A comparison of methods to test mediation and other intervening variable effects

A Monte Carlo study compared 14 methods to test the statistical significance of the intervening variable effect. An intervening variable (mediator) transmits the effect of an independent variable to a dependent variable. The commonly used R. M. Baron and D. A. Kenny (1986) approach has low statistical power. Two methods based on the distribution of the product and 2 difference-in-coefficients methods have the most accurate Type I error rates and greatest statistical power except in 1 important case in which Type I error rates are too high. The best balance of Type I error and statistical power across all cases is the test of the joint significance of the two effects comprising the intervening variable effect.     

Flank transparency: transparent filters seen in dynamic two-color displays

Flank transparency is the perception of a colored transparent filter evoked by apparent-motion displays containing as few as two colors. Displays of flank transparency contain a random array of line segments placed on a uniform background. Small flanks are added to the line segments if the segments fall in the interior of a moving virtual shape, such as a virtual disk. This leads to the perception of a colored transparent disk with well-defined boundaries moving over the array of lines. Current qualitative and quantitative models of luminance and color conditions for perceptual transparency do not account for flank transparency as they require displays containing at least three different colors.     

Perceptions of an old female eyewitness: is the older eyewitness believable?

Young adults rated an old female witness (82 years) to be less competent but more honest than a young female witness (28 years). The effect of age stereotyping on believability was assessed indirectly by using a variant of Loftus' misinformation paradigm. Regression analysis showed that the more competent a witness was perceived as being, the more influential was the misinformation she provided. This was true, however, only when the witness was young because the relatively lower competence of the old witness was associated with nonsignificant misinformation effects. These data suggest that negative beliefs associating incompetence with old age may compromise the believability of older eyewitnesses.     

The generic-viewpoint assumption and illusory contours

Visual images are ambiguous. Any image, or collection of images, is consistent with an infinite number of possible scenes in the world. Yet we are generally unaware of this ambiguity. During ordinary perception we are generally aware of only one, or perhaps a few of these possibilities. Human vision evidently exploits certain constraints--assumptions about the world and images formed of it--in order to generate its perceptions. One constraint that has been widely studied by researchers in human and machine vision is the generic-viewpoint assumption. We show that this assumption can help to explain the widely discussed fact that outlines of blobs are ineffective inducers of illusory contours. We also present a number of novel effects and report an experiment suggesting that the generic-viewpoint assumption strongly influences illusory-contour perception.     

Structural differences in parent and child reports of children's symptoms of depression and anxiety

Two cohorts of public elementary school children and their parents (assessed 3 years apart) completed child and parent forms of the Children's Depression Inventory (CDI) and the Revised Children's Manifest Anxiety Scale (RCMAS). Assessments were conducted twice, once during the fall (N = 562) and again during the spring (N = 630) of the 6th grade. Factor analyses revealed 3 factors for each measure. Two of the 3 parent CDI factors manifested some degree of congruence with their counterparts from the child CDI. Similarly, 2 of the 3 RCMAS factors were somewhat congruent across informant types. Differences between parent and child factor structures suggest that parents' and children's reports focus on somewhat different aspects of child psychopathology, and they can make qualitatively different contributions to the multiaxial assessment of children.     

Got Pain? Racial Bias in Perceptions of Pain

In the days of slavery, White people assumed that Black people felt less pain than did White people. This belief was used to justify slavery; it was also used to justify the inhumane treatment of Black men and women in medical research. Today, White Americans continue to believe that Black people feel less pain than do White people although this belief has changed from its historical, explicitly racist form. Racial attitudes do not moderate the bias, suggesting that racial bias in perceptions of others' pain is not rooted (solely) in racial prejudice. Moreover, Black Americans too believe that Black people feel less pain than do White people, suggesting that the bias is no longer grounded in intergroup dynamics. Rather, contemporary forms of this bias stem from assumptions that Black people face more hardship and “thus” can withstand more pain and assumptions that Black people's bodies are not only different but also superhuman. Although this new instantiation of the pain perception bias is decidedly more “benevolent”, it can nonetheless lead to negative outcomes. Here, we consider how racial bias in perceptions of others' pain may affect racial disparities in health care and intergroup relations more generally. We also discuss potential avenues for interventions aimed at preventing this bias among children and reducing this bias among adults.     

Extending the Validity of the Personality Assessment Inventory's (PAI) Level of Care Index (LOCI) in Multiple Psychiatric Settings

The purpose of this study was to extend the validity and clinical application of the Level of Care Index (LOCI) from the Personality Assessment Inventory (PAI) in 2 independent psychiatric samples. In Study 1 (N = 201), the LOCI effectively differentiated level of care (inpatients from outpatients), and was also meaningfully associated with risk factors for psychiatric admission (e.g., suicidal ideation, self-harming behavior, previous psychiatric admission, etc.), even after controlling for other demographic variables (range of Cohen's ds = 0.57–1.00). Likewise, the LOCI also incremented other risk indicators (suicide and violence history) and relevant PAI indexes (i.e., Mean Clinical Elevation, and Suicide and Violence Potential) in predicting level of care, and explained an additional 6% to 12% of variance in the target variable. Diagnostic efficiency analyses indicated LOCI scores in the range of 15 to 18 optimize positive and negative predictive power, and classification rate. In Study 2 (N = 96), the LOCI was found to be significantly higher in those with a recent psychiatric admission within the past 6 months (d = 0.64), as compared to those without an admission. Similarly, those who were admitted for suicide risk had significantly higher mean LOCI scores as compared to those who did not (d = 0.70). The clinical implications of these findings and potential application of the LOCI are discussed.