poets are liars, what about psychologists? (or, why poetry and psychology are not so different)

An interviewer once asked poet Diane Wakoski to explain what she meant when she said some people find her “irascible.”

Here is her (abbreviated) reply:

“…I know that it probably starts from the way I critique other people's poetry … I have a set of criteria for reading poems, and I'm not shy about presenting them, so if people ask for my critical response to a poem, I tell them what works and why, and what doesn't work and why. For some reason, this is taken amiss frequently. I don't know why you would ask a person's opinion about poets or poetry if you didn't expect an honest reply. And I don't know why you would ask me, since I'm a very serious reader, if you didn't expect me to give a rather serious and sometimes difficult response."

Like poetry, psychology must be evaluated. We devise set criteria to effectively evaluate (research on) evaluation methods and treatments. And just like Wakoski's students and hopeful apprentices, we (that is, psychology as a field) seem baffled and suprised when we don't get the response we expect. But this, I think, is an indication we are on the right track.

What do I mean by this? Well, if we just use our judgment - our "clinical judgment" perhaps - we are very well likely to continue to use ineffective therapies, or even worse: treatments that harm. Instead, a move toward ESTs forces us to use more actuarial judgments to evaluate treatments. However, why is it that we are not completely relying on actuarial judgments to evaluate treatment therapies? Because, like poets, (some) psychologists are liars. But, quite unlike poets, our goal is not some psychological mythos or some elaborate pseudo-scientific mythology. Our goal is SCIENCE.

Thus, lies may be the basis of a good poem, but they are not the foundation of good science. And yes, not disclosing information is lying. Garb (1999) discusses this very issue in terms of excluding results / data in meta-analyses and selective reporting. Sure, we want the data to tell a story. But this is quite different than making up a story. If the non-significant results were included in West's (1998) meta-analysis on projective techniques and their utility in detecting child sexual abuse, would the same story emerge? Unfortunately, Garb does not answer that question definitively in this paper, but suggests that the answer is "no."

But perhaps even more unsettling is the fact that we (as psychologists, clinicians) are lying to ourselves. We think we can make good, sound, clinical judgments. But no! The research says otherwise. Actuarial judgments tend to fare better in terms of predictive ability. Yet, we use / rely on these decisions far less. But when we ignore science and just willy nilly come up with a treatment based on clinical experience we get things like attachment therapy. And perhaps proponents of attachment therapy will say it has a firm theoretical base (except that it doesn't). But even in studies of actuarial versus clinical judgment that favor the latter in cases where there is a strong theoretical ground, the social sciences do not meet the conditions for which such results were obtained. We don't really know the major causal influences of (insecure) attachment (although we "know" some and have ideas about others...) and not all of our assessment measures are necessarily accurate and/or precise.

diagnosing APD vs. V-Code

I was super excited to revisit the Moffit article. This was not my first time reading this article (obviously, as I just said I was revisiting it). However, after working in the Allen lab for almost a year now, I found myself viewing this article in a new light. I was able to draw more parallels between the developmental patterns of antisocial behavior (ASB) and developmental patterns of substance use. While these two disorders have some major differences (e.g., substance use is virtually never seen in toddlerhood where as ASB/aggression is, or at least, can be). But there are also some overarching parallels. Substance use and ASB both escalate in adolescence. Additionally, problems associated with substance use experienced in adolescence are likely to decline for some. However, some of this is due to the fact that drinking alcohol is no longer illegal after age 21 in the United States. I would be interested in discussing how these developmental patterns relate to and parallel each other, especially considering the strong relationship between substance use and ASB. However, that's getting off topic a bit, so instead I turn to a discussion of APD within Moffit's framework and in terms of diagnostic criteria.

Antisocial Personality Disorder (APD) is not new to me. For two years, I worked as a research assistant (at Pitt) looking at substance use and antisocial behavior in females (and their male partners). I became intimately familiar with the diagnostic criteria for APD (as well as the Adult Antisocial Behavior V-code). However, knowing the criteria and seeing it in real life are very different experiences.

