Using analysis of covariance (ANCOVA) with fallible covariates

Analysis of covariance (ANCOVA) is used widely in psychological research implementing nonexperimental designs. However, when covariates are fallible (i.e., measured with error), which is the norm, researchers must choose from among 3 inadequate courses of action: (a) know that the assumption that covariates are perfectly reliable is violated but use ANCOVA anyway (and, most likely, report misleading results); (b) attempt to employ 1 of several measurement error models with the understanding that no research has examined their relative performance and with the added practical difficulty that several of these models are not available in commonly used statistical software; or (c) not use ANCOVA at all. First, we discuss analytic evidence to explain why using ANCOVA with fallible covariates produces bias and a systematic inflation of Type I error rates that may lead to the incorrect conclusion that treatment effects exist. Second, to provide a solution for this problem, we conduct 2 Monte Carlo studies to compare 4 existing approaches for adjusting treatment effects in the presence of covariate measurement error: errors-in-variables (EIV; Warren, White, & Fuller, 1974), Lord's (1960) method, Raaijmakers and Pieters's (1987) method (R&P), and structural equation modeling methods proposed by S?rbom (1978) and Hayduk (1996). Results show that EIV models are superior in terms of parameter accuracy, statistical power, and keeping Type I error close to the nominal value. Finally, we offer a program written in R that performs all needed computations for implementing EIV models so that ANCOVA can be used to obtain accurate results even when covariates are measured with error.     

The Impact of Covariates on Statistical Power in Cluster Randomized Designs: Which Level Matters More?

Field experiments with nested structures are becoming increasingly common, especially designs that assign randomly entire clusters such as schools to a treatment and a control group. In such large-scale cluster randomized studies the challenge is to obtain sufficient power of the test of the treatment effect. The objective is to maximize power without adding many clusters that make the study much more expensive. In this article I discuss how power estimates of tests of treatment effects in balanced cluster randomized designs are affected by covariates at different levels. I use third-grade data from Project STAR, a field experiment about class size, to demonstrate how covariates that explain a considerable proportion of variance in outcomes increase power significantly. When lower level covariates are group-mean centered and clustering effects are larger, top-level covariates increase power more than lower level covariates. In contrast, when clustering effects are smaller and lower level covariates are grand-mean centered or uncentered, lower level covariates increase power more than top-level covariates.     

When Beliefs Lead to (Im)Moral Action: How Believing in Torture's Effectiveness Shapes the Endorsement of Its Use

Does believing in torture's effectiveness shape the endorsements of its use? Using a multimethod approach across six studies, we provide converging evidence that efficacy beliefs can help increase understanding of individual differences and situational influences on torture support. Studies 1a and 1b found that torture opinions contained more efficacy‐based language than other types of harm and that people relied more on torture efficacy than torture's inherent morality when conveying their views. Study 2 assessed predictors of torture favorability including effectiveness and other key covariates, revealing that efficacy beliefs strongly predicted torture favorability—an association that retained its predictive validity above and beyond individual differences known to influence torture support. Mediation analyses further showed that efficacy beliefs explained key associations with torture support. Studies 3 and 4 used moral dilemmas requiring decisions about torture versus other harm. Results showed that individuals who believed harm would be effective were more likely to endorse its use; this was especially evident for torture judgments. Study 5 replicated the torture‐efficacy effect while also revealing efficacy effects for other interrogation techniques, thus suggesting the effect is driven more by the instrumental objective of torture than harm or moral violations. Theoretical and practical implications are discussed.     

An explanation of conservatism in the bookbag-and-pokerchips situation

An experiment is described in which subjective probability revisions were obtained in a standard probability estimation task, the ‘bookbag-and-pokerchips’situation. Three aspects of probability revision were examined: conservatism, sequential effects, and coherence. Under two experimental conditions, the conservatism effect obtained was closely related to subjects' use of a simple strategy. A recency effect was also obtained. Coherence of the probability estimates was excellent. Conditions under which the observed strategy leads to conservatism are explored and previously published results are reconsidered in the light of this strategy. Conservatism in the bookbag- and-pokerchips situation is explained as an artefact of subjects' strategies     

Type A behavior and stress in college males

Substantial inconsistency can be noted among studies that have considered the relationship between Type A status and subsequent cardiac pathology. This erratic prediction pattern could be partially explained by the failure to consider level of stress in Type As, because only Type As that experience excessive and sustained stress should be at special risk. A stress-vulnerability model for young Type A college students was proposed involving selected Type A characteristics and other personality moderator variables serving to elicit, augment, and sustain stress. We considered self-preoccupation as a stress moderator in Type A college males and found evidence that it serves this function. Self-preoccupied Type As reported excessive stress over the previous year, whereas less self-centered Type As did not. Self-preoccupation was of no importance in moderating stress in Type Bs.     

