Archive for February, 2018
Research Summaries: The Effect of Self-Distancing on Adaptive Versus Maladaptive Self-Reflection in Children0
Today’s research summary is based on a shortish paper [pdf] by Angela Duckworth et al (Walter Mischel of Marshmallow effect fame is a co-author!) which focuses on how viewing oneself from a distance, or from a third person perspective, a previous emotional experience, can lead to better and more adaptive outcomes.
- Bad stuff happens. And we make it worse by brooding about it. There is some research that shows that thinking or ruminating about negative experiences can lead to bad outcomes in the present like compromised health or impeded cardiovascular recovery following exercise etc. Ruminative thinking style is known as a precursor and risk factor for depression.
- On the other hand there is a rich tradition of expressive writing (for e.g. Pennebaker’s work) in which people write about their negative experiences and traumas and seem to benefit (boosts in long term mood and well-being) from such an expressive act.
- Different sort of mechanisms are hypothesized in both the above cases. In the first case, one may be reliving the negative experience or recounting it and thus get overwhelmed once more in the present by such a recollection. In the second case, one may be reinterpreting the situation and making fresh sense of the events or reconstruing the events. So reflecting in a negative experience per se may not be bad or good but may lead to a good outcome only when reconstruing happens more than recounting.
- Putting a distance between oneself or seeing events from a detached third person perspective have been shown to increase one’s self control and control one’s impulses and also helpful in alleviating depression by enabling better cognitions. It has been hypothesized that self-distancing or viewing things form a detached third person perspective will lead to better and more adaptive outcomes while self-reflecting, as one will not recount or relive the experiences but will be better able to reconstrue or make new sense of the experiences.
- The current study looked at ~ 100 fifth grade students and asked them to recollect a negative angry outburst/ interaction which was interpersonal in nature. They were then instructed either to feel the event as of it was happening in the present and they were at the center of the action, or that they were watching the event unfold from a distance and observing the distant self. After they had recalled the experience in both conditions, they filled a brief survey measuring their emotional reactivity (how much power the vent still holds over them) and avoidance behavior (do they avoid talking/ thinking about that issue) . They were also asked to write an essay about their reflection and the essay was content analyzed for recounting thoughts, reconstruing thoughts and blame attributions.
- The results showed that when you put a distance between self while recollecting a negative experience, then the emotional reactivity is lesser than when you feel as if you are reliving the experience. Thus, if you want to make a negative experiences hold smaller on you recollect it while putting a distance from self. Thus it was clear that self-distancing was a more adaptive outcome.
- They also found that those students who had put a distance between their earlier self while reflecting on their angry interaction, had fewer recounting statements in their essays and more reconstruing statements. They also made fewer blame attributions.
- They also did a path analysis and found that self-distancing had its impact on more adaptive outcomes (less negative affect and emotional reactivity) via the mediating variables of more reconstruing statements than recounting statements, which in turn led to lesser blame attributions and thus a closure that led to lesser emotional reactivity.
- The take home message, children can benefit form self reflective exercises that make them reflect on negative experiences as long as they are supported in putting a distance between themselves and their past self, so that they don’t merely recount the experience but are able to reconstrue the experience.
Overall, a pretty decent paper [pdf] that stresses the importance of self-distancing while reflecting about past negative experiences.
Today’s research summary looks at another paper [pdf] by Angela Duckworth et al this time focusing on whether it makes sense to include personality variables in long national longitudinal surveys/studies like the MIDUS/ Dunedin/ HRS.
- Personality differences can be conceptualized to be either differences in ability (like cognitive ability), traits (stable patterns of thinking, feeling, acting) , motives or narratives and this paper focuses on traits to the exclusion of other measures of personality. Even in traits, the traits of concern are the Big Five traits of Neuroticism, Extraversion, Agreeableness, Conscientiousness and Openness.
- Personality, in general, and these traits, in particular, are known to predict a range of outcomes like health, achievement, and relationships. The authors believe that large panel surveys should measure these traits to find the correlations with other outcomes being measured. They review research on how traits predict wealth and health and are predicted by underlying genetic polymorphisms or variations.
- For elaborating the association between traits and genes they look at candidate gene studies as well as GWAS. Extraversion is associated with polymorphisms in Dopamine subsystem related genes. Nueroticism is primarily associated with serotenergic genes. Agreeableness and Conscientiousness are both affected by polymorphsism in genes related to dopamine as well as serotonin. Openness to COMT variation. Read the paper to get additional nuances.
