RESOURCES

Methods Column

Within-Person Methods
Theo A. Klimstra, Tilburg University, The Netherlands, 2017

A vast majority of researchers examining developmental processes in young adults (as well as in other age groups) mainly report on the “average” person and use a relatively limited number of measurement occasions to do so. Often, this suffices to answer the research questions we are interested in. However, for certain types of research questions we need a different kind of data. I’ll provide some examples of how I work with this different kind of data. It should be noted that my expertise is as an applied researcher, not as a statistician.

It is important to note the limitations of typical regression or correlation analysis. Indeed, we cannot conclude from these types of analysis that two variables are associated with each other within a particular person. A positive Pearson (or Spearman) correlation between identity exploration and fearfulness does NOT indicate that when your adolescent child starts exploring, they’d also become more fearful. Because the correlation and regression analyses we typically employ produce between-person associations, they only only tell us that adolescents who explore more than their peers in the sample (i.e., are higher up in the rank order of exploration) tend to be more fearful when compared to other people (i.e., are higher up in the rank order on fearfulness). Often, this will be the kind of information we are looking for, as it is an important task of psychologists to identify individuals who are at risk for developing some kind of problem.

However, to examine whether exploration induces fear in a particular adolescent child, we need a different kind of analysis and a different kind of data. That is, we need frequent measurements of both identity exploration and mood, and examine within-person associations. It has long been quite difficult to collect such data with frequent measurements of the same constructs within the same individuals, but the internet and especially the widespread use of smartphones has made it a lot easier to collect such data. Once the data is in, it can be analyzed with relatively complex multilevel analyses (see e.g., Klimstra et al., 2016, Journal of Research on Adolescence). In fact, in many cases, these will be the preferred method. In my own research, I have also used the somewhat easier-to-apply q-correlation. This q-correlation can be very appealing for examining some research questions.

The q-correlation is a within-person association and is therefore calculated for every single individual within a particular sample. So, after you run the syntax (there are syntaxes available for SPSS, but likely also for R and other software), a new variable appears in your dataset. This variable indicates, for every single individual in the sample, how the two constructs that you are interested in (e.g., identity exploration and some mood variable like fear) are associated. Values might range from –1.00 and 1.00. So, for some individuals this value might be positive (e.g., .70), meaning that more exploration comes together with more fear in these individuals. In others, values might be negative (e.g., -.30), which would indicate that they tend to become less fearful whenever they are exploring. In individuals with values close to zero, exploration and fear would be (almost) unrelated. Because the q-correlations appear as variables in the dataset, they can be used for any kind of follow-up analysis. For example, they can be correlated with other variables to answer research questions like “are fear and exploration more strongly associated in individuals with higher levels of neuroticism”? One could even identify profiles based on q-correlations, to distinguish individuals in whom these variables are positively related, from those in whom these same variables are unrelated or negatively related. Some year ago, I for example used this method to study individual differences in the extent to which mood is related to the weather (Klimstra et al., 2011, Emotion).

The possibilities of data with frequent measurements of the same construct are not limited to calculating within-person associations. Another possibility is to track the stability of certain constructs and examine individual differences in this stability. An easy way to do this is to calculate within-person standard deviations (for an application, see e.g., Klimstra et al., 2010, Journal of Personality and Social Psychology). However, these within-person standard deviations do often not produce the best estimates of within-person variability (e.g., Maciejewski et al., 2015, Child Development). Other measures, such as summing up absolute differences for each individual between all consecutive measurements for a particular construct might be better representation of variability. In addition to just studying variability, it is also possible to examine particular patterns in, for example, mood. For example, Kuppens and colleagues have studied mood inertia (resistance of certain mood states to change; e.g., Kuppens, Allen, & Sheeber, 2010, Psychological Science). Similar phenomena can obviously be studied in other variables that may be of interest to researchers studying young adults.