That being said, I have two issues I want to bring up. First is regarding the diagnostic criteria and a comment made in the Moffit article. Like I said, I had become quite familiar with the diagnostic criteria for APD. So, having temporarily misplaced my DSM (and quite possibly my mind) did not seem like a huge problem (well, losing my mind seems to be a problem...). That is, until I read this on page 678: "Robins (1966, 1978) has shown that there are virtually no subjects with adult antisocial personality disorder who did not also have conduct disorder as children."

Isn't conduct disorder with onset before age 15 a "prerequisite" so to speak, a necessary diagnostic criterion for APD? I realize these studies/papers were published in 1966 and 1978 and thus would correspond to previous versions of the DSM (that is, previous to IV). However, (and this is where my memory is shakey) I don't remember their being a major difference like this between DSM-III and IV. Maybe this was not a criterion in II? I can't imagine that a criterion like having conduct disorder (or not) would be taken so lightly as to say "well, it's ok that Mr. X (or Ms., as the case may be) didn't have conduct disorder as a child b/c he (or she) surely has APD now". And perhaps I am saying this because I know (or at least strongly believe, with empirical evidence supporting such a theory) that the course of APD is different from other antisocial behavior (ASB) that is more contextual.

Which brings me to my second point.

As a research assistant (at Pitt), I hated HATED diagnosing APD. Here's the situation: The Female Adolescent Study (FAS) started when the (female) participants were, you guessed it!, adolescents (Time 1: ages 14-18). However, behavioral and mental health reports were collected (via questionnaire and structured interview) retrospectively from both the teen and the parent / guardian (usually mother) regarding childhood. Thus, information about childhood was collected, but often required the participant to recall things that may have happened up to 18 years prior. (I think we're all aware of how fallible memory is, so I won't go into this issue here.) Thus, our best data about any sort of "childhood" antisocial behavior was about behavior that happened when these girls/women were teenagers. We re-interview these girls at five year intervals twice more, totaling three data collection time points.

The problem is clear if we think about what Moffit is proposing, namely that this developmental taxonomy of ASB requires that we know the participant's or client's preadolescent behavior. For all intents and purposes, I don't think the FAS has a good sense of this preadolescent behavior. But now, we often see these women engaging in ASB. Because the sample was selected based on substance use and ASB / conduct disorder, there is a greater chance of having one of these women meet diagnostic criteria for APD. But even when some of these women seemingly met criteria, as I argued with my supervisor (I tended to stir things up a bit), I do not think they actually "had" APD. Instead, according to Moffit's taxonomy, I would be inclined to say they would fall under the adolescent-limited category. Although, it's not quite that limited as we see from the diagnostic interviews that some ASB continued on to / through adulthood. Instead of some personal trait underlying their ASB, I believe there are contextual (i.e., environmental) factors that can lead ASB to persist past adolescence. However, once taken out of this context, such ASB would diminish. There is no question on the SCID to address this. But, if we take Moffit's proposed taxonomy into account, these women (those whose ASB persists b/c of contextual factors versus those who engage in ASB due to some underlying trait) are likely very different! Putting them both in the same APD category will likely deny differences between the two groups and any analysis using diagnosis as a variable is likely to yield inaccurate information - about both groups!

Yummies for Miss Bea (or Is a Deaf Cat Louder than Fear?)

I really liked the articles for this week. The idea of multiple vulnerabilities for a disorder is becoming more popular – but for good reason: it’s true! That being said, the notion of a genetic vulnerability paired with generalized and specific psychological vulnerabilities was not new to me.

What struck me was the differentiation between anxiety and panic. Of course these are similar concepts. But they are also different, but that difference was not always readily apparent to me. Thus, the explanations offered in the two articles offered some clarity.