Type A Behavior and Stress in College Males

Substantial inconsistency can be noted among studies that have considered the relationship between Type A status and subsequent cardiac pathology. This erratic prediction pattern could be partially explained by the failure to consider level of stress in Type As, because only Type As that experience excessive and sustained stress should be at special risk. A stress-vulnerability model for young Type A college students was proposed involving selected Type A characteristics and other personality moderator variables serving to elicit, augment, and sustain stress. We considered self-preoccupation as a stress moderator in Type A college males and found evidence that k serves this function. Self-preoccupied Type As reported excessive stress over the previous year, whereas less self-centered Type As did not. Self-preoccupation was of no importance in moderating stress in Type Bs.     

The Causal Exclusion Argument

Jaegwon Kim’s causal exclusion argument says that if all physical effects have sufficient physical causes, and no physical effects are caused twice over by distinct physical and mental causes, there cannot be any irreducible mental causes. In addition, Kim has argued that the nonreductive physicalist must give up completeness, and embrace the possibility of downward causation. This paper argues first that this extra argument relies on a principle of property individuation, which the nonreductive physicalist need not accept, and second that once we get clear on overdetermination, there is a way to reject the exclusion principle upon which the causal exclusion argument depends, but third that this should not lead to the belief that mental causation is easily accounted for in terms of counterfactual dependencies.     

On the relationship between stimulus intensity and duration of visible persistence

Stimulus intensity and duration of visible persistence have been found to vary directly in some studies but inversely in others. Hawkins and Shulman (1979) have proposed that this inconsistency can be resolved by separating the studies that employed a decrement-threshold measure of persistence (Type I measure) from those that measured the total duration of persistence (Type II measure). They suggested that Type I measures yield an inverse relationship, whereas Type II measures yield a direct relationship between intensity and persistence. Hawkins and Shulman's model is incomplete in ways that are easily remedied. However, the model is totally contradicted by the experimental evidence. A new resolution of the inconsistent results is proposed in terms of retinal afterimages.     

Stress and Distress as a Function of Jenkins Activity Survey-Defined Type A Behavior and Control Over the Work Environment

Although laboratory studies have demonstrated that situational factors which are particularly threatening to Type A individuals induce both Type A behaviors and cardiovascular reactivity, field studies have failed to demonstrate consistent differences between Types A and B in response to stress. Rhodewalt, Hays, Chemers, and Wysocki (1984) argued that this inconsistency is caused by the failure to take into account the individual's perception of the situation. They demonstrated that Type A individuals who perceived high levels of job stress reported more psychological and physical health problems than Type As under low stress or Type Bs under high or low stress. This study attempted to replicate this earlier finding, as well as identify factors in the situation which lead Type A individuals to perceive their job as stressful. Three hundred thirty-six male school principals (categorized as Type A or Type B by the Jenkins Activity Survey [JAS]) completed questionnaires assessing their degree of control in their work environment, perceived job stress, and physical and psychological well-being. A path mediation analysis indicated that Type A individuals reported more job stress under lower levels of perceived control, and that higher job-stress levels subsequently lead to greater incidence of physical and psychological symptoms, especially for Type As.     

The problem of measurement model misspecification in behavioral and organizational research and some recommended solutions

The purpose of this study was to review the distinction between formative- and reflective-indicator measurement models, articulate a set of criteria for deciding whether measures are formative or reflective, illustrate some commonly researched constructs that have formative indicators, empirically test the effects of measurement model misspecification using a Monte Carlo simulation, and recommend new scale development procedures for latent constructs with formative indicators. Results of the Monte Carlo simulation indicated that measurement model misspecification can inflate unstandardized structural parameter estimates by as much as 400% or deflate them by as much as 80% and lead to Type I or Type II errors of inference, depending on whether the exogenous or the endogenous latent construct is misspecified. Implications of this research are discussed.     

