This article expands on a hospital diary I began after my second coronary bypass surgery in 1987. My wife, son, and daughter were close at hand throughout my second surgery, and this level of support contrasts sharply with my social isolation for my first surgery in 1977. My two surgeries can be construed as a repeated-measures quasi-experiment in which the same trauma is experienced with high and low social support. Noteworthy features of the diary include a surprising amount of concrete instrumental support (aid and assistance), which comprised the proximal stimuli of emotional support, a rich and nourishing intrapsychic life during the second, supported surgery, a prominent role for my computer and a teddy bear (“Bear”), and help with the implementation of decisions.

My interactions with Bear and the Microsoft Word artificial intelligence broadened my conceptualization of the “social” in social support. Cognitive representations of interaction sequences with my supporters from the past were supportive gifts that that kept on giving when the individuals represented in my relationship schemas were physically absent.

June 1987: This sentence is written as I lie in a hospital bed. Today is the fifth postoperative day of my second coronary bypass surgery (CABG). My plan is to use a laptop computer to: (a) distract myself from the pain of the surgery and the depressing, sterile boredom of early-morning television; (b) tell a few stories to anyone who is willing to read them; and (c) rethink, in the light of these recent experiences, my understanding of the mechanisms by which social support facilitates recovery from trauma. I am grateful that Microsoft Word, my Zenith laptop computer, and this manuscript have jetted me from bedridden cripple to bedridden bionic man in three keystrokes. I am able to feel competent as a writer and observer at a time when the incision pain in my chest prevents me from extending my arms more than a foot or so to reach for the telephone or my glass of water.

I am an experimental social psychologist with a clinical psychology postdoctoral. No one knew about my arm pain when my first coronary symptoms appeared in 1977, and I don’t think anyone outside the hospital knew I had been admitted to the coronary intensive care unit (ICU). The second surgery 10 years later stands in sharp contrast with the first procedure, as I am recently married and my adult children are close by. There are, of course, many differences between the 1977 and 1987 surgeries (e.g., different physicians, my age and prior experience with the surgery, etc.). But it is the sharp contrast between the support systems that were in place for the first and second surgeries that was and is, for me, the most salient difference between the two medical victimizations. (My use of the victimization metaphor throughout my diary reflects, among other things, how invaded and angry I felt. It also reflects my hope that the lessons learned will generalize to traumas other than heart surgery).

In the language of my mentor, Don Campbell, my two surgeries provide a “within-subject, single-subject, repeated-measures, quasi-experimental research design.” Comparing the two surgeries contrasts high and low social support for the same patient suffering the same medical trauma. Furthermore, I am gathering data through a distinctive methodological window. I am a psychologically minded, social psychology-sophisticated, participant-observer and self-observer writing in the middle of a crisis. Part of what motivates me to tackle this project is a conviction that I was able to see aspects of social support that would be missed with other methodologies

THE FIRST SURGERY (1977) WITH LOW SOCIAL SUPPORT

A month after my fortieth birthday, recently divorced, without a significant other, separated from my preteen children by 3,000 miles, and alone in the house, I felt a severe pain in my left arm while I was vigorously scrubbing shower tiles. My instant appraisal was that the pain was angina; I immediately crawled into bed, soapy scouring powder and all. I called my internist in the morning. He did a cardiogram and blood tests but the results gave no clue of a heart attack. When I finally asked for pain medication so I could get some rest on my third visit to my internist’s office in three days, he decided to admit me to the coronary ICU. I listened as my internist phoned the hospital to arrange my admission. I had driven to his office but he didn’t think I should drive to the hospital. Because my social isolation at the time was such that I couldn’t think of anyone that I could ask to drive me, he called an ambulance and sent me to sit alone in his waiting room. I heard the clatter of the paramedics as they rolled their stretcher in and out of the elevator again and again, first on one floor then another. They were having trouble locating their passenger. I finally decided to take the elevator down to their current location on what sounded like the second floor and asked, “Are you looking for Collins?” That lonely ambulance ride, flat on my back, from my internist’s office to the hospital, was one of the most traumatic events of my first coronary crisis. I still feel a bit of fear when I recall the ambulance attendant (a stranger), shock paddles in hand, seemingly disappointed that he did not have to “hot pop” me (his expression). That stop-and-go ambulance ride through the side streets of Los Angeles, electric shock paddles in the ready position inches away, is among the most terrifying cognitive representations in my mind.