First, I thought that Barlow’s term “anxious apprehension” for anxiety captures exactly what anxiety is quite well. The anxiety is always about something that might happen, that could happen, that will happen – but (usually) not about what is happening right this very second. Think about it. (I did.) Even if you are experiencing, say, performance anxiety, you are worried about what might happen during the performance – not what is happening at that second. I invite you to prove me wrong here. Is there a situation where the anxiety is about what is occurring right this very second? I say “no,” and so does Barlow who characterizes anxiety as a state of helplessness due to a “perceived inability to predict, control, or obtain desired … outcomes in … upcoming personally salient situations or contexts” (p. 1249, italics added).

What, then, is panic? Panic can be thought of as analogous to fear in that the threat is immediate. Thus, anxiety occurs in anticipation of fear or panic.

So, knowing this, I wanted to know how this all relates to anxiety disorders – especially Panic Disorder. This diagnosis has always been a little tricky for me to wrap my head around, and I’m not sure why. I understand the fundamentals of its manifestation (I have seen it) and I understand the diagnostic criteria (I think), but perhaps it was putting the two together that didn’t mesh well. And perhaps this is an artifact of the current diagnostic system. (I must admit, as an RA at Pitt, when giving the SCID and subsequently writing up diagnostic reports Panic Disorder – whether with or without Agoraphobia – was my arch nemesis.)

Imagine you are in your home. Imagine you haven’t left for days, weeks. You can’t. But you're out of [soy milk / carrots / cereal / toothpaste / tampons / some other necessity]. You need to go out. You need to go to the store. But you can’t go. You can't. You just can’t. Knowing that you need [insert necessity here] makes your heart race. You’ll need to go to the store. You need to drive to the store. What if there’s traffic? What if you need to make a left hand turn? What if there's construction and you don't know where to go? You’re having trouble breathing now. What if the store is crowded? What if you can’t find what you need? What if a store clerk asks to help you? What if you get anxious while you’re at the store? Of course you’ll get anxious while you’re there, you’re sweating already. What if you get so nervous you buy the wrong thing? What if you have a panic attack? What if you have a panic attack and you have to leave? What if you have to leave and you don’t get what you came for? What if you don’t get what you came for and you went out for nothing? And if you don’t get what you need, you’ll have to do it all again.

And what if you never get it right?

Ok. So here’s my issue. In the based-on-real-life-but-made-up example above, there is anxiety about having anxiety and anxiety about having a panic attack and during this anticipatory anxiety, panic-like symptoms appear. So is this panic or anxiety? I feel like I could make the case for either. I understand the distinction theoretically, but in real life not so much.

Another issue which I wish we covered is what it is, once the anxiety and panic is set in motion, that allows the person to function. For instance, the person in the above example had not left the house for about a month. Then this person was asked to cat-sit (for an extended period of time) for a deaf and equally anxious older feline whom we shall call Miss Bea (which is indeed the cat’s real name). Eventually Miss Bea started running out of food. Now, Bea, when she wants something, will ask for it. And being deaf, it seems as though she tries to be loud enough so she can hear herself. She can’t. But when she meows at the volume of Bea, the sound is ENORMOUS. Whatever it was, whether it be the enormity of the need or of the bond, Unnamed Anxious Person went out to get food. And continued to go out several times a week for the next few weeks to get something for this cat: special treats, toys, a comb, etc.

And so it goes Miss Bea is not my kitty anymore. Indeed, the Named Anxious Kitty will sit on the Unnamed Anxious Person’s bed and meow nervously (at least once a week) until her new (still nervous) owner returns from the store.

But… this unnamed anxious person is still anxious. Functionally there has been an improvement, but overall anxiety has improved little. Which brings me to the question that I had at the beginning of this class (and still has not been adequately answered): What does it mean to be “better”? What is improvement?

...i'm too sick to think of a title...

