Behavior Research Methods最新文献

Traditionally, behavioral, social, and health science researchers have relied on global/retrospective survey methods administered cross-sectionally (i.e., on a single occasion) or longitudinally (i.e., on several occasions separated by weeks, months, or years). More recently, social and health scientists have added daily life survey methods (also known as intensive longitudinal methods or ambulatory assessment) to their toolkit. These methods (e.g., daily diaries, experience sampling, ecological momentary assessment) involve dense repeated assessments in everyday settings. To facilitate research using daily life survey methods, we present SEMA³ ( http://www.SEMA3.com ), a platform for designing and administering intensive longitudinal daily life surveys via Android and iOS smartphones. SEMA³ fills an important gap by providing researchers with a free, intuitive, and flexible platform with basic and advanced functionality. In this article, we describe SEMA³'s development history and system architecture, provide an overview of how to design a study using SEMA³ and outline its key features, and discuss the platform's limitations and propose directions for future development of SEMA³.

Online experiments are increasingly gaining traction in the behavioral sciences. Despite this, behavioral researchers have largely continued to use keyboards as the primary input devices for such online studies, overlooking the ubiquity of touchscreens in everyday use. This paper presents an open-source touchscreen extension for jsPsych, a JavaScript framework designed for conducting online experiments. We additionally evaluated the touchscreen extension assessing whether typical behavioral findings from two distinct perceptual decision-making tasks - the random-dot kinematogram and the Stroop task - can similarly be observed when administered via touchscreen devices compared to keyboard devices. Our findings indicate similar performance metrics for each paradigm between the touchscreen and keyboard versions of the experiments. Specifically, we observe similar psychometric curves in the random-dot kinematogram across the touchscreen and keyboard versions. Similarly, in the Stroop task, we detect significant task, congruency, and sequential congruency effects in both experiment versions. We conclude that our open-source touchscreen extension serves as a promising tool for data collection in online behavioral experiments on forced-choice tasks.

Interpretation biases in the processing of ambiguous affective information are assumed to play an important role in the onset and maintenance of emotional disorders. Reports of low reliability for experimental measures of cognitive biases have called into question previous findings on the association of these measures with markers of mental health and demonstrated the need to systematically evaluate measurement reliability for measures of cognitive biases. We evaluated reliability and correlations with self-report measures of mental health for interpretation bias scores derived from the Ambiguous Cue Task (ACT), an experimental paradigm for the assessment of approach-avoidance behavior towards ambiguous affective stimuli. For a non-clinical sample, the measurement of an interpretation bias with the ACT showed high internal consistency (r_SB = .91 - .96, N = 354) and acceptable 2-week test-retest correlations (r_Pearson = .61 - .65, n = 109). Correlations between the ACT interpretation bias scores and mental health-related self-report measures of personality and well-being were generally small (r ≤ |.11|) and statistically not significant when correcting for multiple comparisons. These findings suggest that in non-clinical populations, individual differences in the interpretation of ambiguous affective information as assessed with the ACT do not show a clear association with self-report markers of mental health. However, in allowing for a highly reliable measurement of interpretation bias, the ACT provides a valuable tool for studies considering potentially small effect sizes in non-clinical populations by studying bigger samples as well as for work on clinical populations, for which potentially greater effects can be expected.

Monte Carlo simulation studies are among the primary scientific outputs contributed by methodologists, guiding application of various statistical tools in practice. Although methodological researchers routinely extend simulation study findings through follow-up work, few studies are ever replicated. Simulation studies are susceptible to factors that can contribute to replicability failures, however. This paper sought to conduct a meta-scientific study by replicating one highly cited simulation study (Curran et al., Psychological Methods, 1, 16-29, 1996) that investigated the robustness of normal theory maximum likelihood (ML)-based chi-square fit statistics under multivariate nonnormality. We further examined the generalizability of the original study findings across different nonnormal data generation algorithms. Our replication results were generally consistent with original findings, but we discerned several differences. Our generalizability results were more mixed. Only two results observed under the original data generation algorithm held completely across other algorithms examined. One of the most striking findings we observed was that results associated with the independent generator (IG) data generation algorithm vastly differed from other procedures examined and suggested that ML was robust to nonnormality for the particular factor model used in the simulation. Findings point to the reality that extant methodological recommendations may not be universally valid in contexts where multiple data generation algorithms exist for a given data characteristic. We recommend that researchers consider multiple approaches to generating a specific data or model characteristic (when more than one is available) to optimize the generalizability of simulation results.

Multi-informant studies are popular in social and behavioral science. However, their data analyses are challenging because data from different informants carry both shared and unique information and are often incomplete. Using Monte Carlo Simulation, the current study compares three approaches that can be used to analyze incomplete multi-informant data when there is a distinction between reference and nonreference informants. These approaches include a two-method measurement model for planned missing data (2MM-PMD), treating nonreference informants' reports as auxiliary variables with the full-information maximum likelihood method or multiple imputation, and listwise deletion. The result suggests that 2MM-PMD, when correctly specified and data are missing at random, has the best overall performance among the examined approaches regarding point estimates, type I error rates, and statistical power. In addition, it is also more robust to data that are not missing at random.

