Adobe Systems Autumn 2023 Current issues in research of media and audiences Experimental research in cyberaggression and media aggression Marie Jaroň Bedrošová marie.bedrosova@mail.muni.cz What research methods can we use to study cyber/media aggression? surveys interviews focus groups observations content analyses … experiments Photo by Pawel Czerwinski on Unsplash Why do we use experiments? Investigation of the manipulated effect under maximum control Allows inference of causality Photo by Pawel Czerwinski on Unsplash John Stuart Mill (1805-1873) Method of agreement X causes Y Method of difference If X does not occur, Y does not occur Why do we use experiments? Investigation of the manipulated effect under maximum control Allows inference of causality Why do we use experiments? Does watching violent TV make children behave aggressively? John Stuart Mill (1805-1873) Method of agreement X causes Y Method of difference If X does not occur, Y does not occur ? maximum control Terminology Independent variable X manipulated variable factor it has 2 or more values/levels („experimental conditions“) experimental group control group Photo by Anni Roenkae on Pexels Terminology Independent variable X manipulated variable factor it has 2 or more values/levels („experimental conditions“) What is the effect of X on Y? Terminology Dependent variable Y outcome What is the effect of X on Y? Terminology Extraneous variables, „3rd variables“ Z Variables not of interest to the researcher They might influence the studied effects We hold them constant Standardisation of the research situation Terminology Confounding variables uncontrolled extraneous variables that co-vary with the independent variable and could provide an alternative explanation of the results Terminology Confounding variables they change in the same way that an independent variable changes, its effect cannot be distinguished from the effect of the independent variable we do not know if the results are due to the effects of independent variable / confounding variables / combination of those ð uninterpretable Terminology Confounding variables they change in the same way that an independent variable changes, its effect cannot be distinguished from the effect of the independent variable we do not know if the results are due to the effects of independent variable / confounding variables / combination of those ð uninterpretable Obsah obrázku text Popis byl vytvořen automaticky independent variable ice cream consumption dependent variable number of sunburns confounding variable hot temperature How to deal with confounding variables? 1)Standardisation of the experimental situation 2)We hold extraneous variables constant 3)Random assignement Summary Basic characteristics of an experiment 1.We manipulate independent variable(s) – causes 2.We measure dependent variable(s) – outcomes 3.We observe the co-variance of the independent and dependent variables 4.We control for possible confounding variables – alternative explanations of changes in the dependent variable(s) = we reduce the effect of 3rd variables Summary Basic characteristics of an experiment 1.We manipulate independent variable(s) – causes… X 2.We measure dependent variable(s) – outcomes… Y 3.We observe the co-variance of the independent and dependent variables 4.We control for possible confounding variables – alternative explanations of changes in the dependent variable(s) = we reduce the effect of 3rd variables … Z John Stuart Mill (1805-1873) correlation ≠ causality Method of agreement X causes Y Method of difference If X does not occur, Y does not occur •X precedes Y •X and Y are associated •There is no plausible explanation for Y other than X Group task Think of an example of a media-related experiment about (cyber)aggression. What is the independent („manipulated“) variable? What is the dependent variable (oucome)? What are possible confounding variables? How would you deal with them? Obsah obrázku text, tkanina, kreslení Popis byl vytvořen automaticky Advantages of experiments Investigation of the manipulated effect under maximum control ( = reducing the effect of 3rd variables) Allows inference of causality ( = X causes Y) High internal validity = the degree of confidence that the causal relationship we are testing is not influenced by other factors or variables when we have a lot of control Photo by Pawel Czerwinski on Unsplash Obsah obrázku text, tkanina, kreslení Popis byl vytvořen automaticky Disadvantages of experiments Low external validity – low ecological validity „Lab environment“ Can we generalize to everyday life? To „normal“ media consumption? Balance of internal / external validity Obsah obrázku text, tkanina, kreslení Popis byl vytvořen automaticky Disadvantages of experiments Low external validity – low ecological validity „Lab environment“ Can we generalize to everyday life? To „normal“ media consumption? Balance of internal / external validity more control … less „natural“ Obsah obrázku text, tkanina, kreslení Popis byl vytvořen automaticky Disadvantages of experiments Low external validity – low ecological validity „Lab environment“ Can we generalize to everyday life? To „normal“ media consumption? Balance of internal / external validity more „natural“ … less control Obsah obrázku text Popis byl vytvořen automaticky Types of experiments Photo by Dids on Pexels Lab experiment highly controlled conditions „artifical“ situation, low ecological validity Field experiment everyday environment, reflecting real life, high ecological validity low control, many possible counfounding variables Natural experiment independent variables occurs naturally in real life (e.g., policy changes, weather events, natural disasters, …) researcher has no control over the independent variable(s) Obsah obrázku text Popis byl vytvořen automaticky Types of experiments Photo by Dids on Pexels Between subject design we compare groups of people different people in each experimental condition Example Does playing aggressive video games with personalised avatars cause aggressive behaviour? Experimental stimulus A – group 1 1.playing aggressive video game with a non-personalised avatar 2.measuring aggressive behaviour Experimental stimulus B – group 2 1.playing aggressive video game with a personalised avatar 2.measuring aggressive behaviour Comparing the measured aggressive behaviour between the two groups Uživatel se souvislou výplní Obsah obrázku text Popis byl vytvořen automaticky Obsah obrázku text Popis byl vytvořen automaticky Types of experiments Photo by Dids on Pexels Between subject design we compare groups of people different people in each experimental condition Within subject design we compare the same people in different experimental conditions the same person is in all experimental conditions repeated measurements Example Does playing aggressive video games with personalised avatars cause aggressive behaviour? Experimental stimulus A – same group 1.playing aggressive video game with a non-personalised avatar 2.measuring aggressive behaviour Experimental stimulus B – same group 1.playing aggressive video game with a personalised avatar 2.repeated measurement of aggressive behaviour Comparing the measured aggressive behaviour within the same people Uživatel se souvislou výplní Obsah obrázku text Popis byl vytvořen automaticky Obsah obrázku text Popis byl vytvořen automaticky Types of experiments Between subject design we compare groups of people different people in each experimental condition Within subject design we compare the same people in different experimental conditions the same person is in all experimental conditions repeated measurements Can you think of any problems connected to the second design? different designs are suitable for different research questions Example Does playing aggressive video games with personalised avatars cause aggressive behaviour? Experimental stimulus A – same group 1.playing aggressive video game with a non-personalised avatar 2.measuring aggressive behaviour Experimental stimulus B – same group 1.playing aggressive video game with a personalised avatar 2.repeated measurement of aggressive behaviour Comparing the measured aggressive behaviour within the same people Uživatel se souvislou výplní Obsah obrázku staré, zavřít, špinavé Popis byl vytvořen automaticky Ethics Photo by Dids on Pexels Same rules as in other research designs We need to minimise risks and potential harm for our participants E.g., Do we need to show children realistic images of aggression and violence? How do we measure aggression? What is ethical? We need to obtain informed consent from our participants Obsah obrázku staré, zavřít, špinavé Popis byl vytvořen automaticky Ethics Same rules as in other research designs We need to minimise risks and potential harm for our participants E.g., Do we need to show children realistic images of aggression and violence? How do we measure aggression? What is ethical? We need to obtain informed consent from our participants What is an informed consent? A form where we introduce the research We explain all potential risks and harms from participation But – can we reveal the research question? It might reveal also the nature of the experiment and influence our participants Debriefing Obsah obrázku text, savci, kočka domácí Popis byl vytvořen automaticky Do you have any questions? Experimental design Investigation of the manipulated effect under maximum control Allows inference of causality X precedes Y X and Y are associated There is no plausible explanation for Y other than X Obsah obrázku text, savci, kočka domácí Popis byl vytvořen automaticky Examples of experiments researching media aggression and cyberaggression Obsah obrázku text, klipart Popis byl vytvořen automaticky Bobo doll experiments – social learning Albert Bandura (1925 – 2021) Social learning theory / social cognitive theory When observing other people, we are more likely to imitate their behaviour if it has been rewarded and if we perceive those people as important to us (parents, siblings, role models, celebrities) Albert Bandura – bobo doll experiments (Bandura et al., 1963) Several experiments – real-life / on TV observations Children observed adults being aggressive toward the doll ð imitated adults‘ behaviour (stonger effects for boys) Effect of reward and punishment 3 experimental conditions Aggressor was rewarded Aggressor was punished No reward/punishment Obsah obrázku různé, stejné Popis byl vytvořen