It is becoming quite clear that context, especially the interpersonal context, is important in terms of the maintenance and relief of a number of syndromes / disorders / presenting problems / what-have-you. Last week the Marlatt article touched on the importance of context in terms of drug use and addiction and here in the Coyne chapter about depression it has come up again.

The Coyne chapter brought up some particularly interesting issues. A more or less minor issue in the chapter – but a major issue for me – is regarding attachment style ( ß note italics) and depression. The first mention of this appears on page 368 where Coyne reports about a study that asked college students to fill out a questionnaire on attachment style. At this point you may be saying to yourself, “There are those pesky italics again. Why am I taking note?” Coyne almost explained why a questionnaire on attachment is problematic. Almost.

In the attachment literature, there are more-or-less two ways of conceptualizing attachment: the right way and the wrong way … er … I mean, the developmental approach and the social approach. The developmental perspective of attachment is often considered more “pure” in terms of Bowlby’s theory and even the measures of attachment in both childhood and adulthood reflect this. That is, adults are not asked to access and report on their internal working model (IWM), as this would be counter-intuitive in the sense that the IWM consists of unconscious processes that guide behavior. This would be further problematic for the Avoidant / Dismissing sub-group in that they would have little insight to their dismissing/avoidant behavior. Instead, adults are asked questions that tap into their internal model and their discourse is what largely reveals their current states of mind in regard to early attachment relationships. Note here that the emphasis is on early attachment (and the current state of mind).

The social approach looks largely at adult attachment, especially regarding romantic relationships. Measurement is usually done via self-report questionnaires. Although the resulting attachment categories were derived from Bowlby’s work, this sort of measure largely ignores the notion that, especially for those with insecure attachments, the attachment system works hard to keep unconscious processes unconscious. Even more to the point, as we see in the sample statement in the Coyne paper, participants are usually not asked to think about early relationships. In this case, most people would respond as they are now – which may or may not reflect their attachment in childhood.

That being said, it is also important to note that there are mixed results in terms of the correlation between the AAI (developmental measure) and self-report measures, such as the Haven & Shaver (1990) measure. Some studies report a modest correlation, while other report none at all. Obviously these two types of measures are getting at different things. Thus, just because a measure claims to measure “attachment” does not mean that it is measuring “a lifelong trait that developed early in childhood” (p. 369). Indeed, attachment stability can be variable. Even Bowlby proposed the internal mechanism of attachment as working.

**********************************************************************************

so i'm reading ppl's blogs this morning and looking at comments and notice: half of my response is missing! (as allison knows due to a psyc 250 syllabus mishap, i have issues cutting and pasting text.) additionally, there is some weird sort of beta looking character in the first half that should have been an arrow ... i'm not sure what happened there either. below is the rest...

***********************************************************************************

Back to what I was saying…

The literature does suggest that early insecure attachment is a major risk factor for depression, but I would like to argue that this is not what the Roberts, Gotlib, and Kassel (1990) study is saying. Instead, it is looking at more current relationships / relationship experiences. However, I agree with Coyne that perhaps an additional measure that directly addresses current experience is necessary. Coyne goes on to address two more studies that were reported on in the Roberts et al. article, but this part seems unclear to me. I’m not sure why Coyne thinks that effects of attachment should show up in an initial distress score. Theoretically, we may see differences, but it is probably more likely that any given person is likely to experience varying levels of distress at any given time and that distress level at one time point would not be indicative of anything. Because in adulthood attachment can be better conceived in terms of emotion regulation, it would make more sense that attachment is linked with duration of distress (rather than a score at a single point in time). For example, insecure attachment would likely be related to prolonged distress (that is, inability to relieve distress). However, again I agree with Coyne in that the data are more suggestive of how the participants’ relationship experiences (esp. as influenced by attachment) affect level of distress. Indeed, the effect of attachment on current relationships / relationship experiences (rather than early experiences in and of themselves) may be why insecure attachment is a risk factor for depression.