Events got worse when we arrived at the hospital. The nurses stood back, intravenous (IV) needles in hand, because the ambulance attendant wouldn’t leave until I paid him. My stress and arm pain were such that I couldn’t write. Finally, we negotiated a plan in which he would write the check, and I would approximate my signature holding the pen, mitten-like, between a numb thumb and rigid fingers. Compared with that ride, the heart attack that evening was easy. All that I remember of the heart attack is being partially awakened by the tinkling of IV bottles in the middle of the night.

A motion picture imaging of my heart (an angiogram) and some slight chest pain a few days later provoked an emergency coronary artery bypass (CABG, pronounced cabbage) surgery in that same hospitalization. My chest was split lengthwise with a saw, my ribs split apart with a jack, my heart and lungs were stopped, a vein was harvested from my right leg, and two lengths of leg vein were attached to the arteries of my heart so as to bypass occlusions in the original 40-year-old plumbing. I had signed a consent form only an hour after I first learned the procedure existed. I remember only one thing from outside the hospital 10 years ago: I called a friend and asked her to buy me a radio. I wore the earphones (often with the radio turned off) and thus forestalled any attempted social interaction in the hospital. I withdrew into my isolation and ignored the strangers and roommates who passed through my room, including one who died.

THE SECOND SURGERY (1987) WITH HIGH SOCIAL SUPPORT

Almost 10 years exactly from the onset of my first chest pains, my chest pains began gradually to increase in frequency over four days on a trip to the East coast to participate in my daughter Kathy’s graduation from Yale. I resolved to call my cardiologist on my return to Los Angeles. Shortly after settling into my seat for the flight from New York to Los Angeles, however, there was a change in the pattern of the pain. The pain was totally unresponsive to nitroglycerin and was to wax and wane for six hours. In my mid-air appraisal, these symptoms bespoke a serious coronary event.

GETTING TO THE HOSPITAL (WITH HELP)

From the first twinge of chest pain, things went very differently the second time around. I got help from my family implementing my decision about what to do on board the airplane. The pattern of my chest pain, including the fact that the pain was unaltered by nitroglycerin, led me to conclude that I needed medical attention. But it would be four hours before the plane arrived in Los Angeles. I thought about notifying a cabin attendant but I anticipated that the resulting commotion would be stressful and I wanted to make the decision to land or wait for Los Angeles myself. I chose to wait for the familiar doctors, hospital, and home for my family in Los Angeles. (Thus, from the beginning, I was constructing a social support system for the second surgery.)

This was all decided through individual problem solving. My family members were inches and feet away but they did not know I was experiencing chest pain. My social support system was there in the plane with me but I had not yet interacted with it. In fact, they seemed to be a part of the problem, rather than a part of the solution. Would they panic and need support from me? Would one of them get the crew involved and get the plane landed at some foreign Midwest airport? A consensus plan was reached quickly. My wife, Mary, was sensitive enough to my nonverbal cues to ask me what was wrong and I informed her of my analysis and my decision. We had discussed this eventuality previously. We had defined our roles and so we brought to this mid-air crisis with a previously rehearsed cardiac emergency scenario. It was my task to interpret my chest sensations but I was obligated to keep my wife fully informed. Mary and I planned the logistics that would get me to the emergency room together. We would not say anything to the crew unless the pain got worse. I would ask my daughter Kathy and my son Kevin to decide who would ride with me to the hospital and who would pick up the luggage. Free of these tasks, my wife would call the emergency room (ER) and the cardiology practice from the airport to inform them of the symptoms and let them know I was on the way. I could then skip the triage process in the ER.

During this interaction with my social support network, my basic decisions remained unchanged but my ideas evolved from something like a few words scribbled in the margin of an article to something like a manuscript under review for publication. The reality of my interdependence with my family forced me to refine and verbalize a behavioral plan that might have remained half-baked and indeterminate if their physical presence had not prevented me from ignoring them. I am now convinced that the closure and consensus extracted in my interaction with my social support network allowed me to meditate and lower my heart rate. I now remember those four hours of suspense and suspended animation as a pleasant interlude. But without my family at my side, the mid-air environment of my coronary symptoms could easily have created an aversive and dysfunctional stress reaction. I think it is my family that cut off those particular cognitive demonsf.

