Comm 113 Midterm
Lazarsfeld study undermined "powerful effects" model by emphasizing 2 things (these are the two basic premises of the limited effects paradigm)
(a) selectivity - Notion of selectivity calls into question significantly the idea that the nature of the causal relationship is that media affects behavior; it's just as likely that your predisposition is what causes you to select different types of media (selectivity) - Some things that might look like powerful effects are evidence of selectivity at work - This runs directly counter to the basic premise of mass society theory o The mass part of mass society theory pretty much suggests that media consumers are relatively passive consumers of whatever information is thrust upon them, that we are not selective. o Limited effects paradigm instead posit that we're not passive; we're actually active (b) interpersonal connections (and "cross influences") - Study refuted this mass society notion that we live in a type of social vacuum, disconnected from other ppl - The study notes that ppl live in rather dense social networks and we have friends who have different opinions to us and diff experiences. These are called cross influences o Yes, the media might influence us in one sense or in one direction, but we're always able to balance that influence against other things, opinions, and other ideas that we get from our diverse network of interpersonal connections o We aren't these socially isolated individuals who are totally at the mercy of media content in terms of formulating political opinions; we do have a body of other interpersonal connections that we can add to inform our social and political attitudinal identity - This also runs counter to another basic premise of mass society theory which is this notion that we as a society are atomized and disconnected from each other and from important socializing institutions o Limited effects model says this isn't true, we are pretty well connected to other sources of influence besides the mass media
Payne Fund studies correlational findings
- 8 year olds retained a lot of info for at least 3 months; tend to accept things that they see as true; most attentive to action o Looked at survey data and found that children who go to the movies retain a lot of the info they see for at least a fairly long time o Younger kids (8 yr olds) tend to accept things they see as true; they're not great at distinguishing fact from fiction o They're more attentive to action, which they are still (this is the concern about the marketability of media violence) o They retain info they learn from films - Movie watchers scored lower on many ratings like academics, cooperation, morality, and stability o The main impulse for a lot of this concern was that it might be that film viewership among adolescents might be associated with truancy or delinquency and lower academic performance and classroom citizenship o According to these correlational studies, they did find that the more movies adolescents watch, the higher they score on that variable, the lower they scored on academics, the lower they scored on measures of cooperation, morality, and stability - The main finding from the Payne studies is that there was this thought of there that somehow films are a contributor to juvenile delinquency and this research suggests that this may be the case o This is merely correlational; not saying that kids who watch movies do worse in school. It cannot say that conclusively, just saying these two variables go together. o these findings were correlational/observational in nature; can't say that A caused B, just that they correlate - Kids are emotionally affected (such as frightened) by films; such effects may be long lasting, and may be unrecognized by parents o Some level of childhood traumatization going on o Parents seem to be unaware of how their children are being affected by media content
Types of mass media effects or theories: direct vs conditional
- A direct effect would be something where you say you can observe a broad statement, which is like media violence makes ppl aggressive o Ex 2: here is a drug that lowers ppl's BP - A conditional effect would be where you say media violence makes people more aggressive if they are male, but not female, and if they live in an environment that already tolerates some levels of aggression as opposed to them living in an environment which tends to not tolerate aggressive norms o Whatever that media effect is, it's gonna be conditional on who that person is and what other circumstances are o Ex 2: here's a drug that lowers ppls BP, but only in a certain condition, so only if they already have high BP. it doesn't necessarily lower your BP if you don't have high BP, but maybe it lowers my BP bc i'm in that condition that has high BP o Media effects tend to be fairly conditional. It's not as though everyone is affected exactly the same way, it's going to partly depend on what other conditions exist.
Types of mass media effects or theories: Micro vs. macro
- A micro focus will be where you're looking at what happens to individuals o how does the media fit? How does the media, either in terms of content or technologies, how does it affect us as individuals and inside our brain? o this class is the more individual, micro psychological version of media effects, but we dabble in macro from time to time - macro effect would be looking at how does the mass media as an institution function within society generally? o You don't necessarily have subjects coming into a lab to fill out questionnaires, but you're discussing things on more sociological terms. o Talking on an institutional level; not necessarily talking about exactly what happens in ppls brains
results of Payne study experiment
- Anti-black "birth of nation" significantly reduced "favorability" rating regarding blacks o Negative outlook got worse towards African Americans - Pro-chinese "son of the gods" significantly reduced "unfavorability" rating regarding Chinese immigrants o Attitudes got better towards Chinese immigrants - These findings map on this idea that there might be some attitudinal effects due to the media exposure to certain types of content - Results persisted for months (or year+) and sometimes grew stronger over time (sleeper effect?) o Suggests they're not transient. They sometimes grew stronger over time, which is the sleeper effect.