Many interesting phenomena can be examined if frequent measurements are collected among the same individuals and then analyzed with within-person analyses. The current review was meant to provide an illustration of some of these possibilities, however it is by no means an exhaustive list. When diving in to this kind of research, it is very important to read your way into the literature and consult methodologists if you do not fully understand certain analyses. Also, it is important to reflect on whether it makes conceptual sense to examine the construct you’re interested in on a daily level. For example, it doesn’t make much conceptual sense to examine short-term fluctuations in a supposedly stable construct like intelligence or some personality trait. Finally, I would not recommend examining within-person associations in datasets with just a few (e.g., 3 to 5) measurement occasions. Like with any type of method, only use within-person analyses if it matches your research questions AND you’ve got the right kind of data.


 

The Self-Perception Profile for Emerging Adults
Susan Harter, University of Denver, 2016, sharter@du.edu

Susan Harter is a Professor of Psychology at the University of Denver. She obtained her Ph.D. from Yale University where she joined their faculty before moving to Denver. She has been listed in two separate international surveys as one of the 50 top developmentalists in the field. She is well known for her life-span battery of Self-Perception Profiles as well as two developmental books (1999, 2012) on the Construction of the Self (Guilford Press).11:45 AM 7/24/2016


Introduction and Rationale

Why do we need another new multidimensional instrument, the Self-Perception Profile for Emerging Adults, described here, to tap self-perceptions during the period between adolescence and adulthood? Research reveals that most 18 to 25-year-olds do not yet consider themselves to be adults (Arnett, 1997). They realize that they are no longer adolescents but have yet to embrace adult roles and responsibilities in US culture. Thus, Arnett (2000, 2007, 2010) introduced the term “emerging adulthood,” bringing credibility to a new and distinct transitional period of development. This stage occurs between the age of 18, when most individuals graduate from high school, and around the age of 25-plus, when young people are presumed to enter “true” adulthood. During emerging adulthood, youth are typically no longer as dependent upon parents as they were during childhood and adolescence. However, most have not yet taken on the full and more enduring responsibilities of adulthood, for example, an occupational or career commitment, financial independence, marriage, home ownership or parenting.

Individuals during the period of emerging adulthood undergo many transitions (see summary in Harter, 2012). A major developmental task for emerging adults involves occupational exploration, often leading to education beyond high school, for example, college, vocational school, graduate school, or further professional training and degrees. There is often experimentation in the realm of romantic relationships. During this period, romantic liaisons typically last longer than during adolescence and many emerging adults engage in an exploration of a more serious intimate relationship and commitment that, for some, leads to the consideration of cohabitation or marriage.

Peer social relationships also undergo change. There are no longer the ready-made high school niches that comfortably provided friendship opportunities. The emerging adult must put forth more individual effort to make friends and garner acceptance in the workplace, on the college campus, or in other wider societal contexts.
Relationships with parents also undergo challenging transitions. During the high school years, the norm is still dependence on parents who meet many of the adolescent’s needs. For emerging adults, there is a developmentally-appropriate need to establish increasing independence from parents, and to become a more autonomous individual. However, the healthy emerging adult must also remain more maturely connected to parents (see Harter, 1999; 2012).

Global concepts of self may also change or be in flux during emerging adulthood. For example, a person’s level of self-esteem may undergo alterations. Another self construct, the perception that one is displaying one’s true self versus donning a false self, can also be subject to alteration (see Harter, 2012, for a more detailed description). Other new domains also arise during emerging adulthood or take on new developmentally-appropriate manifestations.

Domains of the Instrument

Thirteen different domains, each defined by their own subscale, were identified. Given that the very definition of emerging adulthood implies a sense of “becoming,” self-perceptions in each domain reflect the fact that these judgments are evolving. The emerging adult is a psychological “work in progress.” Thus, the actual wording of these thirteen domains, as well as of the actual items themselves, should reflect this normative developmental process.