We had a good plan and we executed it well. The division of labor among me, my wife Mary, my daughter Kathy, and my son Kevin expedited my trip from somewhere in the Midwestern sky to a coronary intensive care unit in West Los Angeles. My son and I took a taxi to the hospital from the airport. My wife called both the ER and the cardiologist from the airport. While my wife was on the phone, my daughter collected our baggage. When I arrived to wait at the ER waiting room, the triage nurse in the ER admissions needed only my name to lay me down and start an electrocardiogram (EKG). The cardiologist on call from the cardiology group practice (a stranger, as it turned out) was already in the hospital and had spent some time reviewing my chart before I arrived.

My social supporters shortened my wait for medical attention by at least an hour, compared with what it would have taken to get the job done working alone. How different that 1987 time on the airplane was from that isolated night in 1977 when I lay alone, wet, and covered with scouring powder, pondering the implications of a novel pain in my left arm. Scar tissue from the first surgery renders a second (“redo”) surgery technically more difficult and objectively more invasive than the first. In contrast to my vivid characterization of the first surgery, which I wrote while recovering from the second, there are no references to the second surgical procedure in the diary I brought home from the hospital. Cognitive representations of the second surgery are noteworthy by their absence from my mind. My account begins with the recovery room.

BEAR
I found a teddy bear (my first, a gift from my surgeons) in my surgical intensive care bed as I meandered into reality after the second, 1987 surgery. My image of the bear, nestled against my pillow, is the first (and cherished) memory I have from those surreal, Bergman-like intervals in the slow-motion, undulating transition into consciousness from the semi death of the general anesthesia and the heart-lung machine. There is an episode of the television show Eight is Enough in which young Nicholas brings his teddy bear to his troubled, oldest brother and says, “He is torn but he still works.” The line struck me at the time, but I couldn’t figure out what, exactly and concretely, it is that a teddy bear does when it “works.” How could I create functional and nonfunctional teddy bears in the lab? Now I know. Bear slept in my arms those first four nights as I strove to regain my balance from the surgical blow. He is the only entity who saw my tears in the hospital. Other than my wife and the nurse on duty the night before surgery, he is the only source of data revealing how angered I was by my coronary victimization.

With Bear as a surrogate therapist, I experienced and worked through negative affect to the point where I started writing about it, finding my feelings more interesting than devastating. Bear provided an audience not threatened by my anger and sorrow, and what Bear heard, I overheard. Whatever the advantages of human interactions, there are real costs to having other people in your room or talking on the phone. The mere presence of others is indeed stressful. Physical people wake you up, make you miss the second half of the television show you were watching, make you remember and relive medical details you would just as soon suppress at that moment, use your bathroom when you need it, and argue about what to watch on television.

There is something mystical about social support. How, exactly, do the mystical benefits of social support manifest themselves in experience? I have set for myself the task to watch and record the conscious manifestations of my healing spirit. (With a more contemporary take on Norman Cousin’s legacy to health psychology, I might now recast the “healing spirit” metaphor I created in the hospital to “What is the phenomenology of a healthy immune system?” Whatever the metaphor for the search, what I found was a rich and nourishing emotional life.

REPLAYING MEMORIES

I reconstructed episodes of previous interactions with my family throughout the pain-filled, drug- and trauma-jumbled nights and the sterile, isolated, early mornings. Those intrapsychic dramas I wrote and produced were among the most nourishing gifts I received in the hospital. I reconstructed tactile and visual sensations I had previously experienced holding my wife in my arms and I replayed excerpts from stock footage of our loving and intimate conversations. I replayed, on my biological videocassette recorder, my son’s kindergarten response to my tears as my father was dying: “That is okay dad, I am here.” I re-experienced my pride occasioned by my daughter’s honors at her graduation from college the week before. I visualized the three-foot, heart-shaped pillow now on the wall of my study that my daughter had made years ago and stitched with the words, “I love you.” I remembered my sister’s instant decision to fly to Los Angeles from South Carolina for my surgery when I called her after the angiogram, and I re-experienced the closeness we shared canning plum jelly while my mother was dying. I replayed these experiences on and on, tape after tape. Although these cognitive scenarios were closely modeled on historical reality, my affective responses were in real time. The cognitive rehashing of past events generated strong feelings in the present.