Concern about powerful media effects driven partly by WW1 propaganda fears, general fear/misunderstanding of new technology
- During this era, many nations including our own were successful in propaganda and used the media to demoralize citizens of enemy nations, and opposing armies, so there is this sense that we the government could use the media to brainwash us or brainwash someone - So there's a little bit of paranoia concerning the possible effects of mass media as it was starting to develop
The scientific method is an epistemological approach to truth that values empiricism
- Epistemological means it's a way of knowing and learning about the world we live in - Empiricism means its based on careful observation of facts generated by various research methodologies - Rules of the scientific method are designed to guard against natural human biases in observation/interpretation o One of those biases would be selective attention; We have built into the scientific method a system for guarding against those selective processes - " like any game, science has rules. Playing outside those rules is cheating.. Whether and how you punish the cheaters is up to you.." (Steven Milloy, Cato Institute) o In all scientific domains, theres going to be different sides. Often its literally a political or ideological drive where some ppl one one side think one thing and they can generate research and promote research that supports what they think, and the same can happen on the contrasting side o CATO institute is a conservative/liberatarian think tank, and their philosophy is that a good ton of government policy and regulation is ill advised and is premised on bad science + If you want to undermine certain govt regulations and policies, one way you can do that is by undermining the science on which they rely + Lots of examples where junk science has led to bad public policy + Many ppl cheat in the scientific method even though in abstract its a great idea and national practice
Using control variables is a VERY GOOD practice, but can also facilitate fishing for the results you want
- Ex: researcher hypothesizes that more TV viewing is associated with lower satisfaction of one's body image - You're aiming to get p under .05, so you keep controlling for different variables... you may eventually find the magic one [age and self-esteem finally nudged under .05] o Your conclusion: "Controlling for subject age and reported level of self esteem, we found that number of hours of TV viewing per week was significantly related to body image attitudes" + Statement isn't wrong, but it doesn't include the fact that you spent all this time to make it significant ^^^ it must all be disclosed/reported with transparency... but oftentimes ppl just take away that main conclusion
Stats can be very useful in controlling (covarying) "third variables"
- Ex: study between correlation between violent video games and aggression [you identify a positive correlation] o Someone could point out a third variable like gender [if you're a boy, you're more likely to play video games and also be naturally aggressive] o So redo the initial correlation (between just two variables), control for the third variable (gender) + On this redo... If controlling for C kills the original correlation between A and B, that means the initial correlation was "spurious" [merely due to the third control variable] + But if you redo it and the original two variables are still significantly correlated AFTER CONTROLLING FOR C, then we are more sure it is a meaningful relationship - Therefore, you can improve the interpretive value of correlations by controlling for other types of explanations/third vars ( 2nd example pg 35)
Significance is heavily dependent on sample size
- Ex: women have barely higher GPA than men (3.12 vs. 3.10)... such a slight difference easily occurs due to chance o This difference is likely not significant with n = 100 participants o But if n = 4000 participants... then that tiny difference is likely significant [*with a huge n value, almost anything comes out as statistically significant] - Ex: if the GPA greatly differs between the genders (2.41 v. 3.29), but the sample size is only 20 people... this big GPA difference still isn't significant
Film industry criticized for pandering to lower class tastes, contributing to societal decline, undermining social order
- Film industry was criticized for targeting and exploiting lower class tastes - Idea here is that film could have the eventual consequence of driving down the cultural curve - Socio Intellectual cultural hierarchy: 3 different hierarchical classes at time (Upper, middle, lower [middle class didn't exist before industrial revolution, was a consequence of IDR] , and they all have different tastes and therefore like different types of content o At the top, they might like opera o In the middle and below they might like things that are less sophisticated, things more like sex and violence - If i'm in the film industry, it doesn't make much sense for me to target the upper class bc they're not a big market in any industry. There's not too many ppl who live up there in the upper class, so they're not gonna provide me with a great cash flow - There are more ppl in the middle class than the upper class, but then most ppl are at the bottom of the sociocultural pyramid, so I'm gonna create content and themes in my films to appeal to what these lower class ppl like. The idea is that overtime, the film industry might drive the cultural curve down by promoting those lower cultural themes - Barriers between each class are rigid and obvious; it's hard for someone to migrate across barriers. If I'm up there at the top, I like that there are barriers that preserve my identity for being in that class. The idea here is that film created a mingling of these classes; ppl from all social classes go to the same movie theaters, created a blurring of class distinctions that was once a very impermeable barrier, now it becomes somewhat blurry and more permeable, this caused concern for ppl higher on sociocultural spectrum
Stats also used to state effect size
- If correlation of r=.25, r^2=.0625, which means shared variance is about 6% - Coefficient (r) determination: what portion of the variables is overlapping? o How much variance in one variable is shared with variance in another variable? - Ex: 6% of the variance in aggression is explained by/accounted for by/shared with variance in TV violence exposure - Overlapping venn diagram circles = correlation/relationship... NOT causation
Ex: someone says it's very possible that kids who watch a lot of tv are fat. Is that due to the message effects or media effects?
