Copyright - Antonia D'Onofrio - 2001/2002/2003

**Introduction:** This web page outlines how the written assignments
might be structured and organized. These are the sample write-ups
that we completed last Thursday evening. These are impressionistic
but they are models for the type of essay that I would like to see you
write in response to the SPSS applications that we have been doing.

**Assignment #1 Correlation**

The purpose of this study was to analyze the correlations between undergraduates' scores in critical thinking and Miller's Analogy scores and the total score on the College Reading Test. The sample used for this study consisted of 40 juniors at a liberal arts centered university in an urban area. The rationale for this investigation was to determine whether measures of abstract reasoning predict reading for this cohort. Descriptive statistics were computed to describe the sample and to test assumptions of normality in the distributions of the three measures.

Table 1 reports the descriptive statistics.

**Table 1. Descriptive statistics for Critical Thinking, Miller's Analogies
and the College Reading Test (N=40)**

__________________________________________________________________________________________

Test | Critical Thinking | Miller's Analogies | College Reading Test |

Statistics | |||

Mean | |||

Mode | |||

Median | |||

Min | |||

Max | |||

Standard deviation |

The assumption of normality was not met in the data because each distribution indicated that the mean, mode and median were not centrally located. Rather the distributions tended to be negatively skewed. Based on an inspection standard deviation values it was clear that the greatest amount variability could be found in the Millers' Analogy Test Scores (s = 24.9). In addition, the mean score for Critical Thinking (12.34) showed that this was the more difficult test and conversely the College Test (19.78) was the easiest based its mean value.

A Pearson Product Moment Correlation Coefficient was calculated for each pair of test scores. The results of this analysis are reported in Table 2.

**Table 2. Intercorrelations of Critical thinking, Miller's Analogy
and College Reading scores (N=40)**

Critical thinking | Miller's Analogy | College Reading | |

Critical thinking | - | .78* | .88** |

Miller's Analogy | - | - | .79* |

College Reading | - | - | - |

* p<05

** p<01

The matrix of correlations shows that the greatest association was between Critical Thinking and College Reading. All the correlations between scores were statistically significant. Thus, abstract thinking was a good predictor of college reading for this sample.

The practical implications of this research can be summarized as follows. Based on these findings university academic support services might test critical thinking earlier in an undergraduate's time in college in order to provide reasonable support services aimed at raising abstract thinking abilities leading to better college readers, and ultimately better graduate students.

-----------------------------------------------------------------------------------------

Assignment #2 - Multiple regression

The purpose of this study was to determine the best weighted linear combination of predictors for College Reading Total Test Scores. A sample of 40 undergraduate students were tested in their junior year using a battery of reading, performance and reasoning measures. The predictors included the following tests: Critical thinking, Word Recognition, grade in the Advanced Placement Seminar. The rationale for this study was as follows. First, good college readers are also likely to be successful critical thinking which involves both inductive and deductive reasoning. Second, they are also likely to do well in word recognition tasks. Finally, they are probably going to achieve high grades in academic work.

Include a paragraph and Table 1 to review the descriptive statistics.

Include a paragraph and Table 2 to report on the correlations between
all variables.

Multiple regression was used to determine which of
the three measures were the best predictors of College Reading. A
stepwise multiple regression analysis was performed, yielding the following
linear equation:

A multiple R (89) was calculated (p<.001). This correlation was associated with a coefficient of determination RCollege Reading Test = A + b (Advanced Placement grade) + b(Critical Thinking) + b(Word Recognition) + residual.

**Don't forget to include statistical information from the McNemar's
analysis, Mean Squares for Explained and Residual variation and the F ratio
with its probability of error.**

**Table 3. B weights for predictors in the equation**

Advance Placement grade

Critical Thinking

Word Recognition