automaticky Bobo doll experiments – social learning Desensitisation theory Long-term influence of media Gradual habituation to repetitive violent content – e.g., over time, we do not perceive it as emotionally strong Media content creators increase the quality and quantity of violence to gain attention Evolution of the Zombie Desensitization and hate speech What is cyberhate (online hate speech)? Hateful and bias-based expressions via ICT Attacking group characteristics or group membership Motivated by an intergroup bias (connected to stereotypes and prejudice) Exposure to hate speech increases prejudice through desensitization Soral et al., 2018 N = 75 students, between-subject design Laboratory experiment Study about the relationship between web design and memory processes – reading 5 pages from discussion fora and assessing esthetics of the page design Control group + experimental group Neutral comments Hateful comments + Study about reception of Internet content Sensitivity to hate speech Outgroup prejudice Exposure to hate speech increases prejudice through desensitization Results Even relatively short exposure to hate speech desensitized participants to its offensiveness Exposure to hate speech increased the level of prejudice (mediated by desensitization) Comfortably numb: Desensitizing effects of violent media on helping others Bushman & Anderson, 2009 Study 1, N = 320 students, between-subject design Lab experiment 2 conditions Survey about videogames Overhearing staged fight … how long does it take to help the victim? If not during 3 minutes, did the participant admitted hearing the fight? How seriously did they rate it? Violent videogame (Carmageddon, Duke Nukem, Mortal Kombat, Future Cop) Nonviolent videogame (Glider Pro, 3D Pinball, Austin Powers, Tetra Madness) Comfortably numb: Desensitizing effects of violent media on helping others Results Participants playing violent game •Took significantly longer to help •Were less likely to notice the fight •Rated the fight as less serious Comfortably numb: Desensitizing effects of violent media on helping others Bushman & Anderson, 2009 Study 2, N = 162 adult moviegoers, between-subject design Field experiment 2 conditions Staged emergency: young woman who dropped her crutches + control (emergency before violent/nonviolent movie) … how long does it take to help the woman? Violent movie (The Ruins) Nonviolent movie (Nim‘s Island) Comfortably numb: Desensitizing effects of violent media on helping others Results People who saw violent movie •Took longer to help • • Limitations? Violent video games and hostile expectations: A test of the General Aggression Model What was the study about? Bushman & Anderson, 2012 N = 224 students, between-subject design 2 conditions Three ambiguous stories … What happens next? What would the character do/say, think, or feel? Results: violent videogame – increase in hostile expectations Violent videogame (Carmageddon, Duke Nukem, Mortal Kombat, Future Cop) Nonviolent videogame (Glider Pro, 3D Pinball, Austin Powers, Tetra Madness) BUT – do media cause violence? Experimental studies Low ecological validity Short-term effects Small effect sizes Possible effects of many confound variables Effect of media contents Rewarding violence and aggression Not/realistic violence Not/showing the consequences for victims Positive portrayal of the aggressor Avatar personalisation Effect of individual characteristics Individuals‘ aggressivity, hostile attribution bias, normative belifs about violence, empathy, moral identity, … Effect of social environments Family conflicts, parental mediations, parental media consumption, … Aggression and Antisocial Behavior - ppt download BUT – do media cause violence? Vulnerable population The influence of the media on violent behaviour is less than the influence of socio-demographic characteristic At the same time, these characteristics predict the preference and frequency of consumption of violent content Other examples of experiments: Bystanders of Instagram aggression and their moral disengagement Cyberbystanders – witnesses of cyberaggression Moral disengagement – selective deactivation of the self-regulatory system and self-sanctions for immoral behaviour (Bandura 1999; 2002) Victim blaming – rationalising the aggression as being provoked by the victim or as justified due to the victim’s behavior or characteristics Minimizing consequences – reframing of the harmful effects that aggression can have on its victims by ignoring them or minimising them Bystanders of Instagram aggression and their moral disengagement Online experiment, between-subject design N = 658 Czech adolescents 2 conditions Negative comments about the girl How do bystanders (our participants) evaluate the incident? What is their moral disengagement? What is the role of „anti-fat attitudes“, exposure to body-positive online content, and gender? IG post from a girl who is thinner + negative comments IG post from a girl who is plus-size + negative comments Bystanders of Instagram aggression and their moral disengagement Obsah obrázku text, voda, snímek obrazovky Popis byl vytvořen automaticky Obsah obrázku text, voda, snímek obrazovky Popis byl vytvořen automaticky Bystanders of Instagram aggression