- If we are characterizing it as a message effect, we might say bc they watch a lot of tv, they watch a lot of junk food ads, and this content causes them to eat more junk food (content effect) - A medium effect would be saying, well i don't care what messages they're watching, they might be watching sports all day, but just the fact that they're sitting in front of that technology means they're not exercising for that many hours per day. Not the content making them fat, but the very existence of the medium or how they're using that medium (the technology) that is responsible for whatever effects we think we're observing
Lazarsfeld study: Democrats used "alternative" media of radio, Republicans used newspapers
- Important conclusion from this study - Radio was the new tech at this time, newspapers were the traditional technology - Younger ppl more likely to be democrats, older ppl more likely to be republicans o Younger ppl more open to alternative new media, older ppl more reliant on traditional old media - Not a case of media brainwashing here; its basically the case that younger audiences were already likely to be dems and they were already likely to be more radio reliant. Older ppl already more likely to be repubs and already more likely to rely more heavily on newspapers - What you're seeing here (important point) is a selectivity effect o Younger ppl select a type of alternative media, older ppl select a more traditional form of media + There are political consequences associated with that o Not as though radio brainwashes ppl to be dems and newspaper brainwashes ppl to be repubs; theres a selectivity here + Ex: fox news doesn't brainwash ppl to become republicans; they already were conservatives and that's why they choose fox news (selectivity effect)
Hypodermic/"Magic Bullet" model of the early 1920s
- In the 20s, you have film developing and some people were concerned about the effects that the film industry might have on audiences and particularly youth audiences - In early 20s, there was the hypodermic/magic bullet model o Hypodermic refers to a hypodermic needle/syringe. If you get this drug in a needle form from a doctor, you can't avoid getting the drug that's gonna hit you. That drug is also probably gonna affect everyone the same way. You can't evade the effects of that drug once you've been exposed to it o Similarly, the implication of the magic bullet is that if someone shoots a bullet at a bunch of ppl one at a some and it hits them all in the stomach, chances are no one is going to evade that bullet; they're exposed to that bullet and they're going to be hit by the bullet, and the bullet is going to possibly have powerful effects, and furthermore people are going to be similarly impacted. Some ppl might survive, others might die, but there's not going to be anyone who says "i got shot in the stomach, didn't bother me at all and i barely feel it" -
Neumann and Guggenheim (1.) article on GS suggest minimal vs powerful effects narrative has impeded the Communication field's development
- Instead they note that the 'health' of a scientific field has less to do with disputes over effects size, and more to do with theoretical accumulation and sophistication o Basically they say in the article that like any other healthy, mature science, what happens is the field will focus their attention for a period of time on some things that the field thinks are relevant and timely and unresolved, until they think "Ok, we more or less have a handle on this range of theories and topics and constructs, let's now move on to something else" - In this article they look at publications that happened over the last 80 years and what kinds of studies were published, because that's an indication of what the field was looking at, and they have 6 stages where they discuss the kinds of theories and research domains that were happening at this time
Payne Fund studies (1929-1933): Ambitious (through flawed) application of quantitative investigation of media effects
- It was a good try and a decent starting point, and it represents an attempt to study something social scientific (like attitudes and human behaviors) utilizing similar tools that the hard sciences would use - It was a bunch of studies that were published during this period - Took two forms o One bundle of studies that were correlational in nature + In correlational study, you look to see how two or more variables relate to one another. Just looking at how variables go together o another study that was experimental
in the early 1920s, you have this notion here of a "Mass Society" perspective
- Mass implies that the populace, a body of consumers, an amorphous glob of anonymous humanity who are not very active or selective in their consumption behaviors or exposure to patterns that they hear. Seen as very passive recipients of what the environment might offer - If we talk about ppl getting shot with the bullet and everyone pretty much responding in the same way, it sounds like you're talking about a mass - Nowadays people don't experience the media the same way or respond to media content in the same way; we're a lot more idiosyncratic than a mass society perspective would suggest - Mass society perspective characterized early part of the 20th century; tended to regard society as atomized (individuals disconnected from socialization institutions and others) and thus vulnerable to media effects o Since they were so disconnected, the idea is that they're wandering around lost in the social landscape, unaware of what the appropriate social norms are and how they should behave and who they are o This notion began during the industrial revolution. After the industrial revolution, the city of london is the economic, political, social, and cultural headquarters of that nation. Power and assets consolidated in that city bc this is where the factories are. So when there's industry, initially its going to become concentrated in some of these urban environments like london as this is where the work is. So ppl leave from the countrysides and elsewhere in the country and move to london (city). + Lets say that before the industrial revolution, i and my family have always lived in the same country side of northern england, so theres a continuity there of living in my community my whole life, and consequently it provides me with a lot of social support and identity. Will prob have my same job my whole life and will stay within the same socioeconomic strata. So all of these socializing institutions (family, community, occupation) provide me with a great sense of who i am + Once you have factories, work, commerce, and career opportunities down there in london and i move out of my village to london, now im disconnecting myself from these important socializing institutions - So the mass society theory was a sociological notion a little concerned about having ppl who are vaguely disconnected from important socializing institutions at an exact point in time when now you have other socializing influences like film, radio, and maybe one day TV
By the 60s, the Powerful effects model returns because:
- Mass media became scapegoat for social problems (like violence) around 1960's o One thing that was driving this impulse is that by the late 50s-early mid 60s, there's a ton of social change/disruption, some good and some bad (civil rights movement, counterculture revolution which was good and bad [drug parts were bad]). o Some scholars who had been more sociologically minded, wanted to blame for some of this (what they considered to be) destructive, dysfunctional change, and looked to the usual suspects, such as the institution of the mass media o Although Lazarfeld found that media is not that powerful to begin with, these scholars believed that mass media had important influence on society and on individuals, and thought that maybe Lazerfeld didn't identify these powers correctly -Energizing this impulse to look for these powerful effects were things like improved stats and research sophistication that developed throughout the 50s-60s that helped 'discover' subtler effects o used 'path analysis', which is where you basically take correlations but dress them up with fancy statistical dress, and out of those correlations you say, although all we have is correlational data, when I process all that data as a path analysis, i can make a causal conclusion. Correlation CANNOT be used to identify a causal relationship, and yet thats exactly what path analysis does and it may cause scholars to see things that don't actually exist - More focus on pervasive, socialization issues (things that happen over one's developmental lifespan) o By the 60s, TV became a major component of ppl's lives; 9% owned TV in 1950, 65% by 1955, 93% of 1965 + Messages and ideas constantly seeping out of the box in ppl's living rooms and influencing your developmental history, like an additional parent + This is the idea that you are now becoming socialized not because of one exposure, but because of a bunch of diff kinds of exposures across your developmental lifespan, which can make these effects harder to detect. Tv by this time was all around, everyone was exposed to TV constantly
"War of the Worlds" widespread panic is a very good examples of a societal lack of "media literacy"
- Media literacy is an individual level construct o Means that i as an individual, if im media literate, am capable of reading the media, deconstructing it, understanding it. I speak the language and im not easily deluded and i use it constructively o How good an individual does or does not do with respect to judiciously consuming media content - Here we can bring this concept up to a societal level o At this point in time society wasn't universally great at understanding the nature of the particular new medium o It takes a while before we as a culture and a society and as individuals really learn how to judiciously and constructively use a new media tech o Radio was only 20 yrs old at time, we hadn't gotten used to developing a societal skill set for it
Types of mass media effects or theories: Message effects vs medium effects
- Message effects pertain to the content/the nature of the message o Ex: if we are talking about the effects of violent tv, well violent content is what we're talking about - A medium effect does not have that much to do with or indeed anything to do with the content, but just the very existence of the media technology o What effect does the very existence of that media technology have and usage of that technology apart from the content you're absorbing via that technology?