- When it comes to economic outcomes, more introverted and more emotionally stable (less in neuroticism) individuals were more likely to save over the lifetime and borrow less; reverse was found for those high in agreeableness. Emotional stability was the best predictor of earnings; extraversion had a complex relation but overall positively predicted earnings; while agreeableness had a very slight negative impact on earning.
- In terms of health, traits like Conscientiousness had a direct effect on health as well as indirect effects mediated by healthy behaviors and educational attainments. In general it is safe to conclude that personality traits do not affect health outcomes directly but by their impact on problematic or protective behaviors. Personality traits have also been linked to mortality.
- The authors recommend that personality traits should be measured in large panel studies, and measured as far as consistently, using say BFI, so that they can be used to predict important life outcomes. Moreover they recommend that as personality traits can change , they should be treated as dependent variables too and measured in each subsequent measurement time.
- One recommendation they have is to keep such trait measures short and relevant; also they recommend multiple measures using informant reports or cognitive tests like go-no go task. However I ‘m not sure if that may be practical in large surveys.
- They also highlight the concerns about ‘flush-right’ responding where some unmotivated participants who are juts going through the motions of filling the survey may keep choosing the extreme right option making the survey results suspect. The instruments should have something inbuilt to detect such responding just like one detects social desirability.
Overall its a pretty decent paper to understand some of the antecedents (genetics) as well as consequents (health and wealth) of Big Five traits and makes a strong case for incorporating big five measures in such large scale studies and surveys. Check the paper here [pdf] .
Research Summaries: Establishing Causality Using Longitudinal Hierarchical Linear Modeling: An Illustration Predicting Achievement From Self- Control0
Today’s research summary is slightly technical. It is based on this paper [pdf] by Angela Duckworth et al that shows a causal relation between self-control and academic achievement.
- Some personality variables like self-control predict important life outcomes. It is well know that self-control as measured at age 4 (using the marshmallow test) can predict important life outcomes years later. However, prediction may not imply causality as a third factor may be responsible for causing both the phenomena under consideration.
- The test for causality is a) causal variable must precede the effect in time; b) the causal variable and outcome variable should be correlated; and c) any third party confound or variable should be ruled out. This is easy to achieve in double blind randomized placebo controlled experiments, but personality traits like self-control are hard to manipulate as trait variables in experimental settings.
- Typically personality traits and their outcomes are studied using a longitudinal study design where changes in say self control at time T1 are correlated with outcomes like academic achievements at a later time T2, of course measure other confounding variables and factoring their effects; thus self-control, along with IQ, may be measured at the beginning of a school session and at the end of session the CGPA obtained will be used to find whether and how much self-control led to academic achievement. This however cannot establish causality in a strict sense as not all variables of interest can be identified and measured. Often the dependent variable (CGPA in our case) is itself controlled for to ensure that a higher CGPA at point T1 does not lead to higher CGPA at time T2 independent of self-control at T1.
- To take care of third party confounding variables, Angela et al used growth curve analysis with Hierarchical Linear Modelling (HLM). This involves taking multiple measures of say self -control at different times and also multiple measures of the outcome say CGPA. The independent variable is considered a time varying co-variate and used to figure the within-person relationship between the two variables of interest. Consider a between subjects confound like socio economic status (SES) that could potentially lead to different outcomes (CGPA) – if not controlled for the self control- CGPA relation arrived at by analysis of between subjects data might lead to erroneous conclusions. However, a stable thing like SES (which doesn’t change with time and is constant for an individual) will have no impact on the correlation or causal relation between how changes in self-control affect CGPA over time in the same individual.
- The direction of causality can also be ascertained by using HLM with reversed time lagged, time varying co-variates. What this means os that we can try to see of the causal arrow runs in other direction by taking measures of CGPA as predictor and self control as outcome variable.
- In this study, self control was measured using self-report, parents and teachers ratings of students for four consecutive academic years (as they moved from fifth grade to eighth grade) using the Brief Self-Control Scale ; CGPA was measured each year as the outcome variable. Self-esteem and IQ was also measured and so was gender, ethnicity etc.
- They found that self control measured 6 months earlier predicted CGPA six months later; average self-control predicted the baseline CGPA as well as the slope of CGPA changes (how fast the CGPA increased or decreased over time). Howsoever, the reverse analysis whereby short term CGPA was used to predict self-control gave negative results thus establishing the causal direction.
