标题：Fifty psychological and psychiatric terms to avoid: a list of inaccurate, misleading, misused, ambiguous, and logically confused words and phrases
source：journal.frontiersin.org | by Scott O. Lilienfeld, Katheryn C. Sauvigné, Steven Jay Lynn, Robin L. Cautin, Robert D. Latzman and Irwin D. Waldman
欲翻译的小伙伴，可评论『领稿』，并把已完成的译文『贴在评论』处或『发在自己的心理圈内』。请领稿的小伙伴，尽量在一个星期内完成翻译，谢谢！！~~
(16) Love molecule. Over 6000 websites have dubbed the hormone oxytocin the “love molecule” (e.g., Morse, 2011). Others have named it the “trust molecule” (Dvorsky, 2012), “cuddle hormone” (Griffiths, 2014), or “moral molecule” (Zak, 2013). Nevertheless, data derived from controlled studies imply that all of these appellations are woefully simplistic (Wong, 2012; Jarrett, 2015; Shen, 2015). Most evidence suggests that oxytocin renders individuals more sensitive to social information (Stix, 2014), both positive and negative. For example, although intranasal oxytocin seems to increase within-group trust, it may also increase out-group mistrust (Bethlehem et al., 2014). In addition, among individuals with high levels of trait aggressiveness, oxytocin boosts propensities toward intimate partner violence following provocation (DeWall et al., 2014). Comparable phrases applied to other neural messengers, such as the term “pleasure molecule” as a moniker for dopamine, are equally misleading (see Landau et al., 2008; Kringelbach and Berridge, 2010, for discussions).
(17) Multiple personality disorder. Although the term “multiple personality disorder” was expunged from the American Psychiatric Association’s (1994) diagnostic manual over two decades ago and has since been replaced by “dissociative identity disorder” (DID), it persists in many academic sources (e.g., Hayes, 2014). Nevertheless, even ardent proponents of the view that DID is a naturally occurring condition that stems largely from childhood trauma (e.g., Ross, 1994) acknowledge that “multiple personality disorder” is a misnomer (Lilienfeld and Lynn, 2015), because individuals with DID do not genuinely harbor two or more fully developed personalities. Moreover, laboratory studies of the memories of individuals with DID demonstrate that the “alter” personalities or personality states of individuals with DID are not insulated by impenetrable amnestic barriers (Merckelbach et al., 2002).
(18) Neural signature. One group of authors, after observing that compliance with social norms was associated with activations in certain brain regions (lateral orbitofrontal cortex and right dorsolateral cortex), referred to the “neural signature” of social norm compliance (Spitzer et al., 2007, p. 185). Others have referred to neural signatures or “brain signatures” of psychiatric disorders, such as anorexia nervosa (Fladung et al., 2009) and autism spectrum disorder (Pelphrey and McPartland, 2012). Nevertheless, identifying a genuine neural signature would necessitate the discovery of a specific pattern of brain responses that possesses nearly perfect sensitivity and specificity for a given condition or other phenotype. At the present time, neuroscientists are not remotely close to pinpointing such a signature for any psychological disorder or trait (Gillihan and Parens, 2011).
(19) No difference between groups. Many researchers, after reporting a group difference that does not attain conventional levels of statistical significance, will go on to state that “there was no difference between groups.” Similarly, many authors will report that a non-significant correlation between two variables means that “there was no association between the variables.” But a failure to reject the null hypothesis does not mean that the null hypothesis, strictly speaking, has been confirmed. Indeed, if an investigator finds a correlation of r = 0.11 in a sample of 20 participants (which is not statistically significant), the best estimate for the true value of the correlation in the population, presuming that the sample has been randomly ascertained, is 0.11, not 0. Authors are instead advised to write “no significant difference between groups” or “no significant correlation between variables.”
(20) Objective personality test. Many authors refer to paper-and-pencil personality instruments that employ a standard (e.g., True–False) item response format, such as the Minnesota Multiphasic Personality Inventory-2 (MMPI-2), as “objective tests” (Proyer and Häusler, 2007), ostensibly to contrast them with more “subjective” measures, such as unstructured interviews or projective techniques (e.g., the Rorschach Inkblot Test). Nevertheless, although the former measures can be scored objectively, that is, with little or no error (but see Allard and Faust, 2000, for evidence of non-trivial error rates in the hand-scoring of the MMPI and other purported “objective” personality tests), they often require considerable subjective judgment on the part of respondents. For example, an item such as “I have many headaches” can be interpreted in numerous ways arising from ambiguity in the meanings of “many” and “headache’ (Meehl, 1945). So-called “objective” personality tests are also often subjective with respect to interpretation (Rogers, 2003). For example, even different computerized MMPI-2 interpretive programs display only moderate levels of inter-rater agreement regarding proposed diagnoses (Pant et al., 2014). Not surprisingly, clinicians routinely disagree in their interpretations of profiles on the MMPI-2 and other “objective” tests (Garb, 1998). We therefore recommend that these measures be called “structured” tests (Kaplan and Saccuzzo, 2012), a term that refers only to their response format and that carries no implication that they are interpreted objectively by either examinee or examiner.
(21) Operational definition. The credo that all psychological investigators must develop “operational definitions” of constructs before conducting studies has become something of a truism in many psychology methods textbooks and other research sources (e.g., Burnette, 2007). Operational definitions are strict definitions of concepts in terms of their measurement operations. As a consequence, they are presumed to be exact and exhaustive definitions of these concepts. Perhaps the best known example in psychology is Boring’s (1923) definition of intelligence as whatever intelligence tests measure.
