Return to Index page
  Research Weekly Handouts
EDIN 650 -- Research

Return to Weekly Handouts

EDIN 650 Research Methods
Dr. Robert Slotnick
 
 

Hand-out #4
     __________________________________________________

Based on Frankel and Wallen -- Variables and Hypotheses

Importance of Studying Relationships

A good res q can be descriptive, that is, collect data about a problem w/o studying relationships.  But, invariably, after you have information on a subject you want to know what else it is related to, what causes it, etc.

What is a Variable?

A variable is a concept that stands for variation w/i a class of objects
Any noun or concept can be a variable; for example boys, girls, height, weight, software, reinforcement can all be variables.  Variables can be qualitative, quantitative and constant.
If members of a class are identical (do not vary) they are called constants.
In a res study some characteristics will be variables and others ... constants.
In an exper. to test learning outcomes in 9th graders as a function of type of reinforcement -- praise, money, extra points; type of reinforcement is the variable, 9th graders are the constant.
What is the outcome?

In the experiment, using IMM to improve achievement in 10th grade biology students, what is the variable?  what is the constant? and what is the outcome?

Quantitative versus Qualitative Variables

Quantitative var. vary in degree along a continuum from less to more.  We can assign numbers to the degree different individuals possess the variable.  Height and weight are quantitative variables, but so are concepts which are measured along a scale, from 1 to 5 for example, so that individuals can be assessed on that variable.  Achievement is usually measured as a quantitative variable.  Anything that can be measured on a scale is a quantitative variable.

Categorical variables refer to different categories which describe the concept, not differences in amount.  Gender, religious preference, occupation, ethnicity, political party affiliation are examples of categorical variables.

Res will study the rel.ship bet. some combination of quantitative and qualitative variables.  Statistical tests have been designed to detect differences between different variables.  These tests are based on probabilities of  occurrence and the type of variable.  Therefore, different combinations of variables will require different kinds of statistical tests.  For the relationships below based on different variables, what is the best test?

Type of Variable                                                                     Statistical Test
Two quantitative variables                                                                                   Correlation
Age and interest in school
Reading ach and math ach
 

One categorical and one quant. var                                                                         t-test
Method used to teach reading and reading ach
Counseling approach and level of anx

Two Categorical Var                                                                                                Chi Square
Ethnicity and father's occupation
Gender of teacher and subject taught

See Syllabus notes and text, ch 5.  Selection of Measuring Instruments.

Assignment 3: Variables and Statistical Tests.
For the following variables and relationships select the appropriate statistical test.  Focus on the type of variable.  Identify variables,  categorize and select type of statistical test.

 Example                                                  Type of Variable                            Statistical Test
Gender and pol party voting preference

Amount of time watching TV and amt of agg beh

Score on test 1 and score on test 2

Nationality and verbal achievement

Comparison of teachers, administrators, and students on computer use

Comparison of boys and girls on math achievement

Administrative style and faculty morale

Religious pref and pol party membership

Identify the type of variable and the appropriate statistical test to use!!
 

Manipulated versus Outcome Variables
Select a variable for study, e.g., amount of reinforcement
Use different amounts of that variable
You have manipulated the experimental variable
Seek relationship between the manipulated variable and the outcome
Manipulated variable is called the Independent Variable and the outcome variable is called the Dependent Variable.

Independent versus Dependent Variables
Independent variable is the presumed or possible cause ------ of the ---- dependent variable, the presumed result.

Extraneous Variables
In an experiment there are many possible (independent variables) that could have an effect on the dependent variable.  Extran. var. are indep. var. that have not been controlled.  The aim is to anticipate all extran. var. and control them.  One way is to keep extran. var. constant.  In an exper. on the effect of amt. of reinf. on learning, age or prior knowledge is an extran. var.  One way to control them is to select only 9th grade Ss of a particular learning level.
 

Hypotheses

Difference between Research Questions and Hypotheses
Quest. is more general.  Hypoth. has commitment, directionality, and specificity.
Ex: How do teachers feel about special classes for the educationally handicapped? (Question)  Teachers believe that students attending special classes for the educationally handicapped are stigmatized? (Hy)  or the same quest. can generate a different hypothesis: Teachers believe that special classes for the educationally handicapped will help students improve their academic skills.
Hypotheses become a part of scientific body of knowledge.  Hy. lead to the study of relationships.

Analyze the following example: Hypothesis and its variables
"Among stu of the same age and intelligence, skill performance is directly related to the number of practice trials and more so for boys than girls"  (Interpret and analyze)
..
IV = # of trials
DV = skill performance
Moderator = gender
Control Variables = age, intelligence
Intervening Variable = Learning
 

Moderator var is a new term and refers to a special type of IV.  It is selected  because it is believed to affect the IV -- DV relationship.  Like an IV it can be measured, selected, or manipulated.  If an E suspects that the nature of rel. bet. X and Y is altered by the level of a 3rd variable Z, then Z is the moderator variable.
 

             
 
Practice writing hypotheses!!
 
 

General Experimental Learning Model
 

Causes           Relationship            Effects

IV

Moderator     Intervening Variables  Dependent Variable(s)
Control Variable(s)
 

Example 1:
Analyze and Graph: Exper. 1.  Compare the effectiveness of a visual approach and an auditory approach for teaching a unit on ecology.  If you suspect that type of learner will affect the results you have a moderator variable and it should be accounted for in the design and results.

What is the IV? DV? Moderator Var? Control Var?
Let us look at a graphical expression of the above w/o a moderator variable.


 


 
 

In this case the independent variable, method of presentation of ecology unit, has no effect on the dependent variable, amount of learning.  Therefore, it truly has no effect or there is another variable which is masking the real effect.  If we divide the Ss into two groups; one group which learns visually and one group which learns auditorily, we might see the effect of learning presentation and learning style.
 

On the following page graph the experiment with the addition of the moderator variable.
 
 
 
 


 
 
Method of Presentation of Ecology Unit
 

The point is that Visual Learners do better on the task when it is presented as a Visual Task and Auditory Learners do better on the same task when it is presented as an Auditory Task.  Type of learning style is a moderator variable which influences the results.  In this case it masks the difference between learning and achievement
 
 

Example 2.
Analyze the next example : What is the rel.ship bet. test performance and two types of test conditions (ego orientation and task orientation)?  You suspect that test anxiety may influence the results.  Outline the IV, DV, Moderator Var.

In this example, anxiety is the moderator variable.  Test anxiety interacts with ego and task orientation to influence the results.

Moderator var are a type of IV and if not selected, measured, or manipulated they can obscure the affects of the IV on the DV.  Genl. types of moderator variables are age, sex, IQ, social class, income, learning style, motivation-interest
 


 
 
 
 
 
 

Issues to discuss
Internet library search

Research plan

Hypothesis testing
 

Return to top of page