- It was thus established that self-control does indeed cause or lead to higher academic outcomes like higher CGPA. A limitation of the study was that a time varying third variable that increased and decreased in tandem with self-control can still account for the relationship between self control and academic achievement.
I liked the paper, though its more methodological. You can find the full paper here [pdf].
Pathological mental health problems in children and young adults have been classified into externalizing (substance abuse, conduct disorder etc) and internalizing disorders (depression , anxiety etc). Today’s post will try to work out the structure of this internalizing spectrum.
The first major difference, that is made in say DSM, is between Mood disorders (disturbance in mood) and Anxiety disorders (characterized by anxiety and avoidance behaviors) . However, Watson in this article (pdf) emphasizes that this classification is not proper and in many cases these disorder say depression (say MDD) and Anxiety (say Panic disorder) are co-morbid with each other.
To explain this as well as other genotypical and phenotypical findings, Watson has developed a structure of these ‘distress disorders’ – however the road was long, an intermediate stop was tripartite model of depression/anxiety.
According to this tripartite model (developed by Watson and Clark), both depression (MDD, dysthymia etc) and anxiety disorders (phobia, panic etc) share a common non-specific factor called Negative Affect (NA) which is characterized by things like preponderance of negative emotions like sadness, fear, guilt, anger etc as well as irritability, difficulty concentrating etc.
Depressive disorders meanwhile are specifically characterized by lack of Positive Affect (PA), which means less emotions like happiness, interest etc, but also Anhedonia or inability to derive pleasure from earlier pleasurable experiences. Anxiety disorders, on the other hand, are characterized by physiological hyper arousal (PH) (shortness of breath, dizzyness etc) .
This model however was also found wanting and replaced with an hierarchical integrative model, which posited that there was a generic non-specific factor of NA common to both anxiety and depressive disorders, and a lower order low PA factor characterizing depression and more specific multiple low order factors (instead of one PH hyperarousal factor) associated with the different types of anxiety disorders like panic/ agoraphobia, Phobia-specif stimuli, phobia social etc .
However , Watson further modified the structure and came up with this model shown below: One broad factor of distress/NA; two specific factors of anxious-misery and fear and then further unique factor specific to individual diagnosis.
To summarize and also extending it a a bit,
- At top there is an internalizing spectrum and associated with it a non-specific NA factor.
- In middle there are four spectrum:- a depressive spectrum , a Fear spectrum and a bipolar spectrum and an Obsessive compulsive spectrum.
- each of these can be further divided into discrete diagnosis along two factors/dimensions (I will not eb focusing too much on bipolar or OCD for the purposes of this post) :
Lets focus more closely on Depressive and Fear Spectrum and try to see alignment with ABCD model. MDD/Dysthemia imho are mainly about mood or Affect; GAD/PTSD are more Cognitive (reaming stuck in a thought loop) ; Panic/agorophobia more Physiological or Dynamic in nature and Phobia (both specific and Social) more Behavioral in nature (avoiding people, places and animals).
Each of these in turn splits into four factors; for ex PTSD splits into four factors- Dysphoria (A), Intrusions (C), Hyperarousal(D) and Avoidance (B). Similarly, recent research has shown that MDD is itself heterogeneous made up of four neural subtypes- one way to list those would be – marked primarily by Anhedonia (A), Anxiety (C) , Psychomotor retardation (D) and Fatigue (B) . Similar analysis should be possible for other discrete diagnosis.
For now, we will turn to the structure of Bipolar and OCD spectrum by analogy to dep/anxiety spectrum.
- Biploar spectrum:
- Euphoria (Affective)
- Flight of ideas (Cognitive)
- OCD spectrum
- Obsessions (Dynamic)
- Compulsions (Behavioural)
Within this OCD can be seen to be comprising of four factors: Hoarding (A?) , Order and symmetry (C), Obsessions and Checking (D) and Washing and cleaning (B).
Another way to think about the depressive and anxiety spectrum is to say that Depression rgoup 1 is characterized by Low PA, depression group 2 by high PH; Fear group 1 by High PH and Fear group 2 by low PA. What distinguished Fear spectrum from Depression spectrum is the fact that much more variance is explained by High NA for depressive syndromes and only moderate variance explained by NA for Fear syndromes.
What do you think is missing from the above model? Where might it be wrong? where might it be correct? If correct what are the implications?