Table 1

Helping Students Understand the Article

Section

 Tips, Comments, and Problem Areas

Abstract

Note that correlates are not necessarily consequences (effects). Therefore, the first correlational study did not tell about consequences of newspaper reports. It was necessary to do experiments (Study 2 and Study 3) to find out the consequences (effects) of newspaper reports.

Endorsement of: agreement with

Mutability: ability to change, flexibility

Political bias in the media

Political ideology: system of political beliefs (e.g., conservative ideology, liberal ideology)

Status quo: existing conditions; the way things are.

Sociocultural explanations: social and cultural causes; reasons that are due to culture and society; due to the effects of our social and cultural environment

Research objectives

Immutable: unchangeable (by their environment)

Selection of Newspapers and Articles

The authors did not want to examine all 50 newspapers, so they used random sampling.

A Boolean search (which is something you will probably use when you search for articles) might have been something like this:

“Pub date = January 1994 to February 2001 AND sex differences AND human behavior * AND research AND cause AND not animal”

The authors did not want to read 28, 717 articles so they used random sampling.

For one paper, they compared the population (every article that met their criteria) with a random sample of 10 articles taken from that population.  Their sample was similar to (seemed to reflect) the population.

Results

R: Like a Pearson r, the multiple correlation (abbreviated R) is a measure of the association between one variable (proportion of biological explanations) and a second variable.  However, with R, the second variable is not a single observed variable (presidential beliefs) as it would be in calculatingr. Instead, in calculatingR, the second variable is the predicted value of the first variable (the estimated proportion of biological explanations). This second variable is a combination of several predictor variables. Specifically, it was calculated by weighting a combination of at least two variables (in this case, presidential endorsements and sex role beliefs) in a way that would make the predicted value of the criterion (expected proportion of biological explanations) as close to the actual value of the criterion (proportion of biological explanations) as possible. Thus, one difference between  r and R is that R is looking at the correlation between a set of predictors and a to-be-predicted variable rather than looking at the correlation between  one predictor and a to-be-predicted variable. Another difference is that whereas r can vary between –1 and +1, R can only vary between 0 and 1: R cannot be negative.

R²= .29: Just as we can square Pearson r to determine  how well scores on one variable predict scores on another variable (specifically, r² gives us the coefficient of determination: the percentage of variability in scores on one variable that can be predicted by knowing scores another variable),  we can square R to calculate how well  scores on one set of variables predicts scores on another variable. In this case, we know that 29% (.29) percent of variability in the extent to which newspapers use biological explanations for sex role differences can be predicted by knowing both the newspaper’s candidate endorsement and the newspaper’s editorial position on the admission of women into military academies.

b: Like Pearson r correlation coefficients, bs give you an index of the relationship of the predictor with the to-be-predicted (criterion) variable. Indeed, if the regression equation has only one predictor, b will be the same as r. However, interpreting bs is tricky when you have more than one predictor and the predictors correlate with each other.