Types of mass media effects or theories: Nomothetic vs. idiographic
- Nomothetic theory is one that is trying to cover a very broad range of experience. Looking for a general theory of media effect o As if you were to say "a certain type of political communication affects everybody the same way" o In the olden days (1970s-80s) you could be looking at effects of certain type of TV content, and there was an impulse towards nomothetic theories bc back then, everyone was watching the exact same kind of TV content bc there was only like 3 TV channels and they were all the same. Easier to come up with theories about how that content that everyone consumed affected everyone pretty similarly - Idiographic theory is the idea that whatever the effect of this media content or technology is, it's gonna be heavily dependent on who that person is. Not gonna be some on size fits all kinda theory that affects everybody equally o Today we live in a much more fragmented mediascape and we have heavily drifted in the idiographic direction in the last 20 years . Everyone lives in their own media silo, so we don't have the same assumption as we did in the 1970s of everyone doing and watching the same things and absorbing the same kind of content and having the same kind of effects o We have to bear in mind that not everyone's media experience is likely to be similar to somebody else's and they're likely to be very differential types of effects depending on who they are aka on their idiosyncratic identities
Content Analysis
- Observations regarding frequency, context (surrounding those events), and nature of content o No participants/study subjects - Doesn't say anything about effects
results of Lazarsfeld's study
- One thing he noted is that voting behavior is almost entirely explained by SES, occupational class, religious affiliation, rural vs urban, and age o If you know these things about someone, theres a very likely chance you can tell who they voted for in an election. This means its almost predetermined, and that means that there's not much variance left to explain o Critically important result here: basically whatever we say about media influence, overwhelmingly, its demographic things that determine the voting behavior - Ppl who are less interested in the political process were less committed to their political choices and they were late deciders
The People's Choice: How does the media influence voting behavior?
- Panel study conducted by Lazarsfeld in Erie County, Ohio, during the 1940 presidential election o Panel study means that he observed subjects at multiple data points across most of 1940 so he could track their changes in their attitudes and behaviors as they existed across that time frame; pretty good methodological paradigm - This is a carefully considered, carefully thought out and applied methodological approach to that very question: how do people incorporate media content in terms of forming political opinions? Asking "how does the media influence voting behavior? - Rigorous application of scientific methods to social questions o Rigorous application of scientific methods to social questions o Started study around January of 1940 and returned to these ss periodically to track over time their media usage and the development of their political attitudes and voting behaviors in that timeframe between Jan 1940 and when they eventually voted in November of 1940 o This is the study where you can see this is a nice piece of science scholarship; very important study + Took social science questions v seriously and applied credible scientific methodology to an illumination of those questions - One reason for this study is that ppl noticed that ppl who tended to rely on the radio for political information (radio reliant ppl) tended to be democrats. Ppl who relied on the newspaper for current events, info, and politically relevant info tended to be republicans. Some ppl thought this is another form of powerful effect; radio brainwashes people into becoming democrats, newspaper brainwashes people into becoming republicans
Search for subtler media effects was energized by...
- Popular appeal of 'medium theorists' like Marshall McLuhan o Talked about media effects in an engaging way so that regular ppl who weren't media scholars could understand and consider them. He did this in an interesting and charismatic way, caused attention by normal ppl o Said that the medium is the message, which is the basic premise of medium theory + Idea that people so often talk about the content (sexual, violence, political, etc) as being important, but he said the bigger issue is not the message but rather simply the existence of the medium. Says its less important whats coming out of ur tv as the message as compared to the fact that the TV lives inside ur house - Critical cultural theory ("An 'a priori'... no need for proof... framework based on Marxist ideas and critical political economic theory") o Critical cultural theory is not particularly empirical; it is an a priori framework for understanding societal effects o You have an idea about the way society is and how it should be, and then they pick bits and pieces of research or anecdotes to document something they already believed; it's not where the data is necessarily driving the theory, it's more like the theory is being propped up by bits and pieces of evidence you used to support it - Critical political-economic theory o Theory is critical of the fact that our mass media institutions exist within a capitalist corporate economic framework, which may well have it's own political agenda o Idea being is that if im a media corporation, i might be inclined to use little subtle ways to filter out messages that are unfriendly to a corporate/capitalist agenda, and filter in messages that are supportive of corporate/capitalist agenda o They're looking at not so much the influence of one's kind of message, but rather what is the subtle influence of the fact that these media corporations exist within a capitalist corporate structure? To what extent is that somehow filtering information in and out of the mediascape?