Many psychologists appear unaware that the notion of operational definitions was roundly rejected by philosophers of science decades ago (Leahey, 1980; Green, 1992; Gravetter and Forzano, 2012). Operational definitions are unrealistic in virtually all domains of psychology, because constructs are not equivalent to their measurement operations (Meehl, 1986). For example, an “operational definition” of aggression as the amount of hot sauce a participant places in an experimental confederate’s drink is not an operational definition at all, because no researcher seriously believes that the amount of hot sauce placed in a drink is a perfect or precise definition of aggression that exhausts all of its potential manifestations. Operational definitions also fell out of favor because they led to logically absurd conclusions. For example, an operational definition of length would imply that length as measured by a wooden ruler cannot be compared with length as measured by a metal ruler, because these rulers are associated with different measurement operations. Hence, the fact that both rulers yield a length for a of say, 27 inches, could not be taken as converging evidence that the is in fact 27 inches long (Green, 1992).
Psychological researchers and teachers should therefore almost always steer clear of the term “operational definition.” The term “operationalization” is superior, as it avoids the implication of an ironclad definition and is largely free of the problematic logical baggage associated with its sister term.
(22) p = 0.000. Even though this statistical expression, used in over 97,000 manuscripts according to Google Scholar, makes regular cameo appearances in our computer printouts, we should assiduously avoid inserting it in our Results sections. This expression implies erroneously that there is a zero probability that the investigators have committed a Type I error, that is, a false rejection of a true null hypothesis (Streiner, 2007). That conclusion is logically absurd, because unless one has examined essentially the entire population, there is always some chance of a Type I error, no matter how meager. Needless to say, the expression “p < 0.000” is even worse, as the probability of committing a Type I error cannot be less than zero. Authors whose computer printouts yield significance levels of p = 0.000 should instead express these levels out to a large number of decimal places, or at least indicate that the probability level is below a given value, such as p < 0.01 or p < 0.001.
(23) Psychiatric control group. This phrase and similar phrases (e.g., “normal control group,” “psychopathological control group”) connote erroneously that (a) groups of ostensibly normal individuals or mixed psychiatric patients who are being compared with (b) groups of individuals with a disorder of interest (e.g., schizophrenia, major depression) are true “control” groups. They are not. They are “comparison groups” and should be referred to accordingly. The phrase “control group” in this context may leave readers with the unwarranted impression that the design of the study is experimental when it is actually quasi-experimental. Just as important, this term may imply that the only difference between the two groups (e.g., a group of patients with anxiety disorder and a group of ostensibly normal individuals) is the presence or absence of the disorder of interest. In fact, these two groups almost surely differ on any number of “nuisance” variables, such as personality traits, co-occurring disorders, and family background, rendering the interpretation of most group differences open to multiple interpretations (Meehl, 1969).
(24) Reliable and valid. If one earned a dollar for every time an author used the sentence “This test is reliable and valid” in a Method section, one would be a rich person indeed, as the phrase “reliable and valid” appears in more than 190,000 manuscripts in Google Scholar. There are at least three problems with this ubiquitous phrase. First, it implies that a psychological test is either valid or not valid. Much like the testing of scientific theories, the construct validation process is never complete, in essence reflecting a “work in progress.” As a consequence, a test cannot be said to be have been conclusively validated or invalidated (Cronbach and Meehl, 1955; Loevinger, 1957; Peter, 1981). Hence, authors should similarly refrain from using the term “validated’ with respect to psychological measures. At best, these measures are “empirically supported” or have “accrued substantial evidence for construct validity.” The same caveat applies to psychological treatments. When Division 12 (Society of Clinical Psychology) of the American Psychological Association put forth its criteria for, and lists of, psychotherapies found to work in controlled trials for specific mental disorders, it initially termed them “empirically validated therapies” (Chambless et al., 1998). Nevertheless, in recognition of the fact that “validation” implies certainty or finality (Garfield, 1996; Chambless and Hollon, 1998), the committee wisely changed the name to “empirically supported therapies,” which is now the term presently in use (Lilienfeld et al., 2013).
Second, the phrase “reliable and valid” implies that reliability and validity are unitary concepts. They are not. There are three major forms of reliability: test–retest, internal consistency, and inter-rater. Contrary to common belief, these forms of reliability often diverge, sometimes markedly (Schmidt and Hunter, 1996). For example, scores derived from the Thematic Apperception Test, a widely used projective technique, frequently display high levels of test–retest reliability but low levels of internal consistency (Entwistle, 1972). There are also multiple forms of validity (e.g., content, criterion-related, incremental), which similarly do not necessarily coincide. For example, a measure may possess high levels of criterion-related validity in multiple samples but little or no incremental validity above and beyond extant information (Garb, 2003).
Third, reliability and validity are conditional on the specific samples examined, and should not be considered inherent properties of a test. Hence, the notion that a test is “reliable and valid” independent of the nature of the sample runs counter to contemporary thinking in psychometrics (American Psychological Association and American Educational Research Association, 2014).
(25) Statistically reliable. This phrase appears in over 62,000 manuscripts according to Google Scholar. It is typically invoked when referring to statistical significance, e.g., “Although small in absolute terms, this difference was statistically reliable, t(157) = 2.86, p = 0.005” (Zurbriggen et al., 2011, p. 453). Nevertheless, despite what many psychologists believe (Tversky and Kahneman, 1971; Krueger, 2001), statistical significance bears at best a modest conceptual and empirical association with a result’s “reliability,” that is, its replicability or consistency over time (Carver, 1978). Indeed, given the low statistical power of most studies in psychology, a reasonable argument could be advanced that most statistically significant results are unlikely to be reliable. The statistical significance of a result should therefore not be confused with its likelihood of replication (Miller, 2009).