Response latency is a critical parameter in studying human behavior, representing the time interval between the onset of stimulus and the response. However, different time between devices can introduce errors. Serial port synchronization signal can mitigate this, but limited information is available regarding their accuracy. Optical signals offer another option, but the difference in the positioning of optical signals and visual stimuli can introduce errors, and there have been limited reports of error reduction. This study aims to investigate methods for reducing the time errors. We used the Psychtoolbox to generate visual stimuli and serial port synchronization signals to explore their accuracy. Subsequently, we proposed a calibration formula to minimize the error between optical signals and visual stimuli. The findings are as follows: Firstly, the serial port synchronization signal presenting precedes visual stimulation, with a smaller lead time observed at higher refresh rates. Secondly, the lead time increases as the stimulus position deviates to the right and downwards. In Linux and IOPort(), serial port synchronization signals exhibited greater accuracy. Considering the poor accuracy and the multiple influencing factors associated with serial port synchronization signals, it is recommended to use optical signals to complete time synchronization. The results indicate that under the darkening process, the time error is - 0.23 ~ 0.08 ms (mean). This calibration formula can help measure the response latency accurately. This study provides valuable insights for optimizing experimental design and improving the accuracy of response latency. Although it only involves visual stimuli, the methods and results of this study can still serve as a reference.

Researchers increasingly study short-term dynamic processes that evolve within single individuals using N = 1 studies. The processes of interest are typically captured by fitting a VAR(1) model to the resulting data. A crucial question is how to perform sample-size planning and thus decide on the number of measurement occasions that are needed. The most popular approach is to perform a power analysis, which focuses on detecting the effects of interest. We argue that performing sample-size planning based on out-of-sample predictive accuracy yields additional important information regarding potential overfitting of the model. Predictive accuracy quantifies how well the estimated VAR(1) model will allow predicting unseen data from the same individual. We propose a new simulation-based sample-size planning method called predictive accuracy analysis (PAA), and an associated Shiny app. This approach makes use of a novel predictive accuracy metric that accounts for the multivariate nature of the prediction problem. We showcase how the values of the different VAR(1) model parameters impact power and predictive accuracy-based sample-size recommendations using simulated data sets and real data applications. The range of recommended sample sizes is smaller for predictive accuracy analysis than for power analysis.

Rapid adaptation to sudden changes in the environment is a hallmark of flexible human behaviour. Many computational, neuroimaging, and even clinical investigations studying this cognitive process have relied on a behavioural paradigm known as the predictive-inference task. However, the psychometric quality of this task has never been examined, leaving unanswered whether it is indeed suited to capture behavioural variation on a within- and between-subject level. Using a large-scale test-retest design (T1: N = 330; T2: N = 219), we assessed the internal (internal consistency) and temporal (test-retest reliability) stability of the task's most used measures. We show that the main measures capturing flexible belief and behavioural adaptation yield good internal consistency and overall satisfying test-retest reliability. However, some more complex markers of flexible behaviour show lower psychometric quality. Our findings have implications for the large corpus of previous studies using this task and provide clear guidance as to which measures should and should not be used in future studies.

This manuscript presents a novel geofencing method in behavioral research. Geofencing, built upon geolocation technology, constitutes virtual fences around specific locations. Every time a participant crosses the virtual border around the geofenced area, an event can be triggered on a smartphone, e.g., the participant may be asked to complete a survey. The geofencing method can alleviate the problems of constant location tracking, such as recording sensitive geolocation information and battery drain. In scenarios where locations for geofencing are determined by participants (e.g., home, workplace), no location data need to be transferred to the researcher, so this method can ensure privacy and anonymity. Given the widespread use of smartphones and mobile Internet, geofencing has become a feasible tool in studying human behavior and cognition outside of the laboratory. The method can help advance theoretical and applied psychological science at a new frontier of context-aware research. At the same time, there is a lack of guidance on how and when geofencing can be applied in research. This manuscript aims to fill the gap and ease the adoption of the geofencing method. We describe the current challenges and implementations in geofencing and present three empirical studies in which we evaluated the geofencing method using the Samply application, a tool for mobile experience sampling research. The studies show that sensitivity and precision of geofencing were affected by the type of event, location radius, environment, operating system, and user behavior. Potential implications and recommendations for behavioral research are discussed.

To detect careless and insufficient effort (C/IE) survey responders, researchers can use infrequency items - items that almost no one agrees with (e.g., "When a friend greets me, I generally try to say nothing back") - and frequency items - items that almost everyone agrees with (e.g., "I try to listen when someone I care about is telling me something"). Here, we provide initial validation for two sets of these items: the 14-item Invalid Responding Inventory for Statements (IDRIS) and the 6-item Invalid Responding Inventory for Adjectives (IDRIA). Across six studies (N₁ = 536; N₂ = 701; N₃ = 500; N₄ = 499; N₅ = 629, N₆ = 562), we found consistent evidence that the IDRIS is capable of detecting C/IE responding among statement-based scales (e.g., the HEXACO-PI-R) and the IDRIA is capable of detecting C/IE responding among both adjective-based scales (e.g., the Lex-20) and adjective-derived scales (e.g., the BFI-2). These findings were robust across different analytic approaches (e.g., Pearson correlations; Spearman rank-order correlations), different indices of C/IE responding (e.g., person-total correlations; semantic synonyms; horizontal cursor variability), and different sample types (e.g., US undergraduate students; Nigerian survey panel participants). Taken together, these results provide promising evidence for the utility of the IDRIS and IDRIA in detecting C/IE responding.