Causal reasoning and backtracking

I argue that one central aspect of the epistemology of causation, the use of causes as evidence for their effects, is largely independent of the metaphysics of causation. In particular, I use the formalism of Bayesian causal graphs to factor the incremental evidential impact of a cause for its effect into a direct cause-to-effect component and a backtracking component. While the “backtracking” evidence that causes provide about earlier events often obscures things, once we our restrict attention to the cause-to-effect component it is true to say promoting (inhibiting) causes raise (lower) the probabilities of their effects. This factoring assumes the same form whether causation is given an interventionist, counterfactual or probabilistic interpretation. Whether we think about causation in terms of interventions and causal graphs, counterfactuals and imaging functions, or probability raising against the background of causally homogenous partitions, if we describe the essential features of a situation correctly then the incremental evidence that a cause provides for its effect in virtue of being its cause will be the same.     

Type a behavior,field dependence,and hypervigilance: Toward increased Type A specificity

Field dependence and hypervigilance are two modes of sensory information processing that have been related to cardiovascular functioning. Two studies examined the relationship between these variables and Type A behavior. In both studies, Jenkins Activity Survey-defined Type A and B subjects were assessed for field dependence by the Embedded Figures Test (EFT) and then performed a size estimation task—the measure of hypervigilance. Results indicated that field-dependent Type A's were most hypervigilant. The possible contribution that perceptual style might make to the greater predictive specificity of Type A measures was discussed.Portions of this research were presented at the meeting of the American Psychological Association, Toronto, Canada, August 1984. The authors express their appreciation to Lois Haggard for the design of materials for Study II, and to Rick Smith and Kathy Connell for assistance in conducting Study II.     

Truly Inclusive or Uniformly Liberal? An Analysis of the Politics of the Emerging Church

The emergent church movement has fashioned itself as an alternative for Christians who do not want to walk away from their faith, but feel uncomfortable with the dogmatic conservatism found in mainstream evangelicalism. The emerging church movement has portrayed itself as diverse and inclusive, which is a direct result of evading ingroup‐outgroup boundaries. However, despite the desire for a plurality of opinions, the movement's leaders have been known to take political positions that are largely left‐leaning. We use the first dataset known to gather this identity from a sample of Protestant clergy, and assess whether denominationally connected emergent church clergy do, in fact, present a distinctive political profile. Emergent clergy are what they say they are—diverse and inclusive—while they are, on average, more liberal than nonemergent clergy in the sample.     

Toxic mold and mycotoxins in neurotoxicity cases: Stachybotrys, Fusarium, Trichoderma, Aspergillus, Penicillium, Cladosporium, Alternaria, Trichothecenes

Presented is the argument that psychologists and neuropsychologists have no scientific basis for rendering opinions about causation given the current state of the literature. The critical question is whether in a residence or office inhalation of mold spores or mold metabolites, including mycotoxins, causes neuropsychological impairment or mental and emotional disorders. There has not been sufficient research to support such conclusions. Nonetheless, in the context of litigation, speculative opinions are rendered in lieu of scientifically well-founded conclusions. Resources for recognizing and coping with pseudoscientific arguments are suggested.     

Absence Causation and a Liberal Theory of Causal Explanation

For the framework of event causation—i.e. the framework according to which causation is a relation between events—absences or omissions pose a problem. Absences, it is generally agreed, are not events; so, under the framework of event causation, they cannot be causally related. But, as a matter of fact, absences are often taken to be causes or effects. The problem of absence causation is thus how to make sense of causation that apparently involves absences as causes or effects. In an influential paper, Helen Beebee offers a partial solution to the problem by giving an account of causation by absence (i.e. causation in which absences are supposed to be causes). I argue that Beebee's account can be extended to cover causation of absence (i.e. causation in which absences are supposed to be effects) as well. More importantly, I argue that the extended Beebeeian account calls for a major modification to David Lewis's theory of causal explanation, usually taken as standard. Compared to the standard theory, the result of this modification, which I shall call ‘the liberal theory of causal explanation’, has, among other things, the advantage of being able to accommodate causal explanations in which the explananda are not given in terms of events.     

