1 / 38

7±2 Deadly Sins in Media Violence Research

7±2 Deadly Sins in Media Violence Research. Craig A. Anderson, Distinguished Professor. Digital Media & Developing Minds Congress Cold Spring Harbor Laboratory, Oct. 15-18, 2018. Overview. Much existing controversy results from: Methodological errors Statistical errors Conceptual errors

marotta
Download Presentation

7±2 Deadly Sins in Media Violence Research

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. 7±2 Deadly Sins in Media Violence Research Craig A. Anderson, Distinguished Professor Digital Media & Developing Minds Congress Cold Spring Harbor Laboratory, Oct. 15-18, 2018

  2. Overview • Much existing controversy results from: • Methodological errors • Statistical errors • Conceptual errors • Interpretational errors • Such errors tend to lead to: • Replication “failures” • Underestimates of effects • Grist for the denialist gang • For each Sin, there is a corresponding Virtue • Most MV studies are well-done • Following lists may be useful for A. Researchers B. Reviewers C. Editors D. Research consumers.

  3. Experiments (Randomized Control Design) • Underpowered (sample size) • Inappropriately matched violent & nonviolent stimuli • Lack appropriate cover story • Lack of suspicion check • Weak dependent variables, e.g., • Single item measures • Tangentially relevant DV (relational aggression in VG study) • Inappropriate DVs, e.g., • Trait measures in a short-term experiment • Argument with friends.

  4. Cross-sectional & Longitudinal Studies • Underpowered (sample size) • Overcontrolled • Lack appropriate cover story • Lack of suspicion check • Weak measure of media violence exposure, e.g., • Time on media, not time on media violence • Too few items, too short time period • Weak measure of DV, e.g., • State measure when trait measure is best • Single item • Inappropriate time lags between longitudinal assessments.

  5. Statistical Errors • Mindless formulaic analysis (including preregistration short-sightedness) • Failure to check for outliers, non-normal distributions... • Over-control by use of inappropriate covariates, e.g., • Control for trait aggression when the DV also is trait aggression • Over-control for what conceptually is a mediating variable • Under-control by lack of appropriate covariates.

  6. Conceptual & Interpretation Errors • Misunderstanding definition of media violence, e.g., • Does not require blood, gore, screaming… • Confusing theoretically distinct concepts, e.g., • Aggression, aggressive cognition, aggressive feelings • Misunderstanding role of correlational studies in studying causal questions, e.g., • Testing plausible alternative explanations • Part of triangulation process • Jumping levels of analysis (ecological fallacy) • Using theoretically inappropriate population • Interpreting a lack of correlation as proof of lack of causation.

  7. Final Notes • None of the preceding lists should be construed as comprehensive • Considering these 7±2 deadly sins will help you design better studies, provide better reviews of others’ manuscripts, and understand what the research literature says and doesn’t say • To download pdf version of this talk, • Go to: www.CraigAnderson.org • Click on the link that is directly under the photo • Enter the password: dmadmc

  8. Video Game Violence and Childhood Aggression: The American Psychological Task Force Assessment Sandra L. Calvert Children’s Digital Media Center Georgetown University Paper presented at the 2nd Digital Media and Development Congress. Cold Spring Harbor, NY, October, 2018, NY, NY. This presentation is based on an American Psychological Task Force on Violent Media report chaired by Mark I. Appelbaum and the following subsequent article: Calvert, S.L., Appelbaum, M.I., Dodge, K.A., Graham, S., Hall, G.C.N., Hamby, S.L., Fasig-Caldwell, L., Citkowicz, M., Galloway, D.P. & Hedges, L.V. (2017). The American Psychological Association Task Force Assessment of Violent Videogames: Science in the Service of Public Interest. American Psychologist, 72, 126-143. DOI: 10.1037/a0040413

  9. Background • A task force of experts was convened by the American Psychological Association (APA) to update the knowledge and policy about the impact of violent video game use on potential adverse outcomes. • This APA Task Force on Media Violence examined the existing literature, including the meta-analyses in the field, since the last APA report on media violence in 2005. • Because the most recent meta-analyses were published in 2010 and reflected work through 2009, the task force conducted a search of the published studies from 2009-2013.

  10. Literature Review • To undertake the literature review, the task force adopted a two-pronged approach to identify the literature to be included. • First, we conducted a comprehensive search of PsycINFO, MEDLINE, ScienceDirect, Social Science Research Network, ERIC, Google Scholar, ProQuest, Wilson Social Science Index and miscellaneous journals and references discovered through the search process. • Used the following key words: violent video games, violence, violen*, aggressive behavior, aggression, aggress*, prosocial*, prosocial behavior, computer games, video*, game*, video games and media. • Literature to be considered restricted to research focused on violent video game use separately from other forms of violent media.

