Table 1

A Student Guide to Understanding the Article

Section

 Tips, Comments, and Problem Areas

Abstract

Five-factor model (FFM); sometimes referred to as the “OCEAN” model, O=openness to experience, C=Conscientiousness, E=Extraversion, A=Agreeableness, N= Neuroticism. These five factors are explained in the results and discussion section of the article.

The Riverside Behavioral Q-sort will be described later.

To see a more informal summary of the study,  go to the following URL: http://children.camden.rutgers.edu/seminars/associate/p_markey.htm

Introduction

Page 423—

You probably will not have any problems with this page.

Page 424—

The second full paragraph is fairly difficult to understand—and you can understand the article without it. The main point is that whereas  infants can display only a limited range of behavior, adults can display a wider range of behavior. Despite the fact that infants display a limited range of behavior, there are individual differences in how infants behave. Traditionally, psychologists have explained differences in how individual infants typically behave in terms of different infants having different temperaments. Traditionally, psychologists have explained differences in how individual adults behave in terms of different adults having different personality traits.  The authors also note that adults are able to express a wider range of behavior because adults are more developed than infants. The authors suggest that one reason adults are more developed than infants is that adults have been exposed to more complex environments (a job rather than a crib).

The third paragraph expands on an idea mentioned in the second paragraph—that traits may just be temperaments displayed in a more complex and sophisticated way.

Fourth paragraph—

      Robust: holding under a wide range of conditions

      Parsimonious: simple, not requiring many principles (in this case, only five factors to describe personality)

Fifth paragraph—

      Post hoc: after the fact (usually theories should predict events)

Sixth paragraph—

       “Factor structure similar…”: same number of factors (in this case, five factors) 

Seventh paragraph—

Idiosyncratic: an error specific to a particular individual (rater)

Systematic errors: errors that may push scores in a particular direction (e.g., giving consistently high ratings).

Halo effects: rating a person who was high on one dimension high on other dimensions based on the assumption that if a person is high on one dimension that person must be high on other dimensions.

True substantive relation: real relationship between the variables

Method variance: similarities in scores between measures of variables that are not due to the variables being correlated but instead are due to the measuring methods being similar. For example, if both measures are rating scales, rating errors will affect scores on both measures. Thus, if a rater tends to give high ratings, this tendency will cause the two measures to be correlated (the target would get high scores on both measures).

Item overlap: questions (items) may be the same or very similar (e.g., the parent might rate how outgoing the child is as part of the personality inventory and then might be asked how outgoing the child is as part of the “measure of behavior.” Not surprisingly, the parent’s answers to the same question will correlate. 

Eighth paragraph—

Pragmatic constraints: practical reasons (too difficult to do)

Final paragraph

“Judges’ diagnosticity ratings”: judges ratings of the degree that a behavior is characteristic  (typical, diagnostic) of a person with a given trait (e.g., the extent to which judges believe a behavior is typical of a neurotic).

Method

The Participants section is fairly clear.

Five-Factor  model of personality  subsection—

If you want a better idea of what the items on the NEO-FFI might look like, try the following URLs:

http://ipip.ori.org/ipip/newQform50b5.htm

http://www.outofservice.com/bigfive/

http://www.personalitytest.net/ipip/ipipneo300.htm

Riverside Behavioral Q-Sort subsection—

To see all the items in order, go to this URL:

http://www.faculty.ucr.edu/~funder/lab/CRBQitems.htm

In some ways, the Q-sort was not that different from rating each behavior (e.g., “offers advice”) on a 1 (extremely uncharacteristic of child) to 9 (extremely characteristic of child) scale. However, with the Q-sort, instead of rating each behavior on a 9-point scale, each behavior was put on a card, and each rater had to order the cards from extremely uncharacteristic of the child to extremely characteristic of the child and then put the cards into one of nine piles. The researchers had the raters put the items in piles to avoid the two rating errors described below.