Another issue that Coyne brought up that particularly intrigued me was regarding the “kinds of people” (p. 371) who are drawn to or tolerate people prone to depression and with “stormy” interpersonal styles. I am not aware of any literature on the topic, but I do believe this is an area that warrants attention. Indeed, as important as the interpersonal context is in depression, including support systems, it would be important to examine whether or not people with or vulnerable to depression are more likely to interact (in close relationships) with certain types of people and how these characteristics interact with vulnerabilities to depression. I think it is also important to consider the “other” person’s / people’s vulnerability to depression. As Coyne mentioned, with repeated depressive episodes, a patient’s social support system becomes taxed and is less likely to respond sympathetically, supportively, etc. Obviously, this has important implications in terms of remission and relapse. (This also closely parallels what Marlatt touched on last week in terms of substance users’ non-using social support systems breaking down with increased drug use.)

One last thought here about the interpersonal context (in terms of depression, or even other disorders such as SUD) is that we may be able to gain a wealth of insight by looking at some of these issues in terms of relationship research, a la Hinde. Looking more closely at particular relationships may help identify contexts and experiences that are more likely to elicit onset of depression, relapse, or remission.

desperate deranged talking in my sleep again … breakdown amidst the mixtures avoid addictive plea

I am somewhat familiar with the substance use literature; however, I have always approached the issue in terms of what factors contribute to initial use, problem use, and the continuity of problem use. While discontinuity is inherently intertwined with continuity of use, it is a different perspective. It is quite different to think about what triggers relapse (and how to prevent it) than to think about what turns use into problem use. However, there is quite a bit of overlap as well, especially when it comes to abstinence vs. harm reduction. (But I’ll save this for another time…)

Despite the fact that I had to constantly remind myself that I was looking at a relapse model rather than a use (per se) model, I found multidimensional model outlined by Witkiewitz & Marlatt (2004) fairly comprehensive. While not explicitly stated, it does encompass some important factors that are often ignored. For instance, substance availability is a huge factor in terms of frequency and amount of use, and I would assume in terms of potential relapse as well. While the article indirectly implicates substance using friends, peers, etc. as a high-risk factor in terms of relapse, it may also be useful to think about the availability of a substance locally and regionally. Of course, this (likely) has greater implications for illegal drugs than for alcohol and tobacco. There can be quite dramatic fluxuations in the availability of street drugs, which does indeed influence individual use. For instance, there was a time in Pittsburgh (sometime in the 90s, I believe) when heroin was widely available and relatively inexpensive. There were more heroin users at that time in the area than there had been before, and consequently more deaths. When the supply dwindled, drug use changed. Anecdotally, I have been told that this lead some people to stop using heroin altogether (and did not substitute another drug) while others began using other drugs (e.g., benzodiazapines) to simulate or replace the high once provided by heroin. In terms of relapse, when the market is flooded with a cheap substance there is a greater likelihood that a person will be in a high-risk situation where availability and therefore use / relapse is more probable. When thinking about illegal drugs it may also be important to consider the potency of the drug locally and regionally available. As Witkiewitz & Marlatt suggest, relapse is likely to occur – especially early on in or after treatment. When the drug available is highly potent and highly addictive, as was the case about a year ago when Midwest and Eastern areas were flooded with heroin cut with phentonyl, the potential to continue to use after the first relapse is heightened. Not only is the likelihood of continued use heightened, so is the likelihood of death.

However, this is not to downplay the importance of affect, coping behavior, and social support. I believe these factors are hugely important – and not just because I willy nilly decided this was so! There is a whole literature on coping behavior and alcohol/drug expectancies that agree with me! (or maybe I agree with the literature?) Turning the relapse issue around a little bit … Kandel and Ravies (I think) found that substance use was more likely to continue and to become problem use when the substance was used to cope with negative affect versus for social reasons. While this research looked at the continuity of use (rather than treatment and relapse specifically), it seems logical that those who had been using substances socially, even though perhaps problematically, would be in a better position to avoid relapse, especially in terms of coping. In this case, coping is fundamentally paired with substance use for one group and not for the other, implying that the “social” group may have better coping resources. Litt et al. (2003, as reported in Witkiewitz & Marlatt, 2004) report that the relationship between “readiness to change” (I’m assuming here this means quit or reduce substance use) and drinking outcomes is mediated by coping.

