Professional Documents
Culture Documents
Prepared By: 106 Pranav Jain 167 Surbhi Koshal 151 Aayush Agarwal 506 Harsh Gupta
Terminology
Response variable Measured output value
Factors Input variables that can be changed Levels/Treatments Specific values of factors (inputs) Replication Completely re-run experiment with same input levels Used to determine impact of measurement error
Interaction Effect of one input factor depends on level of another input factor
ANOVA
Statistical technique used to investigate and model relationship between a response variable and one or more independent variables
Each independent variable is called as factor
Hypothesis
H0: Population means of each level are equal
H1: At least one of the level means are not all equal
N ni
i 1
Factorial Essentials
Notation System E.g. 2x3, 2 IV, one with 2 level, one with 3 and 6 total conditions
Factorial Matrix
2x2 presentation rate
2-sec/word 4-sec/word
type of training
Imagery Rote
Main Effects
Overall effect of an Independent Variable (IV) on a dependent variable
Main effect of training effect = 20-15 = 5 Imagery produces better results than rote by 5 words Main effect of presentation rate = 14.5-20.5 = -6 2-sec/word produces a worse result than 4-sec/word by -6
Interactions
Teaching Style
Lab
Science Humanity
80 70
Lectures
70 80
Interaction Effects
*No main effect is present in the above case Interactions Major at Lab 80 70 = -10 80 the size or direction of the simple main Major at Lects 70 80 = 10 75 IV1 changing various levels of IV2 Major at Science at80 70 = -10 Major at Humanity 70 80 = 10 70
65 Science Lab Lectures Humanity Mean 85
effect on DV of
Interaction effect
Interaction effect occurs when one factor effects the results differently depending upon the second factor
12
10
Intensity
0 A B
Exercise
Dr ug 2
Dr ug 1
Dr ug 3
Dr ug 4
Subtle Effect
30.0 25.0
20.0
Intensity
15.0
10.0
5.0
0.0 A B C D
Exercise
Dr ug 1 Dr ug 2 Dr ug 3 Dr ug 4
Dramatic Interaction
8.0 7.0 6.0
5.0
Intensity
4.0
3.0
2.0
1.0
0.0 A B
Exercise
Dr ug 2
Dr ug 1
Dr ug 3
Dr ug 4
Experiment
Experiment
Information on the earnings of men and women for different occupations was collected. Suppose that the aim of the experiment is to investigate whether there were any differences between the weekly salaries ($) of men and women employed as computer programmers, contractors and doctors.
A sample of 10 men and 10 women was selected from each of the three occupations, and the weekly salary for each individual in the sample was recorded.
Considering 0.05 level of significance, we need to test for any significant effect due to occupation, gender and interaction.
Terminology
Response variable Weekly Income ($) Factors A) Gender B) Occupation Levels A) Male | Female B) Comp Programmer | Contractor | Doctor Replication 60/(2x3) = 10 replications Interaction Dependency of effect of gender on different occupations or Dependency of effect of occupations on different gender
Hypothesis
Gender H0: Average salary for males and females are same. H1: Average salary for males and females are different. Occupation H0: Average salary for doctors, contractors and computer programmers is same. H1: Average salary for doctors, contractors and computer programmers is different. Interaction H0: There is no interaction effect taking place. H1: There is some interaction effect taking place.
Data Set
Between-Subjects Factors N
Comp Programmer Occupation Contractor Doctor Gender Female Male 20 20 20 30 30
Data Summary
Descriptive Statistics Dependent Variable: Weekly Salary Occupation Comp Programmer Gender Female Male Mean 741.30 796.00 Std. Deviation 78.942 86.340 N 10 10
Total
Female Contractor Male Total Female Doctor Male Total Female Total Male Total
768.65
634.70 979.40 807.05 930.90 1046.90 988.90 768.97 940.77 854.87
85.267
109.283 104.198 205.104 100.507 91.001 110.674 155.878 140.991 170.932
20
10 10 20 10 10 20 30 30 60
Analysis
Tests of Between-Subjects Effects Dependent Variable: Weekly Salary Source Corrected Model Intercept Occupation Gender Occupation * Gender Error Total Corrected Total Type III Sum of Squares 1230024.533
a
df 5 1 2 1
F 26.901
233602.300
493822.400 45571668.00 0 1723846.933
2
54 60 59
116801.150
9144.859
12.772
.000
.321
Crossing of 2 lines (Programmer and Contractor) indicate significant effect on wages due to gender in the occupations. This might be attributed to the job requirements of the occupation
Inverse movement of 2 lines (Programmer and Contractor) indicate significant effect on wages due to gender in the occupations. This might be attributed to the job requirements of the occupation
Conclusion
Gender Reject H0 -> Average salary for males and females are different Occupation Reject H0 -> Average salaries for doctors, contractors and computer programmers are different Interaction Reject H0 -> There is some interaction effect taking place between gender and occupation
Thank You