My choice to reconstruct these particular dramas determined, in large part, my emotional experience during times when I was physically alone. Some of these historical scenarios reconstructed in 1987 predate the first surgery 10 years earlier. There are many wonderful memories from the first 40 years of my life before the first, unsupported surgery and these were potentially available. For example, at the time of my first surgery, I had a cardboard box full of heart artwork my children had mailed to me, all embellished with scrawled “Loves,” although none of these came to mind while I was in the hospital in 1977. The difference between the first and second surgery, then, is not in the availability of a library of memories that can trigger nourishing affect, although this would be the case for a victim who had never experienced a supportive social system. For the second surgery, I was on good terms with my adult children. This time my support network was physically present, had been to visit me the day before, and would visit me again shortly after breakfast. The objective existence of my social support system rendered the healing fantasy episodes more accessible, more plausible, more real, and more evocative of nurturing affect.

PLANS

The stage plays in my head were not limited to historically accurate scripts stored in long-term memory. In these periods of physical separation from my family, I set up my priorities for spending the income tax refund. I made plans for the home repair tasks my son and I would tackle over the summer. I resolved to give my car stereo to my daughter and to buy a full-featured model for myself. I made plans to schedule my hospital visits so that I would have time with each member of my family individually. I thought about things my wife and I were willing to part with that would be helpful to my daughter as she set up her first home outside the dormitory. I made a list of the movies I wanted to rent for all of us to watch when I got home. I was concocting cognitive representations of interaction sequences with my supporters that were realistic in the sense that they were consistent with my relationship schemas of my present-day relationships with my supporters. These were not Walter Mitty escapes from reality or primary- process, wish fulfillment run amuck. These were plausible scenarios representing a realistic anticipation of interactions that would, in fact, be available in the future with members of my social support network.

There is a poster on my office wall that says, “A friend is a gift you give yourself.” It is there, in part, because I don’t understand it. Why, exactly, is a friend a gift, and in what sense do you give it to yourself? How could I give friendship in one experimental condition and withhold it in another? It was natural for me as an experimental social psychologist to look for things I could do to manipulate social support in the lab. Sometime on the airplane, I distracted myself from the limbo of chest pains over Cleveland as I formulated a plan to record (exactly and concretely) what my wife, son, and daughter did to help me.

My then-new colleague Chris Dunkel-Schetter (e.g., 1984) had convinced me, before the surgery, that: (a) ‘things would go better" with social support; and (b) emotional support might be more important than the concrete aid and assistance of instrumental support. These research-based propositions are important here because they constitute almost all I knew about social support as scientific phenomena when the chest pain began on the airplane in 1987. I was already surprised, before I got off the airplane, by the number of concrete acts of aid and assistance I noticed. The second of the two working hypotheses was disconfirmed. If I had encoded my victimization without a mindset to record instances of social support and if I had waited until a month after the event to pick a topic to write about, I do not think I would be sitting here trying to wax eloquent about the value of slippers and TV Guide. I suspect, rather, I would be writing the same stories I tell when I am trying to entertain— the midair setting of my symptoms, the evening I spent pondering the similarity between my .03 chance of dying that morning under the knife and the .05 significance cut-off, etc.

FURTHER REFLECTIONS ON THE MEANING OF MY DIARY

Although the above material has been edited, the ideas were in the 1987 hospital draft. Time, feedback, and scholarship have generated further associations. First, the cognitive events I recorded as the proximal cues for my healing spirit were interaction sequences imbedded in my relationship schemas (Pierce, Baldwin, & Lydon, 1997), including scripts, accounts (Harvey, Weber, & Orbuch, 1990) narratives (Baumeister, Stillwell, & Wotman, 1990; Berscheid, 1994), event schemas (Mankowski & Wyer, 1997), and stories. A still picture, a noun, or a string of adjectives (the kinds of stimuli that can be presented with quantified exposure times in the lab) would represent only the tips of these cognitive icebergs. The metaphor I chose in the hospital was a VCR, not a picture album. Second, it was these through-time scenarios that energized the production of nurturing affect. Third, my supporters engineered emotional support in real time with concrete acts of appropriate instrumental support. It was these pebbles of concrete, trivial, non-memorable acts of help that triggered the more noteworthy ripples and waves of warm fuzzies.

