Hypothesis
The independent variables are maternal education and quality of parenting. The dependent variable is the scaffolding behaviors.
Subjects Studied
Participants were mothers/child pairs previously enrolled in a longitudinal study of families with children from 4-5 years old.
The enrollees were originally enrolled in 2002-2003 from southern England. Schools were asked to send letters home with children aged 4-6 years old with older siblings aged 8 or younger. Many schools were uncomfortable with sending letters home to specific students, so they sent letters home to all students. In the first part of the test 173 families participated, in the second phase of the study, 4-5 years later, 106 of the original families were recruited.Research Method
This study used correlational research methods from a longitudinal cohort study using interviews, observational studies, and self-report data sets. There was also a cross-sectional component to the study.
Conclusions
The data collected was analyzed using many techniques. Initial study data (verbal mental age, maternal education) were tested using Gaussian distributions.
Parenting measures were analyzed using Cronbach’s alpha, which is a measure of internal consistency within a study. A higher value for the alpha is desirable as it shows the consistency of the data. As a general rule, a score of 0.70 or higher is wanted. In this case, the alpha value for expression of affection was .78 and.8, for parent-child conflict it was .86 and .82, for the discipline category it was .53 and .74, and for negative discipline it was .53 and .67. Thus, the consistency of the parenting measures was good except for the discipline categories that have questionable reliability.
Factor loading was also used in the studies of parenting quality. This means that data sets were added together using multiplication techniques based on composite scores. Essentially it correlates the coefficients between variables and factors. It is used to get a percent variance score.
A Cohen’s kappa was established to measure inter-rater reliability. It supposedly takes into account chance agreement, but on the whole it is a suspect measurement. The kappa equaled .87, which is considered very high agreement. And is theoretically excellent.
Finally, a Sobel test was used which is a specialized t-test that tests the significance of a mediating effect.
Results
The results partially support the first hypothesis. “Positive parenting” subjects showed more contingent shifting behaviors and negative parenting was correlated with fixed failure feedback and negatively associated with contingent shifting.
Maternal education was positively associated with positive parenting and contingent shifting, and negatively associated with harsh parenting and fixed failure feedback. The reviewers found some support for their second hypothesis in the data with a .30 correlation between positive parenting and contingent shifting.
Finally, the authors claim that they also found support for the third part of their hypothesis: that parenting quality mediates the association between scaffolding and maternal education. They showed that it was the combined effects of harsh and positive parenting that mediated the quality of scaffolding.
The problem with all the results is that they were subjected to questionable statistical methods and furthermore they seemed to try account for too many variables. Good science tries to use one variable in order not to confuse and distort data.
Implications of Results
These data sets are not particularly useful. Firstly, the data had to be overly manipulated to test the variables that the researchers were using. The use of Cohen’s kappas, eigenvalues for parenting quality, Sobel values to determine mediation, and factor loading techniques yield suspect statistical analyses. All these methods are routinely critiqued in the statistical literature. Furthermore, the use of regression models on the data for the 2nd and 3rd hypothesis is confusing and indicates that the reviewers were looking at too many variables to make useful scientific discoveries. As well as the possibility that altogether unstudied variables are responsible for the “discoveries.”
The data would have been much more reliable if they had done three independent experiments rather than trying to combine the three. Furthermore, their analyses are ultimately subjective and unreliable as indicated by the low alpha score that the positive and negative parenting received. The entire study is likely confounded by a bias for what is “good parenting.”
Some useful research would look at a child’s actual academic achievement. Furthermore, the authors didn’t take into account that socialization is a “two way street.” A next step in research would be to look at the nature of the child’s disposition and intelligence levels aside from just verbal maturity age.
Future research should be simpler and more exact. For instance, what is the relationship between harsh parenting and academic achievement, or positive parenting and academic achievement? The way the mothers imparted information on putting together blocks is fascinating, but all in all, playing legos doesn’t make for a smart, socialized, well-rounded individual.
Sources
Carr, A., Pike, A. (2012). Maternal Scaffolding Behavior: Links With Parenting Style and Maternal Education. Developmental Psychology, 48(2), 543-551.