Why voting behavior may not be the best indicator of effects
- Putting yourself in one box or another (ex voting for Hillary or trump); these kinds of categorical/dichotomous measurement items are usually not that good. It's better if they can exist along a continuum - methodological lesson: the way you measure your outcome variables (dichotomous vs continuous variables ), you can detect there was a difference o This is one reason why the question "who did you vote for in 2016?" is maybe not the best question, maybe better one would be "how likely do you think you are to vote for Clinton? How likely on a scale from 1-7 do you think you're going to vote for ____?" more informative way to ask the Q so that if they are any effects of some manipulation, some exposure to political ads for ex, you're more likely to observe those
Why did people panic after War of the World radio broadcast? (Cantril post hoc survey research)
- Radio accepted as a medium for important information o People thought the radio told the truth o ppl were very used to hearing important info and possibly horrific bulletins on the radio from war in europe o People hadn't fully understood what radio's job description was, what radio was capable of - Historical timing o war broadcasts in europe o more importantly, this was in 1938, the year before was the Hindenburg disaster, which was a Zeppelin that was from Germany, that left germany and was supposed to touch down in NJ (May 1937). There was a broadcaster covering this event, said it was beautiful and magic and bigger in person, and then he suddenly started screaming about how it blew up. + In the world of wars BC, they told the voice actor to make yourself sound like the guy who BC'ed the Hindenburg disaster - People tuned in late, missed warning that this was a dramatization o People tuned in late because on a different radio station, there was a popular entertainer who had his own radio show and a lot of ppl were watching that show. On that other show, there was a singer that wasn't too popular (Edger Bergen), so they went to see what else was on. So they tuned away from other radio show and accidentally tuned into war of the worlds, but missed warning - Realistic format and "dramatic excellence;" use of real places ("Grover's Mill".. Specificity effect) in NJ/NY o Certain amount of dramatic excellence in this show; very well done and very realistic o Martians landed in a small, not well known town in NJ, Grover's Mill o If they landed somewhere like LA, viewers would figure that aliens in fictional shows/films always seem in land in major cities like LA o Could be related to persuasion technique called the specificity or a precision effect + Means i'm prob more persuasive if i'm giving you very specific information rather than what appears to be generic info + These little details where you don't get say, it landed in NYC or just NJ, but instead pick a little specific detail like Grover's Mill, it gives us a sense that this really maybe did happen - Use of "experts" o Heavy reliance on ppl who have titles and from other names like Dr or Professor and when people are uncertain they rely on ppl who have authoritative titles - Use of internal/external checks with various degrees of success o Some ppl where better than others at examining the narrative internally and externally + Internal checks are things where you don't have to look out the window or call your friend; there's just things internally in the BC that don't make sense, that strongly suggest this isn't real (Ex: martians getting to earth in only 15 minutes from Mars, seems internally wrong) + Externally means if i live in the vicinity at all, and the broadcaster is saying "clouds of smoke rising over Grover's mill..", and i look outside the window and i dont see smoke, well im doing an external check on the reality of the nature of the broadcast with the reality of the real world. If they dont match, i start to get suspicious that this is a piece of fiction - Personality differences were important: (1) religious, (2) phobic, (3) emotionally insecure, (4) low confidence, (5) fatalistic o Ppl who were more religious were more likely to believe it, particularly if their religion was fatalistic o People who were more phobic, generally fearful of things or irrationally fearful of things o People who are insecure are often more persuadable o Ppl who have low confidence more likely to believe o People who were fatalistic and believed the world was gonna end already
Publication bias is also a very real concern ('file drawer fate')
- Researcher knows they have a better chance of getting published if they find a significant finding that the field thinks is the correct answer - Ex: research war between opinions on media violence o Both sides produce a meta analysis (analysis of analyses) o Journal editors prefer to publish a significant finding (although they say they don't), so researchers are inclined to produce results that align with what the rest of the field thinks - Ex: you conduct a study and find that violent games aren't associated with aggression... you're concerned that no journal will be interested in publishing, so you just hold onto the study and "file it away" like it never happened = "file drawer effect" - Ex: you conduct study that finds that violent video games makes ppl LESS aggressive o You're worried editors think you're whack, because there's just no way your finding is true - Therefore, the world of PUBLISHED research may not accurately reflect the world of CONDUCTED research - publication bias, fishing for results via using a bunch of control variables, and doing a bunch of analyses to find statistical significance are probably why the reproducibility rates are so poor [MULLIN'S OPINION], more so than fake-data
Experimental
- SS are randomly assigned to conditions o group A and B are essentially clones, then you manipulate something in their environment - These "clones" receive differential treatment (indep. var) o Ex: group A receives violent content, Group B receives amusing content - Any diff on outcome measure (dep. var) must (generally) be due to IV - Advantage: Control, and thus ability to make causal statement - Disadvantage: artificial (poor external validity) o making subjects watch something they typically wouldn't, unnatural circumstances o External validity = to what extent you can generalize beyond the lab to the real world
Types of mass media effects or theories: Short term vs long term (we see a lot of both)