Mixture distribution latent state-trait analysis: basic ideas and applications

Extensions of latent state-trait models for continuous observed variables to mixture latent state-trait models with and without covariates of change are presented that can separate individuals differing in their occasion-specific variability. An empirical application to the repeated measurement of mood states (N=501) revealed that a model with 2 latent classes fits the data well. The larger class (76%) consists of individuals whose mood is highly variable, whose general well-being is comparatively lower, and whose mood variability is influenced by daily hassles and uplifts. The smaller class (24%) represents individuals who are rather stable and happier and whose mood is influenced only by daily uplifts but not by daily hassles. A simulation study on the model without covariates with 5 sets of sample sizes and 5 sets of number of occasions revealed that the appropriateness of the parameter estimates of this model depends on number of observations (the higher the better) and number of occasions (the higher the better). Another simulation study estimated Type I and II errors of the Lo-Mendell-Rubin test.     

The end of the end of ideology

The "end of ideology" was declared by social scientists in the aftermath of World War II. They argued that (a) ordinary citizens' political attitudes lack the kind of stability, consistency, and constraint that ideology requires; (b) ideological constructs such as liberalism and conservatism lack motivational potency and behavioral significance; (c) there are no major differences in content (or substance) between liberal and conservative points of view; and (d) there are few important differences in psychological processes (or styles) that underlie liberal versus conservative orientations. The end-of-ideologists were so influential that researchers ignored the topic of ideology for many years. However, current political realities, recent data from the American National Election Studies, and results from an emerging psychological paradigm provide strong grounds for returning to the study of ideology. Studies reveal that there are indeed meaningful political and psychological differences that covary with ideological self-placement. Situational variables--including system threat and mortality salience--and dispositional variables--including openness and conscientiousness--affect the degree to which an individual is drawn to liberal versus conservative leaders, parties, and opinions. A psychological analysis is also useful for understanding the political divide between "red states" and "blue states."     

Perceived coercion, resistance to persuasion, and the Type A behavior pattern

Three studies are reported, which tested the notion that persons possessing the Type A “coronary-prone” behavior pattern would be more susceptible to the experience of psychological reactance than would Type Bs, i.e., persons without that pattern. In Experiment 1, subjects exposed to a persuasive communication that was either low or high in coerciveness rated the degree of coercive intent behind the communication. Type As perceived greater coerciveness overall than Type Bs, a finding that was most reliable in the low coercion condition. In Experiment 2, subjects' postcommunication opinions were assessed, so that resistance to persuasion could be measured. A gender difference within Pattern A emerged from this study. Male Type As resisted the persuasion attempt to an equivalent degree whether it was low or high in coerciveness; female Type As displayed comparable resistance in the low coercion condition, but tended to be even more resistant in the high coercion condition. In contrast, both male and female Type Bs displayed persuasion in the low coercion condition, and resisted (to a degree comparable to Type As) in the high coercion condition. Experiment 3 further investigated resistance to persuasion among females, using a different issue for the persuasion attempt. Type A women again resisted more in the high coercion than the low coercion condition. Type B women displayed no evidence of resistance in either condition. Discussion centers on the observed gender difference and the possible theoretical importance of the notion that Type As perceive threat at lower objective levels than do Type Bs.     

A Propensity Score Adjustment for Multiple Group Structural Equation Modeling

In the behavioral and social sciences, quasi-experimental and observational studies are used due to the difficulty achieving a random assignment. However, the estimation of differences between groups in observational studies frequently suffers from bias due to differences in the distributions of covariates. To estimate average treatment effects when the treatment variable is binary, Rosenbaum and Rubin (1983a) proposed adjustment methods for pretreatment variables using the propensity score. However, these studies were interested only in estimating the average causal effect and/or marginal means. In the behavioral and social sciences, a general estimation method is required to estimate parameters in multiple group structural equation modeling where the differences of covariates are adjusted. We show that a Horvitz–Thompson-type estimator, propensity score weighted M estimator (PWME) is consistent, even when we use estimated propensity scores, and the asymptotic variance of the PWME is shown to be less than that with true propensity scores. Furthermore, we show that the asymptotic distribution of the propensity score weighted statistic under a null hypothesis is a weighted sum of independent χ² ₁ variables. We show the method can compare latent variable means with covariates adjusted using propensity scores, which was not feasible by previous methods. We also apply the proposed method for correlated longitudinal binary responses with informative dropout using data from the Longitudinal Study of Aging (LSOA). The results of a simulation study indicate that the proposed estimation method is more robust than the maximum likelihood (ML) estimation method, in that PWME does not require the knowledge of the relationships among dependent variables and covariates.