future-Orientation Questionnaire (Nurmi et al., 1990) also was drawn from to assist in the wording of the questionnaire. As in the Future-Orientation Questionnaire, this questionnaire asked how important particular dreams were and how often participants thought about these dreams. However, to be consistent with Studies 1 and 2, the questionnaire asked respondents to consider themselves in 10 years. Most of the questionnaire took the form of forced-choice questions, typically on a 5-point Likert-type scale. For example, respondents were asked, "In 10 years, how important will a career be to you?" The responses were based on the following: 1 (not at all important), 2 (not very important), 3 (somewhat important), 4 (rather important), and 5 (extremely important). An open-ended question invited respondents to add any further comments to the questionnaire once they had completed it.
Participants. In an attempt to obtain consistency in each of the samples, participants used in Study 3 were again recruited from similar groups including university students and social groups. The sample consisted of 100 men and 100 women from Sydney, Australia, of whom 100 were 17 to 22 years old (M = 19.5, SD = 1.7) and 100 were 28 to 33 years old (M = 29.2, SD = 1.7).
Of these, 1% had achieved high school level, 4% had achieved a school certificate, 51% had achieved their higher school certificate, 10% had achieved technical qualifications, 7% had achieved a diploma, 21% had achieved a degree, and 6% had achieved postgraduate qualifications.
Sixty-seven percent of the sample were Australian born, 20% of the sample were born in English-speaking countries, and 13% of the sample were born in non-English-speaking countries. Although the samples recruited for the three studies were slightly skewed toward a more educated and higher socioeconomic class, they were overall fairly similar when compared to each other (e.g., educational levels shown in Table 2). Furthermore, each method required participants to complete the task under similar conditions. Once people granted their consent, they were required to write their story, to be interviewed, or to fill out a survey in one of the labs at Macquarie University. I carried out all three studies. Hence, although having different participants participate in each of the studies could indeed affect the results, similarities in demographic variables and conditions most probably eliminated some of the potential bias.
Findings. The aim of this study was to examine the utility of a story-writing method in researching young people's ideal selves in respect to their hopes and dreams for the future. I did so by comparing some of the themes generated and the age and sex differences elicited by three methods. Because of the breadth of data collected from each method, it is impossible to provide a detailed account of every finding in this article. Instead, I present the results that help answer the research question concerning the usefulness of a story-writing method in analyzing young people's ideal future lives.

ANALYSES Study 1: Story Writing
In the story-writing method, a content analysis initially was carried out to clarify the number and types of dreams and impediments included in participants' stories. An initial reading indicated the presence of many of the themes revealed in past studies (see Nurmi, 1991). However, the overriding impression was the complexity of the interwoven identities. It became a challenge to code the stories without sacrificing too much of their richness. Given that this study also was used to generate themes for the following studies, I considered it necessary to document as many of the themes as possible, even if only a few of the participants mentioned them. The main dreams were concerned with occupation, finances, romance, parenting or having children, travel, leisure, friendship, parents and siblings, and religious and political beliefs. These main themes were next broken down into subthemes, which included further description of the dreams and impediments to these dreams. To check reliability of the coding procedure, 10% of the data were recoded by a trained scorer. Intercoder reliability was high, ranging from κ = 0.75 to 1.00.
Logistic regression was used to analyze the data. Initial analyses for the main categories of dreams (occupation, finance, family, marriage, friends, children, leisure, travel, religion, and politics) included the independent variables age and sex and their interactions. The same variables were used in the analysis of the subcate-POSSIBLE SELVES 219 gories. If the interaction was significant, tests of simple effects were carried out with a conservative p value of .025. If there were any empty cells, the Pearson chi-square was used to test for simple effects. Significance for the variables age and sex was taken to be at p ≤ .05.

Study 2: Interviews
The interview transcripts first were examined for the themes that emerged from the story-writing method. All of the main themes and many of the subthemes revealed in Study 1 were also uncovered in this study. The transcripts then were reread to locate any other pertinent themes. Although no new main themes emerged, some additional subthemes were yielded in this analysis. Logistic regression was again used in this study. Age and sex and the interaction of age and sex were once more included in the analyses. If any interactions were significant, tests of simple effects were carried out by using a conservative p value of .025. If there were any empty cells, the Pearson chi-square analysis was used to test for simple effects. Once more, to check reliability of the coding procedure, 10% of the data were recoded by a trained scorer. Intercoder reliability was again high, ranging from κ = 0.75 to 1.00.