- Short term effects mean that you might have a study where you bring ppl into a lab and then show them porn and right after you ask them some question about their attitudes towards women and you might notice some short term effects, that they're more likely to sexualize women for the 15 minutes after they viewed porn, but 30 minutes later they're back to normal. Those might be considered sometimes priming effects that the content made me think of something / certain attitudes and colored the way I view the world for a short time, but I eventually go back to normal. Effects don't persist o Short term effects are pretty easy to do in a lab and in a constrained time frame and you can manipulate what ppl see in a lab and you can observe their behavior shortly after in a lab o Easier to conceptualize the research on these kinds of things - Long term effects would be more like "well i grew up in a culture where there were lots of ads where women were sexualized and portrayed sexually in films and tv, and this has reinforced a sexualized message impression of females, so it's possible that whatever this effect is, it wasn't from one exposure to 30 minutes of porn images, it was a lifetime of growing up in this culture where those images were widely available and those kinds of effects (socialization effects) might be harder to do in a lab o Cant bring someone into a lab for 10 years and manipulate their exposure to messages. It happens across one's developmental lifespan and the effects of those things may persist for a long time o A lot of the effects that might have more long term importance might be these socialization effects that accrue over a long developmental period of time
"Significance" (e.g., p<.05) lets you know if an observed pattern could be due to chance
- That observation can not be due to chance variation, I have observed something that is real o if fewer than 5 times out of 100, then it wasn't due to chance because it is so unlikely
A "significant" study's observation cannot be due to chance variation in the dataset
- The problem is that anything could be due to chance; we can never be 100% certain that it's NOT - Given that we always have to tolerate some uncertainty.. The cutoff is .05... if the probability is <.05 [it could have occurred fewer than 5 times out of 100]... then it wasn't due to chance o Secretly we know it COULD be due to chance, but that's the cutoff
The sleeper effect
- The sleeper effect is if you absorb information from any source (here we're thinking of a media source), you tend to remember the information, but eventually you kinda forget the source - Ex: if subaru comes out with a hybrid car, and they say " our hybrid battery is better than prius bc ours gives you more miles per charge, etc etc" and you see that in a subaru ad, you are going to discount what they said a bit bc you know they're trying to sell you a car (not objective info). So i remember the info but ive discounted it. Lets say a year later im shopping for a hybrid, trying to decide between subaru or prius. One thing i like about the subaru is i remember they get more miles per charge. I remember the information, i just forgot the source - In the context of this study, if im exposed to a movie about american astronauts who land on the moon and they discover alien life forms there, you come out of movie thinking "thats interesting, alien life forms on the moon..." but ofc it was just a fictional narrative so you tuck away a little bit of info, but you've completely discounted it. Imagine a year later, you start to think about this and you say "i don't remember where i heard that, but i do know there was life on moon" - You can somehow discount information you get from a media source at first, but as you forget the source, the persuasive impact of that information can become greater
Types of mass media effects or theories: Attitudinal vs behavioral vs cognitive vs emotional*
- The types of outcomes we might be looking at would be attitudinal o Your valence towards something, your positive or negative feelings towards something. Do I like it or dislike it? o More than just a belief thing, am I in favor of or against it? - Cognitive is your beliefs and thoughts about something - Behavioral is what you actually do o You can maybe affect cognition and attitudes (for ex towards smoking with ads against it in the media), but it's hard to affect behaviors + Ppl who are smokers are gonna be resistant to giving up that behavior - Emotional outcomes
"Reproducibility project" (Nosek, 2015) raised HUGE questions about to what extent the scientific method is reliably guarding about selective human biases and biases of any type
- They took about 100 studies that had been conducted and published in the previous 10 years and sought to replicate or reproduce those results o After following the same original studies, wanted to find out to what extent will we also produce the exact same findings? o You would want to see about 90-95% of those studies successfully wielding the same original findings - The reproduction rate for social psychology journals was on the order of 25%, which is horrible - Cognitive psych was only about 30%, still horrible and shocking - What could possibly lead to such a rock bottom reproducibility statistic? Is it that ppl are lying/making up data? o Mullin doesnt think this is the case, but it is possible that some ppl are making it up - Overall, some people question how well the rules of science are actually enforced, and many think these rules aren't enforced that well, and if this is true, it helps explain why we have some difficulty reproducing previous results
Experiment design from Payne Studies looked at media effect on attitudes: Subjects (SS) attitudes assessed, then ss given tickets to film with predictable theme; day after viewing, SS atts assessed again (pic on pg 8)
- They were looking at to what extent can film content or do films impact subjects' attitudes on particular ethnicities? o Two ethnicities these studies were concerned with was Chinese immigrants and African Americans o Wasn't just looking at negative effects, but just effects in general - Design was kids attitudes assessed towards those ethnicities and other stuff, ss then given tix to see certain films, so they're controlling for what the ss are going to consume. After viewing ss atts assessed again o Manipulating what content the ss are gonna be exposed to o After second ss assessment, researchers looking for changes - In a causal relationship study, you're hypothesizing that A causes B (not just related to) - IV is the thing you're hypothesizing as a possible cause, DV is the thing you're hypothesizing as the possible outcome - Here they're thinking maybe film exposure (IV) changes/affects attitudes (DV) - This study employed a pretest (not always necessary) - Study used random assignment (RA) to assign ss to three diff groups o On average, the kids in group A are equal in height and age and weight and intelligence to kids in groups B and C o RA is going to create (on avg) equality across those groups on some variable, including attitudes towards afr americans and chinese immigrants o We know at the beginning, even if we dont have pretest, we know a b and c are equal - At the outcome, are they no longer equal? o This is what can establish a causal relationship. if they were equal in beginning and then later not, and if the only other thing that could have possibly accounted for that non equality is what you did to them in the intervening between RA and measuring outcome measure, then that provides pretty good evidence that ur IV caused ur DV o This study did a very good thing by checking up on the kids months later to see if the attitude change had persisted - The films that they used were themed around certain ethnic themes o The antiblack film was D.W. Griffith's The Birth of a Nation, very famous and very very racist film. Heros of movie are the KKK o The Pro chinese film was Son of the Gods, which was a favorable or sympathetic tale of chinese immigrants - The idea is that possibly the antiblack film is going to change attitudes towards african-americans in a negative direction, but maybe the pro chinese film will change ppl's attitudes in a pro chinese immigrant direction, this was their hypothesis o They found exactly what they hypothesized
