Smasharoo wrote:
It's actually a pretty crappy push style study anyway (it makes numerous assumptions about how gender and race roles tie into welfare that I don't agree with and which completely affect the results of the data that is collected). But even then, it ultimately supports my point far far better than yours or Joph's (that his perception is more about his own projection than about the reality).
Can you cite the specific issues you find with the methodology?
Not methodology specifically, but assumptions that were made. First one is here:
Quote:
My assumption going into this study was that the notion of the welfare queen had taken on the
status of common knowledge, or what is known as a "narrative script." The welfare queen script
has two key components welfare recipients are disproportionately women, and women on welfare
are disproportionately African-American.
I don't have direct access to the raw data of the study, but I'm betting that this assumption colors the resulting analysis of that data as presented in the abstract itself. No way to say for 100% sure, but I'm always leery of this sort of thing. He should not have any assumptions, but follow the data. If you assume, for example, that drivers of red cars drive faster than drivers of blue cars, you might analyze data showing that red cars are involved in more accidents than blue cars and conclude that those accidents were caused because of the assumed faster speed (and thus kinda circularly proving your own initial assumption). But you can't ignore the possibility that your initial assumption was wrong, and that the increased rates of accidents are because of something completely different, like a certain model of car is more likely to be red than blue, and that model has some control problem (bad wheels, steering, or whatever). Following your assumption, and crafting a study based on it, will likely cause you to fail to find the actual cause of your data.
Here's a set of criteria he uses that are easily subject to his own initial bias (again, I'd love to see the specifics of the questionnaire and how this was scored):
Quote:
Three different categories of attitudes were addressed. The first pertained to their attitudes about
the causes of and solutions to welfare. I was able to measure the number of people who believe
that individual failings were the cause of welfare. On their questionnaires, this group of viewers
indicated that they believe welfare recipients cheat and defraud the system, that they abuse the
system by staying on too long, that welfare undermines the work ethic, and that welfare
encourages teenagers to have kids out of wedlock. They also tend to indicate a high level of
opposition to various public assistance programs (e.g., AFDC, food stamps, subsidized housing
and health care).
The second set of attitudes is related to racial beliefs. I was able to determine the percentage of
people who endorsed negative stereotypes about African-Americans. I did this by eliciting
responses about perceptions that African-Americans are lazy, sexually promiscuous, not law-
abiding and undisciplined. I also computed the percentage of participants whose views and
attitudes were described in more subtle terms. Included in this category were phrases such as
"blacks don't try hard enough," "they should pull themselves up by their bootstraps," and "there is
not much discrimination nowadays."
Finally, attitudes about gender were numerically measured by the percentage of people who
preferred women to play more "traditional" gender roles. These attitudes emerged from
responses to such statements as "the husband should be the achiever outside of the home,"
"working women do not have as close a bond with their children as mothers who stay at home,"
and "a preschool child is likely to suffer if mom works."
There are a whole set of potentially questionable assumptions in there. So anyone who believes that there isn't much discrimination is identified as being "subtly racist"? And simply saying "they should pull themselves up by their bootstaps" is as well? Um... I'd say that about anyone, not just blacks. But of course (and this is another thing that isn't clear about he underlying methodology) if he only asked this about blacks and not whites, then he's creating the perception that this is some kind of special/bad treatment or view of blacks, when it may not have anything to do with race at all.
I'll also point out that someone might have a negative view of those who use welfare without having a negative view of blacks. This connection only exists *if* our analysis of the data assumes that people view welfare through the narratives of the "welfare queen" script. Which, as I pointed out above, is precisely what is going on with this study. Thus, as far as I can tell, he's concluding that anyone who has a negative view of welfare recipients has a negative view of African Americans. This is relevant later.
I have similar reservations about the correlation between the questions asked about gender roles and the assignment of "traditional" and "liberal" views. Again though, still don't have enough data, but that's kind of the point. I can't really be sure if this guy is just spouting complete BS or not, but the squirrelly nature of the results tends to make me thing he is.
This one is just bad (and obviously biased) reporting of findings:
Quote:
The first finding is that the welfare queen script has assumed the status of common knowledge.
When white subjects were asked to recall what they had seen in the newscasts, nearly 80
percent of them accurately recalled the race of the African-American Rhonda. On the other hand,
less than 50 percent accurately recalled seeing the white Rhonda.
Ok. What about black subjects recalling white versus black "Rhondas"? Is this really about people associating "welfare==black", or possibly that people are more aware of a person of a different race on the screen versus someone of the same race? I can think of several possible explanations of this data that don't support the assumption he's giving. Of course, more complete data on the questionnaires would help clear this up.
