Fall 2014 - Syllabus
PSYC 3090 (Section 200) Introduction to Experimental Psychology
Class Times: MWF, 9:05 a.m. 9:55 a.m.
Class Location: 120 Brackett Hall
Professor: Patrick J. Rosopa, Ph.D.
Office: 410C Brackett Hall
E-mail: [email protected]
Sample - Exam 1 Questions
1. Age in Years is a(n) _ variable.
2. Which index of central tendency is most influenced by extreme scores?
a. Standard deviation
Psych 309 Notes
Friday October 19th
Comparing More Than 2 Groups
No exercise n=10
c=number of comparisons, or how many t tests would you have to conduct
1-(.95)^3 = .14
Correction made to t
o A researcher believes that reaction time to a red light is faster than reaction time
to a green light. Ten participants are measured for their reaction time in
milliseconds to a red light (Y) and a green
For size of an effect when dealing with hypotheses about means, use Cohens d.
An estimate of Cohens d provided by Hedges g:
Expresses mean differences (numerator) in standard deviation units (denominato
Example 1 (cont.)
D = 153.6 =10.94
Also could be calculated by taking the mean of x, taking the mean of y, and finding the
X Y = D
Sample standard deviation
D i D
D D(hyp) 10.940
Characteristics of the t Distribution
Not a single distribution, but a family of distributions. Each differs in its approximation
to the standard normal curve (z)
Exact shape of a t distribution depends on the sample size, or mo
Example 3 Continued
= .076 -> only need this one generally. Just depends on how
the researcher reports it.
o The critical t is entered into the point biserial
Conclusion: Reject H0. The population m
Sociability score (1 = low to 5 = high) of boys and girls from the same family. Below, the
sociability scores for a boy and a girl from each of six families are listed.
Problem #10 in Chapter 14
o Albert Bandura has conducted a number of studies on aggression in children. In
one study (Bandura, Ross, and Ross, 1963), one group of nursery school children
were shown a film o
Population (, )
The values of these statistics will vary from sample to sample (X-Bar, Sx)
The statistics we compute (e.g., mean, difference between two means, corre
Sampling Distribution of the Mean
The distribution of sample means is called a sampling distribution of the mean. We can
construct a histogram of this distribution of means
Statistical simulation in R
o Some interesting characte
Two Independent Groups
Instead of taking one sample and calculating a sample mean, we have two independent
o Elements in one sample in no way affects the selection of elements in the other
Background: A Health Director at a local university believes that students at her campus
are very health conscious and, as a result, tend to consume less sugar than do most people
living in the U.S. The director knows t
Confidence Interval for
D t crit s D
Example: Text Messaging (Chapter 15)
Does text massaging affect driving?
Research study using a driving simulator at Clemson. n = 14
Each participant drove 55 mph u
Row Sum of Squares (SSR)
X R X 2
SS R =
n R=Sample a Particular Row
o Each row gets calculated
This gives us two values
Column Sum of Squares (SSC)
X C X
Factorial ANOVA: Two-Way ANOVA
In chapter 18, we had one IV (with k levels (Independent Variable)
o Exposure to aggression (k = 3)
o Drug/therapy treatment (k = 4)
In chapter 19, we have two IVs where the level
Each of 16 human resource (HR) managers review the performance appraisal of one
The performance appraisal that each HR manager reviews is actually the same except that
the gender of the employee is
Effect Size in ANOVA
o Measure of association between the IV and DV
o Provides the proportion of the total variance due to the IV (Eq. 18.13)
S S bet
o 0.01 = small
o 0.06 = medium
o 0.14 = large
Sampling Distribution of r
Sampling distribution of r has an expected value (mean) equal to (rho)
The sampling distribution of r has a standard deviation approximately equal to
Testing Hypothesis that = 0
H0: = 0
o Ratio of two variance estimates
o This ratio is always positive
F= bet =
o Because effect would be equal to 0, all s are equal
o When H0 is true, the F ratio is approximately 1
Problem with Multiple t Tests
Many statistical tests
o If k = 3, there are 3 pairwise comparisons
o If k = 7, there are 21 pairwise comparisons
Assuming all the tests are independent and that each is conducted at (= 0.05), the
Example (Section 18.7)
Calculate SS bet
ni = sample size within group
X= Group Mean
X= Grand Mean
SS bet = 4(17-
Hypothesis tests result in one of two possible outcomes: reject H0 or fail to reject H0.
Instead of dichotomous decision-making, we can construct confidence intervals.
A confidence interval (CI) is a way t
General Steps for Testing Hypotheses
1. Specify H0 and HA about a parameter of the population
2. Random sample drawn from population and the value of the sample statistic is calculated
3. Examine random sampling distribution of th
Assume a normally distributed population with = 70 and = 20. Your sample size is 25.
What is the probability of obtaining a sample with a mean that differs from the population
mean by 10 points or more?
X x 8070
With 1 sample, we converted our sample mean to a t
o df = n 1
With 2 independent samples, we convert the mean difference to a t
o df = Total Number of Scores 2
Is the reading comprehension of dyslex
Two-way ANOVA Example (Chapter 19)
Modified version of a study by Heilman and Chen (2005) published in the Journal of Applied Psychology.
Title: Same behavior, different consequences: Reactions to mens and womens altruistic citizenship behavior.
Sample - Exam 3 Review
1. The shape of the sampling distribution of r
a. is approximately normal.
b. is positively skewed.
c. is negatively skewed.
d. depends on the value of .
2. You calculated the correlation (r = -0.42)