Two-step flow came out of Lazarsfeld study, meaning ideas flow from media to "opinion leaders", who in turn influence less interested, less informed individuals
- Two step flow: Very often, media effects don't necessarily hit everyone equally, some ppl might be disconnected from media influence, but the way media influence may trickle down to the populace would be that ideas flow from the mass media to the world opinion leaders. Opinion leaders are those whose opinions you trust and are looked up to o Opinion leaders could be parents, a friend who knows way more about politics than you, celebrities - This model says that the path of influence here is a two step thing from the mass media affecting opinion leaders who then in turn affect me
the 'rules of science' are sometimes flexible, and even very ethical researchers may discretionarily bend them in their favor [helping them find things that they want to find, or vice versa]
- Types of analyses, number of analyses, interpretation of results, etc. o if you do enough analyses, one of them will eventually come up as statistically significant (even tho it's due to chance) + research ethics says you must report how many total shots you're taking
Survey/Correlational Research
- Uses pre-existing data or questionnaire data o Ex: socioeconomic level of various communities and how it relates to crime rate o Ex: separate survey questions gauging one's consumption of media violence and their aggression... try to find correlation - Advantage: Naturalistic (ex: Not being manipulated in a lab) o Not forcing subjects to view something they wouldn't see otherwise o No manipulation - Disadvantage: o Can't make causal claim (though some researchers pretend you can) + Ex: if someone is naturally more aggressive then of course they're more likely to select violent content... not causal o Directionality and third variable issues + Perception and behavior can have a bidirectional relationship (being scared at home may make you watch tv more) + 3rd Variable: might be crime in neighborhood - For example Media violence and aggression o Ask about how much tv they watch with violence in it & how aggressive they are - Beware: structural equation modeling/path analysis o Researchers "dress up" their correlations to kinda make a causal claim (even tho you can't) o Ex: ppl feel that alcohol on TV contributes to underage drinking... so greater consumption of TV contributes to more drinking behavior for adolescents + this is NOT supported by research ^, BUT they can dress it up and say TV consumption is positively associated with identified alcohol marketing elements... and the elements are associated with likelihood of teen alc consumption + Correlation between A → B, and B→ C... but NOT A → C + Certain studies take path analysis to conclude that A→ C * The communication field generally embraces structural equation modeling/path analysis, even tho all the data is correlational o MULLIN IS IN THE MINORITY FOR BEING AGAINST THIS ^^^
Inferential stats (guessing population from sample) assumes you have representative sample
- Volunteer samples, poor response rate, other selection biases hurt representativeness o Volunteer samples are horrible/meaningless bc they have no representative value o Poor response rate hurts representativeness even if you started with a good sample + Ex: 200 of 300 men drop out of the study... now it's no longer 50/50 men and women - Ex: if you want to measure every single student's height at UCSB to calculate average height, there would be no guesswork/inference needed o But this is impossible ^, we hope to get a representative sample (get 600 students that are representative of the population, randomly sample off of the UCSB registrar) o Ex: you measure avg height of 5'7" from your sample... statistically establish a range to be applied to the whole population + Confidence interval ["I'm 95% certain it's in this range"]
"War of the Worlds" radio broadcast (1938)
- War of the worlds is a functional book written by a science fiction writer, H.G. Wells about martians who come to earth and start to destroy humanity - The Mercury Radio Theater group did a dramatization of that book, and it was announced as such on the radio and in the newspapers. There was a pretty fair heads up that this was not a real news broadcast and that it was a drama for the radio - Massive panic (1.2 million of 6 million listeners, plus many non-listeners) amongst a wide array of persons, despite warnings, etc o 1.2 million out of 6 mil listeners believed this was an actual martian invasion, caused mass panic o Many ppl who didnt even listen to BC heard from neighbors who did listen also panicked o Widespread panic in response to a radio show that is transparently a piece of fictionalized drama - FCC got involved, hearings followed; Cantril put together a post hoc (after the fact) survey project o The Mercury Radio Theater was in a bit of trouble, got hauled in by the FCC and before congress to apologize for the irresponsible behavior of doing a great job dramatizing a work of function Cantril, after the fact, did some survey research and was able to identify some themes that may have led to this result that some people were overwhelmingly affected by this media content. Had to be post hoc bc he couldn't have known this was all going to happen before it actually occurred, and looking at things post hoc limits you a bit + Maps onto this idea that the media can be super influential and have very powerful effects among some people
within the scientific method, theoretical acceptance is based on testability and parsimonious fit with data
- Within the scientific method there's this idea that there's going to be competing theories; 2-5 theories that explain phenomena and do so with equal ability to predict things. The theory that's gonna win out is going to be the more parsimonious theory, meaning it explains the same amount of stuff but does so with less assumptions and less theoretical baggage you have to assume.
