In their book, Hart and Risley study 42 American families...in Kansas. And even though they do study professional, working class, and welfare families, I keep thinking about the characteristics of the sample used. The authors make a slight reference to the fact that their study might not be 100% generalizable over all cultures, such as Native American populations, yet they still stand by their results and the reliability of their data. While I do not doubt that their study has merit, I just keep coming back to the small sample size and the fact that the participants were all from traditional/stable families in a city in Kansas. What are people's thoughts about the sample the authors used? Do you think the results would be vastly different if the researchers had sampled families in the projects of Baltimore City? Or do you think the results are generalizable across many populations and communities? I'm interested to hear what other people think!
This is a very interesting question, and I think the only way it can really be answered is if people do similar studies but with a different sample set.
ReplyDeleteWhat interests me most is not so much the fact that they're from the Midwest, but that all the students studied from from traditional/stable families. On the one hand, I'm glad that they kept this characteristic standard, because it shows that any differences are not attributed to home life (other than income level). At the same time, I'm curious to know if there are any studies that focus on stability of the family rather than socioeconomic level and whether this has more or less of an effect on students' performance.
So basically, I have no answers. But, like you, I'm very interested in how the sample may have effected the study. I wonder if there are any similar studies out there who have looked at a slightly different sample set and how they compare.
Lindsay
Much like Lindsay, I'm not sure I have answers to the questions you pose either, Jamie, but I find it so intriguing that the sample size is limited to Kansas. I'm sure that your experiences as an early childhood educator have you asking the same questions about the hypothetical outcome and results of a study in Baltimore City.
ReplyDeleteWhile the findings may be similar in our environment, I believe there would be vast differences in the vocabulary used or the strategies parents implemented to encourage and expand vocabulary use based on their everyday experiences in their communities. Vocabulary and language are so closely tied to cultural experiences, and cultural experiences can be unique to the communities from which they grow, despite the possibility of being similar to others. The experiences of the parents and children of Baltimore City would be unique to Baltimore City. That being said, I don’t necessarily agree that the statistical findings would be different, but if educators are going to attempt to address the deficiencies in their students' vocabularies we will need to be acutely aware of what our own students' strengths and weaknesses are of their spoken vocabularies.
Molly
I am honestly not sure if the geographical location would really matter. I don't mean to diminish the importance of geography to different lifestyles and experiences. On the contrary, I think especially in a country the size of the United States, with very, very disperse and different ways of life across our many different regions, geography can play a fundamental role to whom you become as an individual. However, I do not see any reason to assume that these results would not be transferable to nearly any home life. The results are not contingent on those producing the words, but simply the numerical representations of the words produced. If there is a disparity in the amount of words, the results will almost certainly replicate.
ReplyDeleteExcellent question, Jamie.
ReplyDeleteI thought about that a great deal while I was reading the book, and I was surprised the authors didn't dedicated more time to answering that question.
That being said, they did address this issue in more detail in chapter 9 (page 195 in my book) when they said this:
"Most important to emphasize is that we observed only well-functioning families in the mainstream of American culture; none of the families were dysfunctional, severely stressed, abusive, or addicted. None were independently wealthy. No persons with disabilities were present in any of the homes. Although not all the families owned their homes, they were less transient than may be typical...."
and then, interestingly, the authors said:
"The findings...may be less applicable to families living in deteriorating neighborhoods in inner cities where crime is prevalent, to families with children whose language development is delayed or impaired, to bilingual families, and to families who live in rural pvoerty such as that experienced by migrant workers."
She then acknowledges that different cultures is also important.
In the end, though, they do say that they feel confident "extrapolating to other hours of the day and to other months and years for these 42 families." They changed the times of the day they observed, and the families were also always changing. The only thing that stayed constant: their vocabulary data.
The authors finish that section by saying that their numbers and data were comparable to others' findings in different geographic, environmental, and familial locations.
In conclusion, I'm confused and I am still questioning the sample.
Austin
Like Austin, and the rest of the thread it seems, I have gone back and forth on this issue.
ReplyDeleteReading through the responses here and rereading the sections that Austin highlighted has me thinking about the impact of the quality of the vocabulary children are exposed to. For instance, would the results change if children heard high-level vocabulary, but heard less of it?
After rereading chapter nine, I'm also curious about the results that would be observed in dysfunctional families.
At the end of it all though, the focus was numerical and I'm inclined to agree with Aaron that the data probably holds true regardless of the content and other variables that have been discussed in this thread.