Study 3: Questionnaire
The independent variables age and sex and their interactions were used to analyze the questionnaire. Given that most of the dependent variables were numerical, analyses of variance were used in the analysis. Where dependent variables were categorical, a logistic regression was used. The main effects employed a p value of .05, whereas the interactions used a p value of .01 Table 3 are the frequencies at which participants mentioned the main dreams in Studies 1and 2 and those who rated these as important in Study 3. The different ways these results were obtained need to be highlighted here. In the story-writing method, respondents were permitted to write freely about their possible future lives; in the interviews, all the main dreams revealed in analysis of Study 1 were used as prompts; whereas in Study 3, one of the questionnaires required respondents to rate their three most important dreams, given the same prompts used in the interview. Therefore, despite some of the interesting comparisons that can be made with these results, the comparisons nevertheless need to be treated with some caution. For example, it easily could be argued that had the participants in the story-writing method been given the prompts travel, leisure, friendship, politics, and religion, more of the participants might have included these themes in their stories. Across all three studies, occupation and romantic relationships were considered the most important and most frequently mentioned dreams. Nurmi (1991) pointed out, in his review of studies focusing on adolescents' dreams, that occupation is the most frequently mentioned dream, and that romantic relationships also rate fairly highly. It is interesting that when participants were interviewed, religion was rated as the third most important dream (a result atypical of past research). This might be because the interview tapped into expected or ought-to selves: Religion is a dream these participants felt they ought to be considering for their future lives. Furthermore, in the questionnaire, unlike Study 1, financial dreams were seen as not important, whereas dreams about parents and siblings were considered reasonably important. Again, this supports the idea that different methods uncover different selves.

Reported in
Overall, the three studies produced similar findings. In every instance where age and sex differences were found, they were in the same direction. Despite these similarities, there are a number of important differences. For the sake of brevity, only the differences that contribute to the main research question are reported here.

STUDY 1: GENERATION OF THEMES
The intention of using the story-writing method first was to generate themes. This indeed proved to be a wise choice. In addition to the main dreams cited previously, the data revealed numerous subthemes. For example, some of the subthemes that stemmed from occupational dreams included becoming successful in one's career, requiring qualifications for one's job, having a glamorous job, having a high-status and influential job, doing work that would benefit society, being promoted, a need to balance a relationship and children and a career, lack of abilities as an impediment to one's hoped-for job, and a desire to reach the pinnacle of one's career before having children. (There were more subthemes for occupation not listed here.) Past research on hopes and dreams typically has focused on a more broad level. Few studies have moved beyond the main dreams (such as those outlined in Table  3) to include subthemes. Therefore, the story-writing method allowed for both a macro-and microanalysis of people's dreams. In fact, the subthemes yielded important age and sex differences that would have been missed if the analysis simply had focused on the dreams per se.
The interviews were also useful in generating themes. Although fewer dreams emerged in the interview study, it did elicit a few more subthemes not mentioned in Study 1.