  11. Literature Review • Second, based on an initial PsycINFO search, we contacted approximately 130 of the most frequently published researchers in the topic area to request nominations of the 10 strongest empirically based articles addressing violent video game use published between 2000 and 2013. • This process yielded four meta-analyses conducted since the last 2005 APA resolution that were directly relevant to the task force’s charge to evaluate the literature related to violence in video games. 

  12. Prior Recent Meta-analyses of Violent Video Games • The four meta-analytic reviews that the task force identified examined the impact of violent video game use on a variety of negative outcomes • All found reasonably similar statistically significant effects, albeit interpreted quite differently (Anderson et al., 2010; Ferguson 2007a; Ferguson 2007b; Ferguson & Kilburn 2009). • These meta-analyses reviewed more than 150 research reports, including more than 400 effect sizes. • A composite aggression score, aggressive behavior, aggressive cognition, aggressive affect, physiological arousal, reduced prosocial behavior, reduced empathy/desensitization, and criminal violence were included in at least one meta-analysis. • These variables became our focus, as we assessed the new literature after 2009 through August of 2013.

  13. Two Approaches: Systematic Evidentiary Review and Meta-Analysis • Systematic Evidentiary Review (Zief & Agodini, 2012) • Includes a clear set of objectives with pre-defined eligibility criteria used to include studies in the review • Explicit methodology • A systematic search that attempts to identify all studies that meet the criteria • An assessment of the findings in the studies identified • A systematic presentation of characteristics & findings in the included studies • Conclusions based on the evidentiary review

  14. Relevant Literature • Repeated the literature search process to capture all relevant articles that became publicly available between January 1, 2009 (since the last meta analyses by Anderson et al., 2010) and August 12, 2013. • 170 new research reports

  15. 170 Reports Screened using Inclusion Criteria Developed by the Task Force • 1) Does the report include at least one empirical analysis addressing video game violence separately from other media violence? • 2) Does the analysis include complete statistics? • 3) Does the report include at least one outcome variable considered in the earlier meta-analyses: a composite score of aggression, aggressive behavior, aggressive cognitions, aggressive affect, physiological measures, reduced prosocial behavior, reduced empathy/desensitization, delinquency or violence? • 4) Does the report include some measurement of violent video game exposure? • 5) Does the report include some description or assessment to determine that the violent video game is, in fact, violent? • 6) Is the study published in a peer-reviewed academic journal?

  16. Sample Selection • 68 of the 170 articles (n = 78 studies) met all six screening criteria • Each study then rated on a 3 point scale (high, medium, low) for: • Possibility for causal inference • Ecological validity • Sampling validity • Measurement of independent & dependent variables • Created high & low utility groups, Kappa = .789 • Yield was 31 high utility studies

  17. Key Findings from the Evidentiary Review of High Utility Studies (n = 31) • Exposure to violent video games associated with • Increases in aggressive behavior, aggressive thoughts, aggressive feelings • Decreases in prosocial behaviors, empathy & sensitivity to aggression • Findings Robust: Crossed a variety of different kinds of methodologies, including experimental research • Concluded: Exposure to violent video games is a risk factor for aggressive outcomes • Insufficient studies that examine the link between • Exposure to violent video games & delinquency, criminal behavior, or other violent outcomes • An important area in need of additional research

  18. Key Findings from the Evidentiary Review • Even when risk factors such as antisocial personality traits, delinquency, poor academic achievement, parental conflict, child and parent depression, and exposure to deviant peers were controlled: • Relation between violent video game exposure & aggressive outcomes remained robust • Data suggest that the magnitude of relation between exposure to violent video games and aggressive outcomes did not differ: • For adolescents, college students & young adults. • Very few studies on children under age 10.

  19. Evidentiary Review: Future Recommended Directions of Study • Major gaps in the literature about the impact of violent video game exposure as a function of • Gender, ethnicity & socio-economic status. • Although more exposure to violent video games is associated with higher levels of aggressive outcomes • Insufficient studies that examine dose-response relation links with violent video game exposure. • Gaps in the literature • Video game properties (e.g., plots, production features) • Player perspective during video game use (e.g., 1st versus 3rd person perspective, competition or cooperation) • User motivations for using violent video games

  20. Results of Meta-analysis for All Outcomes (n = 31 high utility studies) * denotes meta-analyses with corrections for publication bias, ^denotes aggressive affect analysis, +denotes decreased empathy/desensitization analysis. Note: k = studies, r = correlation metric, d= Cohen’s d; SE = Standard Error

  21. Policy Recommendations • APA engage in public education and awareness activities by disseminating our findings to children, parents, teachers, judges, and other professionals who work with children in schools and communities. • Called for tests of educational interventions directed at children, youth, and their families to determine whether violent video game use can be decreased or the adverse impact of their use can be decreased. • Called for refinement of video game ratings by the Entertainment Software Ratings Board (ESRB).