• Intelligence. This subscale assesses how intelligent or cognitively competent the young adult perceives the self to be at this transitional period of life.
• Job/occupational competence. This subscale taps the extent to which young adults feel that they are successful at exploring job or occupational options where they can perform competently.
• Athletic/physical competence. Given that high school provides a structure for playing sports that will vanish upon graduation, the items on this subscale tap the young adult’s success in finding new opportunities within which to pursue their athletic interests and perceived physical skills post-graduation.
• Physical appearance. This subscale assesses the young adult’s perceptions of his or her current physical appearance, meeting personal or societal standards of attractiveness.
• Peer friendships/social acceptance. This subscale taps whether or not, after high school, the emerging adult has been able to develop new and meaningful friendships and to garner acceptance in his or her current social environment.
• Intimate relationships. This subscale assesses whether the emerging adult feels that he/she has the capacity to develop an intimate relationship, one that is more serious and reflects commitment, in contrast to casual romantic encounters.
• Relationships with parents. This subscale taps the challenge of moving toward increasing autonomy or independence from parents while continuing to maintain a more mature sense of psychological connectedness to parents.
• Morality. This subscale addresses the goal of developing more internalized moral or ethical standards that the young adult has personally come to construct and to own.
• Sense of humor. This subscale is specific to young adults’ ability to laugh at themselves in current life situations that may be unplanned, awkward, or potentially embarrassing.
• Daily life management. This subscale taps a major challenge for emerging adults, namely the practical ability to meet the new demands of managing one’s daily life responsibilities.
• Optimism. This subscale addresses emerging adults’ feelings of optimism or hopefulness about their futures versus pessimism and hopelessness.
• True-self/false-self behavior. This subscale taps the young adult’s ability to be his or her true self in social situations versus the tendency to don a false self in order to cope socially.
• Global self-esteem. This subscale taps the young adult’s current perception of his or her global self-esteem, that is, how much he or she likes himself or herself as a person, in contrast to disappointment with his or her overall self.

The actual instrument, formatted for use, with an explanation of the question format, administration instructions, scoring procedures, etc., are all described on my website: http://portfolio.du.edu/SusanHarter

Data Collection

We are currently planning our own data-gathering strategies. However, given advanced knowledge about this instrument, there has been considerable interest on the part of investigators studying this period of emerging adulthood, including requests to be allowed to administer this measure. I have been ambivalent, given that I think it first needs to have documented psychometric properties. I have discussed this with Jeffrey Arnett who has seen the measure and with his blessing, and a few tweaks of items, we have agreed on the following: For me to release the instrument to thoughtful researchers who will examine its psychometric properties including its validity in studies on this population, and report back to both of us.

I am particularly concerned that it be employed in the service of thoughtful research questions and hypotheses, to which I am told is what the members of this society are devoted. I feel it is essential that researchers include appropriate non-college samples. Too many studies have used university students as “samples of convenience”. However, the majority of emerging adults are not college students. So whom does one select? There are many considerations. One is a person’s socio-economic level. Another is the particular occupation. For example, are we talking about those who are still employed as wait staff or in service jobs vs. those who have become more skilled laborers such as electricians, plumbers, or those employed in the technological world of computers, who do not have college degrees? Are we talking about those who have moved along the path to independence from their families versus those who, for various reasons, are still living in their parents’ home, not uncommon in today’s world?

As Arnett has thoughtfully observed, an emerging adult is not a generic term for anyone in our culture who has graduated from high school and is of a certain age. Nor should researchers race off to other cultures, though cross-cultural issues are critical in our global, expanding or perhaps constricting, world today. I would not advise this. This instrument, like many others that I have constructed, was designed for American youth. I have the utmost respect for many colleagues who are doing thoughtful cross-cultural research, who know and understand the culture that they are investigating. But too many well-meaning researchers take American measures off to other countries or cultures, with little cultural sensitivity or knowledge of the culture, returning with meaningless results. People have taken my measures to cultures where there is not even a term for “self-esteem” in their language. I have written about cross-cultural issues in a chapter in my 2012 book, The Construction of the Self: Developmental and sociocultural foundations.

Jeffrey and I propose that, if people want to use this instrument toward examining thoughtful hypotheses, and if the instrument proves to be reliable and valid, such investigators submit their research, in an article (APA style) to both of us. We will review such articles and select a few to be the basis for a symposium at the October 2017 convention. If there is interest, we could schedule a follow-up session for discussion of the use of this instrument, for those interested.

This is an exciting new venture. I am a believer in emerging adulthood. I raised one of these creatures! We joked about how we would just “freeze-dry” her for a few years, and have her emerge as a fully-developed adult! Such is obviously not to understand development. Of course, I imagine myself to have been fully-evolved, skipping this tortuous period of development. Witnesses to the contrary, please do not come forth.
Please join us in this endeavor.