Grand Dreams Formed in One's Youth
One of the interesting findings revealed by the story-writing method was that some of the younger participants (17 to 22 years old) described many grandiose dreams for the future. For example, many of them seemed overly ambitious in their career dreams, and some aspired to extraordinary achievements (such as finding a cure for cancer). Others wrote that they would be famous (such as being a supermodel or in a rock band), whereas others hoped to marry someone famous. Following are some examples of extracts from these types of dreams: Currently she is in charge of promoting channel 9's new series and current affairs and is working very closely with the Packer family. More specifically, the statistical analysis revealed that older adolescents (17 to 22 years old) wrote more than young adults (28 to 33 years old) about having a glamorous occupation (such as a rock star, supermodel, or a journalist), χ 2 (1, N = 280) = 7.08, p < .01; having a benevolent occupation (such as a human rights activist), χ 2 (1, N = 280) = 6.55, p < .01; and becoming wealthy, χ 2 (1, N = 280) = 18.5, p < .001. Women wrote more than men about having a powerful and successful partner, χ 2 (1, N = 280) = 8.22, p < .01. Moreover, younger women wrote more than older women that they would have a high-status and influential job, χ 2 (1, N = 140) = 5.53, p < .25. Given that we do not know the abilities and potentialities for these participants, it is difficult to argue that these themes are unrealistic; however, the story-writing method does highlight the different tone evident in these participants' personal myths.
This unrealistic tone depicted in some of the younger group's stories is in line with Levinson's (1978) concept of the Dream. Levinson argued that initially the Dream is poorly developed and is not necessarily connected with reality. Consequently, it may start by taking a dramatic or unrealistic form. He also argued that if the Dream is unrelated to the person's life structure, it will fade away and lose purpose, and that the age 30 transition period (28 to 33 years) is a time when people begin to settle down and a de-illusioning process begins (Levinson, Darrow, Klein, Levinson, & McKee, 1976). If the Dream is not well formed, it might become abandoned or reworked at this stage. This conceptualization is supported by the younger groups' unrealistic ambitions, whereas the older group was more cognizant of the real opportunities available and the real impediments to their dreams. For example, the young adults were more concerned than the older adolescents about such things as a lack of money being an impediment to their career dreams, χ 2 (1, N = 280) = 4.55, p < .05; the need to provide for their children, χ 2 (1, N = 280) = 4.29, p < .05; the need to save money, χ 2 (1, N =280) = 3.86, p < .05; and the desire to make financial plans with their partner, χ 2 (1, N = 280) = 13.56, p < .001.
These results can be further explained by Elkind's (1974Elkind's ( , 1978Elkind & Bowen, 1979) notion of the personal fable. Elkind argued that adolescents go through a stage where they develop heroic stories about themselves. He considered that lack of experience in the world partly explained adolescents' egocentricity. This idea of the personal fable could explain why more of the younger group in this study desired a career in which they would be famous and acquire great wealth. Although Elkind argued that the personal fable is more evident during midadolescence, it is clear from Study 1 that some older adolescents and young adults still develop narratives that resemble a personal fable (Whitty, 1997(Whitty, , 1998(Whitty, , 2001.
As a corollary of these results, Studies 2 and 3 included a focus on whether younger participants develop less realistic dreams. From these results, a somewhat different picture emerges.
The overall tone evident in the interviews collected in Study 2 appeared to be more serious and somewhat less ambitious than in the story-writing method. Congruous with Study 1, the older adolescents stated more often than the young adults about dreaming of becoming wealthy, χ 2 (1, N = 48) = 4.18, p < .05. However, in contrast to Study 1, there were no significant differences for wanting a glamorous occupation or a benevolent occupation. Furthermore, more participants in Study 1 (32.5%) mentioned a glamorous occupation compared with Study 2 (22.9%), and more participants in Study 1 (22.1%) mentioned a benevolent occupation compared with Study 2 (14.6%). Moreover, although 13.9% of participants in the first study wrote about the desire for a powerful and successful partner and 30% wrote about having a high-status and influential job, these themes did not emerge in the second study.
An explanation for these differences could be that the interview method elicited different responses from Study 1 because the interviewed participants were faced with an interviewer, or a real audience. Elkind's (1978) theory of the personal fable was proposed as an explanation for why some of the older adolescents in Study 1 developed stories involving extraordinary subthemes. In turn, this theory could be used to explain the differences in responses between Studies 1 and 2. In conjunction with the personal fable theory, Elkind argued that adolescents have an exaggerated belief that everyone is watching them. This he termed the imaginary audience. He posited that many adolescents believe that because people are being critical of them, acting outrageously makes little difference. It therefore appears that the story-writing method had some of the participants, especially in the younger group, envisioning this imaginary audience. This might explain why more of the stories in the story-writing method contained grand and extraordinary themes. Furthermore, because participants were faced with a real audience in Study 2, they may have been likely to provide answers they saw as suitable to that audience. For example, in Study 2, more men (12.5%) than women (0%) stated that they felt advantaged in their careers because of their sex. Given that the participants in Study 2 were interviewed by a woman, this might have appeared the acceptable response. It is poignant that this theme was not mentioned by any of the 280 participants in Study 1.
Given that the questionnaire presented numerical data, in which respondents rated the importance of each theme on a 5-point Likert-type scale, it is a little more difficult to compare results borne out from the questionnaire with the categorical data generated from Studies 1 and 2. Nevertheless, it is noteworthy that age and sex differences elicited for grandiose themes in Study 1 were typically not significant in Study 3. The only significant difference for the grand themes outlined previously was for the desire to work in a benevolent occupation. Furthermore, rather than an age difference there was a sex difference in Study 3, whereby women scored higher on this subtheme than men, F(1, 199) = 4.88, p < .05.