…and no I have not been sitting here (in my bed) typing this response with a cigarette between my lips wondering how much more I’m going to write before I can go outside.

*Title taken from Skinny Puppy's Addiction, lyrics by Nivek Ogre.

further up the spiral?

The Jacobson et al., 2001 article gave me a positive outlook on the contextual treatment approaches. I was particularly interested in the idea that Behavioral Activation (BA) does not focus on a individual illness / mental illness model whereby unseen, internal pathology is targeted in treatment; rather BA takes into consideration the environmental context of the individual and treatment targets overt behaviors. This seems to be advantageous for several reasons.

Contextual approaches such as BA may reduce number of assumptions (therapists / researchers / clients / others???) need to make. The goal is not to change unseen cognition, which may be difficult to decipher/interpret. Thus the focus on observable behavior reduces the need to interpret the relation between external environmental events and internal cognitions

[however, functional analysis – foundation of treatment – is largely based on assumptions…]

It may also be that, in treatments founded in contextual approaches, it is easier to measure incremental changes in both clinical and research settings. Because we can observe differences in behavior over time, we are not trying to guess whether or not an unobservable cognition has changed (even if that assumption is based on behavior). Important to note that cognitions are not denied; they simply are not the focus of treatment.

However, I question the extent to which cognitions are (or are not) changed by behavioral activation. Does targeting and changing such behaviors as avoidance behavior only change behavior or are there underlying changes to cognition? If cognitions do change, what are these changes and what is the process of change? Do cognitions quickly change such that once positive behavior routines are established we can assume more positive underlying cognitions? Or is the process more slowly such that we may see observable changes in behavior long before there are true changes in cognition?

--and you, Albert Ellis, what were you doing down by the watermelons?

I was particularly intrigued by Ellis’ point #3 on page 156:

Family, social, and cultural influences learned at a young age may strongly encourage people to engage in behaviors that, under different conditions would not lead to addiction. For example, in order to win the approval of peers, youths may ‘willingly’ addict themselves to smoking, drinking, and criminal acts.

While I believe that family, social, and cultural influences may in fact encourage people (esp. adolescents) to engage in destructive behaviors like substance use (and my current research is pointing in this direction), I am not sure I am willing to accept Ellis’ explanation that this is due to a difficulty in distinguishing between strong desires and rigid demands.

I believe that if adolescent substance use was not so culturally normative, we would see less of it overall. But I do not think that this would greatly reduce the number of those that go on to develop substance use disorders (SUDs). To say that there would not be some other circumstances that would likely elicit destructive, addictive behavior would be naïve. We already have a plethora of evidence implicating the importance of biological predispositions and emotional dysregulation in the onset of substance use disorder (note here, I mean addiction, not substance use itself). Thus, I agree with Ellis that there are biological and environmental factors that are associated with the onset and maintenance of self-destructive behaviors like SUD. However, I am not convinced that thinking about it in terms of desires and demands is theoretically correct or even helpful.

That being said, I was disappointed that the meta-analysis did not provide evidence regarding SUD treatment. Ellis would probably be inclined to say that cognitive-behavioral treatments like RET would in fact be effective in treating SUDs, so long as the client was willing to “choose to do the hard work and practice that is usually required for change” (p. 156). I, myself, am inclined to think that some sort of CBT would in fact work for those with SUDs whose use was motivated by emotion regulation. However, there are indeed cases where emotion regulation plays little to no role in the maintenance. In these cases, I am not sure whether or not CBT would be effective.

This leads me to the question of whether or not understanding etiology would make an important impact on the efficacy of CBT.