First, when raters are forced to order the cards from least characteristic of the child to most characteristic,  raters are forced to compare their ratings of each behavior with their ratings of the other behaviors. Consequently, raters probably will not accidentally rate a behavior they think is less characteristic of a child higher than a behavior they think is more characteristic of the child. Had raters merely filled out rating scales, raters might make each rating without comparing the current rating with their previous ratings. Consequently, they might unintentionally give a behavior they saw as less typical of the child a higher rating than a behavior they saw as more typical of the child.

Second, the Q-sort can stop rating errors such as raters rating every behavior as extremely characteristic of the child or every behavior as a 5 (neither characteristic or uncharacteristic of the child). Instead, the researchers forced raters to have their ratings approximate a normal distribution. For example, the researchers could have insisted that one of the ratings be a “1,” three be 2’s, six be 3’s, twelve be 4’s, twenty be 5’s, twelve  be  6’s, six be 7’s, three be 8’s, and one be a “9.”

Last paragraph—

Prototypical: like the typical case; seeming like the perfect example.

Judges rated the extent to which each of 64 behaviors would be characteristic of (a) the neurotic, (b) the extravert, (c) the agreeable person, (d) the conscientious person, and (e) the person who would score high on the openness to experience factor. However, rather than using a rating scale, judges used a Q-sort (see above for the reasons for using a Q-sort). The Q-sort originally led to ratings from 1 (extremely uncharacteristic of such an individual ) to 9 (extremely characteristic of such an individual). The researchers mathematically transformed the judges’ ratings so that (a) 0 would indicate that the behavior was rated as neither uncharacteristic nor characteristic of the trait, (b) a negative score would indicate that the behavior was uncharacteristic of the trait,  (c) a positive score would indicate that the behavior was characteristic of the trait,  (d) 2/3 of the scores would be between –1 and +1, and (e) all scores would be between –3 and +3. In their study, the strongest predicted relationships were between conscientiousness and “displaying ambition” (2.36), between openness to experience and “shows a wide range of interests” (2.31), between openness to experience and “shows interest in intellectual or cognitive matters” (2.31), between extraversion and “is talkative” (2.11), between extraversion and “seems detached from the interaction” (-2.10), between neuroticism and “expresses insecurity” (2.04), and between neuroticism and “shows physical signs of tension or anxiety” (2.03).

Results

     “Effect size … 15”: in this case, the authors mean the absolute value of the correlation coefficient must be greater than or equal to .15. In other words, correlations between  -.14 and +.14 were not included. Usually, researchers use r² instead of r (see pages 164-167 of your text). Also note that most of these correlations are small (to understand how small the correlations between .15 and .20 and between -.15 and -.20 are, see pages 164-167). (The authors refer to the small size of correlations as reflecting “low fidelity.”) Note that, in Tables 1-5, whereas none of the effect sizes are large (the largest r is .36), many are statistically significant (all the correlations that have asterisks next to them are statistically significant [reliably different from a zero correlation] at the p <.05 level).  Thus, effect size is not the same as statistical significance.

      Note that the results, except for openness, are what you would have predicted.

Last paragraph on page 427 (continues to page 428)—

 The researchers had judges predict the degree to which a behavior would be related to a trait. Then, the researchers correlated the judges’ predictions with the actual correlation between the behavior and the trait. The researchers did this for all five traits and for all the behaviors studied. A correlation of 1.00 would mean that the judges perfectly predicted the extent to which the traits and behaviors were correlated; a correlation of 0 would mean that there was no correspondence between the judges’ predictions of how the traits and behaviors were correlated and the reality of how the traits and behaviors were correlated; and a correlation of –1.00 would indicate that the judges’ intuitions were completely opposite of reality.  Thus, the correlations in Table 6 give an indication of the extent to which the results agree with expert intuition.

Discussion

Manifested: expressed, made visible (in terms of behaviors; in terms of what children do)

“Delineate a taxonomy of four higher-order traits”: propose a classification system consisting of four general traits

“Concurrent and longitudinal outcomes”: what is currently happening to the child (e.g., child’s grade-point average at the moment) as well as future outcomes (e.g., how long the child will live).

“Strong situations”: structured situations that dictate how individuals should behave, and thus result in observing fewer individual differences in behavior (e.g., students in a first grade classroom all being told to put their heads on their desk). In contrast, weak situations are less structured (e.g., children at recess).