For example, one of my more vivid memories from the 1987 surgery is a brief moment in which I choked on a pill while talking on the phone. My son, accurately knowing that the chest pain made it difficult to reach out for a glass of water, even without a phone in my hand, rushed across the room and handed me a glass of water. “He knows!” I thought, “I am not alone in this pain.” The memory still brings a warm glow. But would he or any behavior observation system have coded that as emotional support? I think not. Is providing emotional comfort what Kevin thought he was doing? I doubt it. Interviewing Kevin or coding videotape of my hospital interactions would lead to the conclusion that I had been supported with instrumental help. That glass of water was not an example of what is usually meant by “supportive psychotherapy.”

My daughter’s gag gift provides another example. From her trip to the Venice Boardwalk, she brought back an extraordinarily appropriate gift, a dart gun with a ball of foam on the tip and a string tethering the dart to the end to the gun. The string meant I would not have to reach to retrieve the dart after I shot the gun. The gun provided an opportunity to be mischievous and playful, experiences definitely underrepresented during surgical recovery. How many other objects (temporary tattoos, sunglasses, T-shirts, etc.) had she processed before she found just the right one? Clearly, I realized, I had been on her mind while she was out in the sun and clearly she carried an accurate representation of my carved-up self and hospital circumstances.

Such acts of genuinely needed assistance are rarely observed when scientists bring social interaction into the lab and structure that interaction around a verbal discussion. It may be that the instructional set to the supporter that maximizes the receiver’s experience of emotional support would be something like: “Roll up your sleeves, get in there, and help your partner with this simulated grocery shopping task.” Such a set might elicit here-and-now behaviors more highly correlated with perceived support than, say: “Get in there and give some emotional support on this contentious issue” or “Get in there and do some reflective listening to the other person’s feelings.” Fourth, help with the implementation of plans and decisions also provided support.

Much of the concrete aid and assistance I recorded during the transition from symptom appraisal on the airplane to the ER was found in the elaborate dance my wife, son, daughter, and I did working through a course of action. This advantage of group versus individual problem solving is reminiscent of a conclusion I had reached 25 years before (Collins & Guetzkow, 1964): Consensus-based, group decisions made while working together with others are often not superior to individual decision making by single experts working alone; however, group discussion often facilitates the implementation of the decision, especially when the group members are the people who are responsible for the implementation. “Many important tasks that would defy the individual can be mastered by groups or larger social organizations” (Stroebe & Stroebe, 1996, p. 604). It is not clear how such collaborative medical decision making and action identification would fit into the oft-cited House (1981) typology of social support: (a) emotional support (love/concern, understanding, reassurance, encouragement); (b) instrumental support (aid or assistance); (c) informational support (advice, information of a problem-solving nature); and (d) appraisal support (approval). Fifth, we need to extend the range of methodologies used to study supportive interactions.

Although the case study methodology is unlikely to make its way into the scientific literature, other aspects of the present methodology would prove useful. Asking hospital patients to record their experiences into a tape recorder in story format would, I think, uncover data unlikely to be sampled in a structured interview. (The phrase “tell a few stories” is among the first I typed in the hospital.) Ethical cautions often make us skittish about interviewing trauma victims. But, after my own experience, I can strongly endorse composing a story each morning after surviving a night in the hospital as a coping strategy.

The nights of a hospital stay are lonely and stressful and disorientation may peek during that first hour after the nurse wakes you at 5 or 6 a.m. How nifty it would have been to have someone there to help me structure the dark chaos. If he or she brought along a tape recorder, a newspaper, and a glass of juice, all the better. Designs in which the same individual is followed through multiple traumas can also make a contribution to the social support literature. At present, we are struggling to separate the stable-dispositional from the environmental components of perceived social support. Comparing two different traumas experienced by the same individuals is close to the optimal methodology for that purpose.

Tragically, the AIDS epidemic often generates a sequence of multiple traumas over periods of time in which individuals are relatively isolated. Even with heart surgery, the half-life of a CABG is only 10 years and life circumstances often change substantially over that period. Sixth, virtual supporters are important. It has taken some time for me to think through the implications of the central role Bear played in my hospital diary. He was like the empty chair in Gestalt therapy, a surrogate for significant others. In Don Campbell’s language, he functioned as a social “entity.” A contemporary cyber metaphor would characterize Bear as a virtual supporter. On reflection, I realize my diary also documents extensive fantasy interactions with my mental representations of the many others in my perceived social support network.