This pattern of only looking at white responses continues through the entire results section. Which kinda automatically makes me suspicious that this "experiment" is less about finding facts and more about manipulating them to support a starting assumption.
Next:
Quote:
I also contrasted responses among subjects who viewed the welfare story and did not have a
visual cue and those who saw our welfare story featuring Rhonda Germaine (either a white or
African-American image of a woman). I expected that participants who saw a woman in the story
would be more likely to endorse traditional women's roles, oppose welfare spending, and cite
individual causal attributions.
Not only was my expectation wrong but two other results emerged. First, seeing a woman in the
news story actually decreased opposition to welfare spending. Second, exposure to a welfare
queen in the news significantly increased support for negative characterizations of African-
Americans by an average of 10 percent.
He doesn't elaborate on exactly how he was wrong, but we can assume that this means that there was not the correlation between a traditional gender role viewpoint and the assumed negative welfare viewpoint (with it's assumed racial context). Um... Which is the point at which he should be saying "I guess I was just wrong and should start over".
I'll reiterate that based on the earlier stated assumptions, as far as I can tell the "negative characterization of African-Americans" he speaks of really just means "negative characterization of welfare recipients" (cause he's assuming people view them as the same, just as Joph's mother allegedly did). Actually, it not really clear from this sentence what his data actually showed. But I smell spin.
And finally:
Quote:
Finally, I examined the racial effects by comparing those white viewers who were exposed to the
white Rhonda and those who watched the welfare story featuring the black Rhonda. The general
expectation was that exposure to the quintessential welfare queen script (i.e. the black Rhonda)
would increase anti-black sentiments, heighten opposition to welfare spending, and lead more
people to cite individual failing as the cause of welfare.
The results were somewhat mixed. True to form, exposure to the full confirmation of the script
(i.e. black Rhonda) increased opposition to welfare spending by five percent and showed a 10
percent rise in an attribution of cause to individual failings. Likewise, white participants who
watched the welfare story with the black Rhonda were more likely to hold negative views of
African-Americans than those who did have a visual cue. Contrary to expectations, however,
exposure to the white Rhonda produced the biggest increase in anti-black sentiment. That is,
watching a story with the white Rhonda increased negative depictions of blacks by 12 percent
compared to the black Rhonda and by 23 percent over the story without a picture.
I have a sneaking suspicion that a good portion oft he "mixed" and "contrary" results also derive from the bizarre assumptions he's making. If you drop the assumption of the welfare queen narrative, I suspect that the data here actually would make a lot more sense. For example, I'm going to speculate that the reason white exposure to a white rhonda produced the biggest increase in "anti-black sentiment" is because he's assuming that certain forms of anti-welfare sentiment are actually anti-black sentiment (just as Joph did). But if his starting assumption is wrong, then that's really just anti-welfare sentiment, and has nothing to do with race, and the correlation between higher opposition when someone you associate with more strongly (someone of the same race) is shown to be on welfare, makes complete sense.
This was the portion that really made me suspect that his starting assumption was not just wrong, but ******** with his results. Drop the assumption, and the result aren't nearly so mixed or contradictory.
Oh. And then there's this (which was admittedly just me poking a bit of fun at you guys):
Quote:
Most interestingly, people who espouse the most "liberal" views about gender roles turn out to be
the most hostile to blacks when they are exposed to the white Rhonda. Put differently, the most
gender-liberal white participants appear to be most likely to implicitly blame African-Americans for
the plight of their racial peers, and there is early evidence to suggest that this tendency is most
pronounced among women.
Again, of course, if the correlation between "anti-welfare" and "anti-black" is carried through here, this ceases to be a funny racism counterbash at liberals, but actually makes a lot of sense. People who see women as being just as capable in the workplace as men, and who think careers should not be outweighed by child rearing (or homemaking) would logically be more upset about women on welfare than those for whom accepting the "welfare" of a husbands salary to stay home and raise the kids is more normal.
Given how much better the results seem to match up when one rejects the starting assumption, perhaps we should do that? Of course, that is precisely the same assumption that Joph was insisting was commonplace, wasn't it? This is why I said that this abstract didn't say what you seemed to think it did. If you read it and think about the results, if actually suggests that the welfare queen narrative model is wrong.
Quote:
I don't want to imply that you didn't read it and are just making vague guesses about what you think "problems with studies" are, but some people are saying you didn't read it all.
Actually, I was suggesting that you didn't read it. I'm hoping that my response proves that I did.
You're welcome Smash! Keep up the good work. We're all counting on you.
Edited, Oct 3rd 2014 5:26pm by gbaji