Different Selves
The main aim of this article is to elucidate the value of a story-writing approach in the study of young people's hoped-for futures. It appears from the results just outlined, along with some further examples given here, that the story-writing method was more successful at elucidating participants' ideal or hoped-for selves, whereas Studies 1 and 3 tapped more into actual, ought-to, or expected selves.
As described previously, Study 2 yielded more serious narratives than Study 1 and found participants often justifying their responses to the interviewer. For example, the participants who mentioned that they were advantaged in the workplace because of their sex and that religion is an important dream (see Table 3) could be projecting their ought-to or expected aspects of the self. Furthermore, a unique subtheme to Study 2, stated by 10.4% of the participants, was that a financial dream was not a priority in their lives. Hence, Study 2 elicited a more socially accepted response, although one might hope to be wealthy or comfortable in the future, it should not be a major concern in one's life.
Once more, the more socially desirable responses evident in Study 3 show that this study was more inclined than Study 1 to measure respondents' actual, expected, or ought-to selves. It is noteworthy that although no significant differences in the first two studies were obtained for an occupational dream, Study 3 found that younger participants scored higher than the older for wanting an occupational dream, F(1, 199) = 6.78, p < .01. This dream could have been overemphasized by the younger participants in Study 3 because occupation is something younger participants are expected to consider as important. Therefore, in this instance, the questionnaire probably revealed participants' ought-to selves. To give another example, women scored higher than men in Study 3 when asked directly if they were concerned about not finding the right partner, F(1, 199) = 12.21, p < .001. Moreover, another result exclusive to Study 3 was that women scored higher than men for stating they would work hard at their romantic relationships, F(1, 199) = 3.79, p < .05. These results appear to be fairly realistic given the many requirements for a perfect partner and the emphasis they placed on their romantic dreams across all three studies. It could be argued that this reflects women's "actual" concerns.

CONCLUSIONS ABOUT THE UTILITY OF A STORY-WRITING METHOD
The story-writing approach was very successful at elucidating the complexity of people's dreams. This study found that the strengths of the story-writing method were its abilities to generate themes and to uncover ideal or hoped-for selves. Narrative psychologists propose that when we consider identity, we should take into account the stories people have about their lives, rather than how their lives simply actualize. Some have argued for the inclusion in research of both real and unreal aspects of the self (e.g., Bruner, 1986;McAdams, 1993). In fact, Sarbin (1998) claimed: The words real and reality are excluder words. They tell only what something is not, and then only if the context is known. … It is important to recognize that real is employed as a term to convince one's self or another that the credibility assigned to an imagining is warranted. (p. 24) It is likely that some of the Study 1 participants' stories will not in reality eventuate. Does this mean that these narratives speak less truth? If the young woman does not eventually finish up in charge of Channel 9's new series and "work closely with the Packer family," and if the young man does not become a successful soccer player for Manchester United, does this make their narratives any less authentic, truthful, or valid? Simply because narratives appeared to the researcher to be grounded in fantasy does not mean they were not realistic possibilities for these individuals. Perhaps the young man who foresaw being chosen by Manchester United was, in fact, an excellent soccer player who could possibly be chosen to play on one of England's finest soccer teams. However, even if this story does not actualize, it still contains a certain kind of truth. Although the story-writing method produced some narratives with unrealistic possible futures, this does not mean that these narratives are inauthentic. As Sarbin (1998) argued, if we are only focusing on what is real, we are excluding parts of the narrative. Hence, both realistic and unrealistic narratives were considered to fall within the range of the possible future self. It is also noteworthy that not all participants in Study 1 produced extraordinary stories, but rather the story-writing method provided a greater opportunity for those who would like to envisage grand futures for themselves to do so. It is also interesting that the older adolescents were more inclined to write these stories; perhaps this says something about the way we are socialized to dream.
Although one might argue that the exclusion of prompts was a weakness of this specifically designed story-writing method, it can be equally counterargued that this was a strength of the research. Had the participants been prompted with certain dreams, they might have felt less free to write about how they would like to see their future lives.
In conclusion, this article reveals the need to examine identity formation after midadolescence and demonstrates the need to include this story-writing method in future research.