  22. Conclusions of the APA Task Force on Violent Video Games • The use of violent video games results in increases in overall aggression as well as increases in the individual variables of aggressive behaviors, aggressive cognitions, aggressive affect, desensitization, physiological arousal, and decreases in empathy. • Did not find sufficient studies to evaluate whether there is a link between violent video game use and criminal behavior. • Recommend additional research in a variety of areas, including more studies of younger children, ethnicity, gender, dose-response effects, and game characteristics. • Conclude that the use of violent video game use is a risk factor for subsequent aggression.

  23. Neuroimaging in the study of media violence Tom Hummer IU School of Medicine

  24. What is neuroimaging? • Any one of various tools for determining biological and physical properties of the brain (without cutting it open) From: Weighing brain activity with the balance: Angelo Mosso’s original manuscripts come to light; Brain. 2013;137(2):621-633

  25. MRI: Still lying down • MRI: Physical setup limits naturalistic environment • Less realism, nonsocial, head must remain still • Stimuli typically static images/tasks

  26. EEG • Easier for subjects • Better temporal resolution: can differentiate primary sensory & secondary responses • Direct measure of brain activity • Worse spatial resolution than fMRI • Wearable devices for use outside lab? Engelhardt et al., 2011; J ExpSocPysh47.5: 1033-1036.

  27. What can fMRI tell us? • Traditional fMRI measures relative differences in activity between different conditions • Simple designs allow us to isolate responses or processes, but may be less realistic Example: Neural “desensitization”? • Complicated designs more difficult to interpret (e.g., natural viewing/playing) • Difficult to isolate specific events Gao et al., 2017; Frontiers in psych, 8, 650. Szycik et al. 2017; Frontiers in psych, 8, 174.

  28. fMRI during media violence exposure • Better for viewing general effects, rather than response to specific moments Gentile 2016; PoPMC5(1), 39

  29. How do we interpret fMRI signal changes? • We already know the brain is involved! • Changes in energy use in the brain • “Activation” = More relative activity during one condition • Can be difficult to know what precisely is causing differences in activity • Attention, effort, duration of activity, “baseline” differences • Difficult to parse temporal order of activity: Focus on levels, location • Neural efficiency vs. engagement? • Is more activity better? Or less efficient? • Important to connect to behavior measures, clinical models

  30. Functional brain connectivity • View brain as a dynamic system—not just a bunch of “lightbulbs” • Seed-based: Where is activity correlated with region of interest? • Networks: ICA (data-driven) or defined a priori Yeo et al., 2011; J Neurophsyiol 106: 1125- 1165 • Network analysis typically examines underlying “intrinsic” connections during rest or free-viewing

  31. Brain networks during viewing Networks: Violent > Nonviolent • More holistic approach to brain response to media violence • More work needed that embraces data-driven network approaches Zvyagintsev et al.Neuroscience 320 (2016) 247–258

  32. Naturalistic Viewing • Is brain activity consistent across viewings? Schmalzle 2017, bioXriv

  33. Examining brain structure Freesurfer software • Connects gray matter volume, cortical thickness to measures of long-term media use • Info about long-term associations with brain development • Not simply bigger = better • How does it fit in trajectory?

  34. How does media impact brain development? • Longitudinal studies are needed to understand how media exposure fits into broader picture of development • Structure can’t be viewed in isolation Shaw et al., 2012; Biol Psych 72:191–197

  35. White matter • White matter properties develop until 30’s • May more directly reflect neurodevelopment Diffusion Tensor Imaging: WM microstructure Voxel Based Morphometry: WM microstructure Negative association with TV violence exposure

  36. Connectome Imaging • Global and local network measures • Define connectivity between all regions in the brain to form a whole brain network • Human Connectome Project http://www.humanconnectomeproject.org/ • Brian Connectivity Toolbox (Sporns et al.)

  37. Connectomics: Individual fingerprints? • What is connection with media exposure (short- and long-term)? • May provide more info about individual variability, between-network changes Finn et al., 2015; Nat neuro18(11), 1664.

  38. How best to employ neuroimaging • It is only part of the story! • Combine imaging with other methods, models, measures • Don’t settle for a pretty picture • Understand results within larger understanding of pediatric brain development • Clear, testable neural models are needed • More examination of clinical populations

More Related