These mental entities also functioned as virtual supporters, even if they were generated by my biological computer and not by computer chips. I framed the manuscript as another tribute (Brewer & Collins, 1981) to Don Campbell and a public accolade for my family. In that sense, writing the diary was a mental transaction between my physically absent supporters and me. Furthermore, I imagined a virtual audience of professional colleagues for this story from the first keystroke. Mentally anticipating the reaction of future readers made salient all the nourishing interactions stored in my relationship schemas with these others.

The artificial intelligences created by the software in my laptop also functioned as virtual social entities. I experienced the word processor as listening. Those of you who didn’t have “points taken off for spelling” throughout your pre-computer life may not understand, but spell checkers have been a godsend in my life, in general and in the hospital bed. I am certain that I would not have experienced the same sense of self-efficacy (“jetted me from bedridden cripple to bedridden bionic man in three keystrokes”) had I been using pen and paper.

Today, when I leave for the hospital (an altogether too familiar and routine life task), I make sure I have a pair of boxer shorts (hospital gowns are open at the back), some large-print escape reading (the world blurs in the hospital), and my next dose of medications (the hospital staff can take some time to get organized). The hospital will have a toothbrush. A laptop is too cumbersome a companion as one winds through the bureaucratic procedures of triage and the high-drama intake evaluation in the ER, but the computer is the first thing I ask my wife to bring to me when I reach my hospital room.

Seventh, I am no longer surprised that what social supporters do in real time is uncorrelated with perceived social support. The most important gifts my supporters gave me were a history of nourishing social interactions, real-time examples of the same to prime those nourishing schemas, and a schema of the current relationship that kept the nourishing interaction sequences accessible and realistic. To locate “help” in the helping behaviors of the moment is to ignore the context of these behaviors that is provided by both the interaction history of the relationship and the plausible future interactions implicit in the relationship schema of the here and now relationship.

Loving exchanges in the past were gifts from my supporters that kept on giving in a pain- and violation-induced, psychedelic, mental never-never-land; kept on giving in the middle of the night when the closest human could take 20 minutes to appear; and kept on giving in the face of affective displays that would have been inhibited by their physical presence. Furthermore, my schemas of my contemporary relationships were such that I could reasonably expect continued nourishing interaction sequences in the post-trauma future, an image of reality that my supporters found so many creative and compelling ways to confirm with within-hospital exemplars.

REFERENCES
Baumeister, R. I., Stillwell, A., & Wotman, S. R. (1990). Victim and perpetrator accounts of interpersonal conflict: Autobiographical narratives about anger. Journal of Personality and Social Psychology, 59, 994-1005.
Berscheid, E. Interpersonal relationships. (1994). Annual Review of Psychology, 45, 79-129. Brewer, M. B., & Collins, B. E., (1981). Knowing and validating: Attribute to Donald T. Campbell. San Francisco: Jossey-Bass.
Collins, B. E., & Guetzkow, H. (1964). A social psychology of group processes for decision making. New York: Wiley. Dunkel-Schetter, C. (1984). Social support and cancer: Findings based on patient interviews and their implications. Journal of Social Issues, 40, 77-98.
Harvey, J. H., Weber, A. L., & Orbuch, T. L. (1990). Interpersonal Accounts: A social psychological perspective. Cambridge, UK: Blackwell. House, J. S. (1981). Work stress and social support. Reading, MA: Addison-Wesley.
Mankowski, E. S., & Wyer, R. S. (1997). Cognitive causes and consequences of perceived social support. In G. R. Pierce, B. Lakey, I. G. Sarason, & B. R. Sarason, (Eds.), Sourcebook of social support and personality (pp. 141-163). New York: Plenum Press.
Pierce, T., Baldwin, M. W., & Lydon, J. E. (1997). A relational schema approach to social support. In G.R. Pierce, B. Lakey, I.G. Sarason, & B.R. Sarason, (Eds.), Sourcebook of social support and personality (pp. 19-43). New York: Plenum Press.
Stroebe, W., & Stroebe, M. (1996). The social psychology of social support. In E. T. Higgins, & A. W. Kruglanski (Eds.), Social psychology: Handbook of basic principles.) (pp. 597-621). New York: Guilford.