and their moral disengagement Results Experimental condition: Girl who is plus-size / who is thinner Victim blaming Minimizing consequences Anti-fat attitudes Exposure to body-positive online content Gender Bystanders of Instagram aggression and their moral disengagement Results Experimental condition: Girl who is plus-size / who is thinner Victim blaming Minimizing consequences Anti-fat attitudes Exposure to body-positive online content Gender Bystanders of Instagram aggression and their moral disengagement Results Experimental condition: Girl who is plus-size / who is thinner Victim blaming Minimizing consequences Anti-fat attitudes Exposure to body-positive online content Gender (boys more) Bystanders of Instagram aggression and their moral disengagement Results Victim blaming Bystanders of Instagram aggression and their moral disengagement Obsah obrázku text, voda, snímek obrazovky Popis byl vytvořen automaticky Obsah obrázku text, voda, snímek obrazovky Popis byl vytvořen automaticky Limitations? Other examples of experiments: Cyberostracism: Emotional and behavioral consequences in social media interactions Galbava et al., 2021 (Cyber)ostracism – social exclusion, being ignored or excluded (online) H1a. Ostracized people will derive lower satisfaction from belonging, self-esteem, control, and meaningful existence than non-ostracized people. H1b. Ostracized people will be in a worse mood than non-ostracized people. H2. Ostracized individuals with higher social anxiety will experience lower needs satisfaction. H3a. Ostracized people who are less satisfied in the needs of belonging or self-esteem will tend to choose a prosocial response. H3b. Ostracized people who are less satisfied in the needs of meaningful existence or control will tend to choose an antisocial response. H3c. Ostracized individuals with higher social anxiety will tend to choose an evasive response. Cyberostracism: Emotional and behavioral consequences in social media interactions N = 246, CZ Laboratory experiment, between-subject design Survey … „group“ task 1 … survey … „group“ task 2 (interactions with „real people“ - preprogrammed) Task 1 – create profile, meet other „participants“ and see their profiles, give likes (and receive likes from „others“) 2 conditions participants received 1 like (ostracism) participants received 4 likes (control) Cyberostracism: Emotional and behavioral consequences in social media interactions Survey … „group“ task 1 … survey … „group“ task 2 (interactions with „real people“ - preprogrammed) Task 2 – cooperative financial game. Goal was to maximize the groups‘ profit, minimum amount of money to play was 500 CZK, pariticipant received 800 CZK but one other „player“ received only 200 CZK. The „player“ asked participant for a loan. Possible reactions: Results: Most common pro-social reaction (opportunity of re-inclusion) But – people in an ostracized condition choose an antisocial response more often than people in a control condition Loan money (prosocial) Refuse to loan money (antisocial) Choose to not to play the game (evasive) Hi, I need to borrow 300 CZK for the entry fee. Will you loan it to me? Hi, I need to borrow 300 CZK for the entry fee. Will you loan it to me? Obsah obrázku text, savci, kočka domácí Popis byl vytvořen automaticky Experimental research in cyberaggression and media aggression https://padlet.com/cyber_marie/lfln8i99o374qs4 What two main things I learned today at the lecture? Do I have any remaining questions? Adobe Systems References ̶Bandura, A. (1999). Moral disengagement in the perpetration of inhumanities. Personality and Social Psychology Review, 3(3), 193–209. https://doi.org/10.1207/s15327957pspr0303_3 ̶Bandura, A. (2002). Selective moral disengagement in the exercise of moral agency. Journal of Moral Education, 31(2), 101-119. https://doi.org/10.1080/0305724022014322 ̶Bandura, A., Ross, D., & Ross, S. A. (1963). Imitation of film-mediated aggressive models. Journal of Abnormal and Social Psychology, 66(1), 3–11. https://doi.org/10.1037/h0048687 ̶Bedrosova, M., Kvardova, N., & Machackova, H. (accepted). Bystanders’ victim blaming and minimizing consequences of weight-based cyberhate attacks: The roles of anti-fat attitudes, body-positive online content, and gender. Journal of Interpersonal Violence. ̶Bushman, B. J., & Anderson, C. A. (2009). Comfortably numb: Desensitizing effects of violent media on helping others. Psychological Science, 3, 273–277. https://doi.org/10.1111/j.1467-9280.2009.02287.x ̶Bushman, B. J., & Anderson, C. A. (2012). Violent video games and hostile expectations: A test of the General Aggression Model. Personality and Social Psychology Bulletin, 28(12), 1679–1686. https://doi.org/10.1177/014616702237649 ̶Galbava, S., Machackova, H., & Dedkova, L. (2021). Cyberostracism: Emotional and behavioral consequences in social media interactions. Comunicar, 29(67), 1–11. https://doi.org/11.10.3916/C67-2021-0 ̶Goodwin, Kerri A., & Goodwin, C. James. (2016). Research in psychology: Methods and design (8th edition). Wiley. ̶Soral, W., Bilewicz, M., & Winiewski, M. (2018). Exposure to hate speech increases prejudice through desensitization. Aggressive Behavior, 44, 136–146. https://doi.org/10.1002/ab.21737 Adobe Systems