Types of mass media effects or theories: alteration/change vs stabilization/reinforcement
- alteration/change: how does this form of content/persuasive campaign change people's attitudes or alter their behaviors? -stabilization/reinforcement: a good deal of media effects is more aimed at stabilizing attitudes and behaviors and cognitions or reinforcing them o This might be the more important media effect; not changing the behavior or attitude, but reinforcing or stabilizing an existing behavior or attitude
1960s: the Powerful effects model returns
- replaces limited effects paradigm
Types of mass media effects or theories
-Micro vs. macro -Direct vs conditional -Nomothetic vs. idiographic -Message effects vs medium effects -Attitudinal vs behavioral vs cognitive vs emotional* -alteration/change vs stabilization/reinforcement -Short term vs long term (we see a lot of both)
the "Limited Effects Paradigm" era
...
Media campaign functions found in Lazarsfeld's study
1. Activation (14%): arousal of interest, attention, selectivity o For many ppl, it's main function is to activate ppl and arouse their interest in the campaign and the election, grabs their attention, causes them to be more selective in the info they consume or how they interpret that info o Not changing my opinion, but just making me more interested and paying more attention 2. Reinforcement (53%): justification for not changing mind o For over half the ss, the main function of a media campaign is to reinforce existing beliefs, give ppl reasons for not changing their minds o If im a republican, the campaign's info mission is going to give me reasons to not change my partizenship and opinion 3. Conversion (8%) o Changing people from a democrat to a repulican and vise versa o Small point about the media landscape is converting one voting choice to another
3 types of research we'll see:
1. Survey/correlational 2. Experimental * In any research domain, you'd like to have both surveys and experiments so they can compensate for each other's strengths weaknesses 3. Content Analysis - With ALL THREE of these designs, there are always issues with conceptualization and operationalization of variables - Conceptualization o Ex: what "counts" as a responsible media message about sexuality? [STDs, pregnancy, emotional readiness msgs] - Operationalization o Ex: how exactly are you going to actually measure it (coding the responses + Ex: Should a joke count as a response?
History of Media Effects Research: Early 1920s
Hypodermic/ "magic bullet" model era
A 6-stage model of media effects research history (Neumann and Guggenheim article)
I. persuasion theories (1944-1963) - During this era, it was basically powerful effects model kind of stuff o This being I hit you with a message and how do you respond? - Straightforward persuasion; information leads to attitude change II. active audience theories (1944-1986) - Shift towards a minimal effects paradigm due to audience activity and selectivity - Uses and gratifications o Perspective puts the user in the driver's seat o Not focusing on how im affected, but what i'm selecting and why - Minimal effects and selective exposure III. social context theories (1955-1983) - More emphasis on the social context, how info flows thru society o Not just how i as an individual might be affected, but how does information flow through society - Two-step flow o Idea that info might find its way into the social body, not b/c everyones necessarily affected similarity by the information or not at all, but rather that media info may arm opinion leaders with info, who can communicate that to people whom you know, who look up to them - Diffusion of innovations IV. societal and media theories (1933-1978) - 'Media and society' accumulative focus, includes political economy, cultivation o Idea of messages and effects accumulating over one's development lifespan - Cultivation theory o Drip rather than drench hypothesis + Drench hypothesis implies that you see one media message thats so loud and powerful that it might have, on its own, some effect + Drip hypothesis is this idea that the accumulation of tiny effects can have big impacts V. interpretive effects theories (1972-1987) - Interpretive models where salience (more than just attention change) is the focus - Agenda setting theory: idea is that the media may not be overwhelmingly powerful in telling you what to think, but it is overwhelmingly powerful in telling you what to think about. Doesn't necessarily brainwash your brain with respect to the correct answer, but it tells you the things im supposed to pay close attention to and regard as important issues VI. New media theories (1996-) - In mass media context, could include 'demassification' issues (e.g. 'echo chambers') o Demassification: in the 70s, everyone was aware of the same tv shows because there were 3 tv channels, so everyone had a shared cultural experience and we were more like a mass in those days. We now live in a very demassified landscape where we are selecting our own videos and idiosyncratic media. We are living our own little idiosyncratic media lifestyles in a way that may not overlap that well with other ppl's media lifestyles. There's less of a sense of shared cultural experience o Echo chambers: the political implication of this demassification - We've had to really rethink older media theories in the wake of new media technologies o For ex, we have to totally rethink agenda setting in the new media landscape. Nowadays people don't all have same exposure to same informational sources like they did before; they may evaluate the importance of things very differently, depending on where they get their info
Kuhn's paradigm shift concept suggests anomalies can only be explained for so long until a more compelling paradigm gains acceptance
Kuhn's paradigm shift suggests that there will be a dominant paradigm for a while, science will operate in a normal fashion, within that paradigm, anomalies will continue to add up, the scientific community will ignore them for as long as it can to keep that original paradigm alive, but at some point the weight of these anomalies become so great that you have to shift to a new paradigm
Media effects theories graphic from Macnamara article is on page 22 of notes
direct effects-> minimal or limited effects theory-> cultural studies-- neo-marxist ->political economy -> political economy -- 'public sphere' theory' -> cultural studies -- 'new audience research' (similar to uses and gratifications approach' in that it emphasizes role of audience activity
Stats
the language by which observations are processed and communicated