R e s e a r c h

What Is Research

• Dictionary Def “A search for knowledge”

• LAN: “Research is a systematic, empirical and controlled inquirer into a subject. The purpose is to identify the relevant phenomenon, specialized relationships between the phenomena “

• Inquires that are systematic, empirical, controlled

• ID Phenomena

• ID relationships

• Use results to rule out alternate factors

Why do people do research?

In the search for the truth, research is the scientific way to weed out opinion and bias to find what we can prove. You do research to say Rush says “we have more trees in America than when Columbus discovered America” so you research the issue, what is a rational way of estimating number of trees in the past, what is the unit of measure lets projects some measurements. Lets compare that number to present day estimates, make sure your estimates are valid and comparable. Then compare the numbers. Then numbers say yes there are more trees or no there are not more trees. Then we have truthful answers based on facts and proper scientific inquiry
 

Traditional scientific model

• LAN: Set of Interrelated Constructions, definitions, Propositions, That Present a systematic view of an event in order to explain that event and or predict a future event

• Theory is a way of summarizing out history in the interest of predicting the future.

• Theories must be scientific, testable and valid

• Not Scientific if not systematic

• BABBIE: Theories are a systematic set of interrelated statements intended to explain some aspect of an event
   

HYPOTHESES:
 

• BABBIE: A statement predicting that a relationship exist between an independent variable and a dependent variable

• Must be verifiable

• Include a causal relationship (independent variable=current policy, dependent variable = change in policy)

UNIT OF ANALYSIS

• Unit: an individual, a city, a car

• Variable or property of that unit of measure : age weight eyes, : population, size sq miles : make, model year

OPERATIONALIZE:

• Saying exactly what your variables are.
  Defining your terms

• Figuring out the details that must decided for your studies to operate

• Making it crystal clear what is being studied

OBSERVATION:

• Collection of the data, observe, survey Ect
  Validity check

• To prove that the measure accurately reflects the concept you are trying to measure Example using speaking ability to validate presidential leadership ability

Reliability

• That the date collection is consistent and repeatable and not contaminate

Research and Theory.

• Theory is a guess and research proves the guess LOOK THIS UP

• Not all theories are theories: Must be a System peer pressure is a variable belongs to the system of adolescent behavior To say that continental drift effects adolescent behavior in Spain would be wrong because the one is not a variable in the system .

• Not all theories are scientific: Must be predictable and behavioral Astrology is based on a belif in tradition that is un provable

• Not all scientific theories are infallible; they must apply in every instance. Any theory that only applies sometimes is incomplete or false. To say that man always acts in his own self-interest does not address someone falling on a hand grenade.

INTERNAL VALIDITY:

• Does the causal relationship exist?
  Independent variable vs dependent validity

• Example pre and posttest on kids verbal skills after finishing a year head start class

Threats to validity

• History maybe 9/11 happened between pre and post test \\\ new teletubbies teaches verbal skills

• Maturation \\\kids natural increases in verbal skills

• Its natural tendency to change, not stay static must control for maturation or natural change, you can control by adding a control group and more testing points

• Selection two groups very different need to control for difference

• Best way to control is by randomization

• Testing maybe the change found in the research came about from the test only

• Do a control post-test with no pre- test to control for the effects of testing

• 4 groups, control, with and without pre test experimental with and without pretest to control for the testing

• just the act of testing increases skills

• Instrumentation, the change of testing instrumentation /// the program add new software

• Experimental mortality /// what about kids that drop out, mabe just good kids because their parents are satisfied

• Statistical regression the more measurement happens, the more likely scores will group together //iff bobby gets a 0, chances are he will do better

EXTERNAL VALIDITY
Who you sample

Construction validity is it put together in a valid form
1. Definition of your concept, construct or measurement.
2. Close to measurement validity but asking precise definition of construct
3. Inadequate definition of the construct
4. Pre-operational
5. Implementation threat
6. Overcome monar* operation and method bias( *not sure what this is )
7. Expectation bias subject answer they way they think you want to hear

Confounding construction

• Confounding levels of constructs apples and oranges
  im, hungry, lets grow some food , need to do research before it is needed ??? a good scholar does research ahead of time

• Statistical conclusion validity
  What is you level of conference in your statistics, 90%, 95% 99

• Measurement validity
  Stability of measurement, is it repeatable

• Generalization, will data derived from college students apply to rual adults
   

True experimental

• Uses random selection for the participants to hold validity Randomization

• This means they are equal haves and controls selection bias and regression effects // for the head start you have a variety of children with parent ages, background education and income all represented. So you cant say one factor effected results (selection Bias)

• Randomly split the group into a Control group and a Experimental group

• The control group controls for history, maturation and instrument

• Give a pretest to see where both groups are as a group

• Apply the treatment /// in this case head start

• Give a post test to determent changes

• Issues of testing bias and the hawthorn effect

Issues of the ethics of a control group, //how come my child is left out, everything is voluntary, reasons why some do and some do not volunteer (we are looking at child abuse at home, do you want to volunteer) the rotating ticket books is a good example of getting a random sample of offenders with different tickets ,

Does something have an effect or causation on something else, and how do you determent that ?

1. Constant conjecture: the dependent variable is related
2. Temporary Precedence: cause happens before the effect
3. Elimination of alternative explanations

You must be able to explain away all alternative explanations
To prove causation you must be able to isolate alternant variables

Deductive: moves from the general to the specific more powerful

• All men are mortal--- major premise
• Aristotle is a man --- minor premise
• Aristotle is mortal --- conclusion

Inductive: move from the particular to the general exploratory but weak

Aristotle is a mortal
Aristotle is a man
Plato is mortal
Plato is a man
All men are mortal

• hypothesis A statement predicting that a relationship exist between an independent variable and a dependent variable
• independent variable the static non changing variable like a policy
• control variable on that remains unchanged
• internal validity is the question or hypothesis correct
• external validity did you implement the research correctly
• statistical conclusion validity I think it means do the conclusion agree with statistical results
• construct validity is it put together in a way that cross checks and jive with previous research
predictive validity do the results agree with other facts , or is it opposite of what we see
• face validity does you research method seem reasonable on the face
• content validity does you hypothisis measure all elements of the topic you are measuring
• reliability That the date collection is consistent and repeatable and not contaminated

• Reliability verse validity; a bathroom scale that is 10 pounds over, it weight is always reliable, it shows you if you have gained or loss but its numbers are not valid
• Experimental set; group that has treatment
• Control set; group that has no treatment
  Example of controlling for testing
  Head start sample

Thomas Kuhn's theory on scientific revolution

Paradigm shift : A shift in the way people look/solve a problem
Paradigms shift come from looking from outside the box to explain anomalies

Normal Science: based on life inside the parading

Anomaly: a anomaly that cannot be explained by the normal theory/paradigm

• Ignore the anomaly
• Fit it into the paradigm
• Think outside the paradigm

Revolutionary science a discovery that realigns or shifts the paradig

Qualitative conclusions have no bases in numbers

• Ethnography
• Interview
• Participative
• Action studies

Quantitative with numbers

• Randomization quasi experiments
• Objective test
• Based on numbers or mental judgment

Quantitative numbers Qualitative no numbers
Logical
Deductive
Measurement
Data collation
Hypothesis
Dynamic theories
Try to separate emotion Emotion is part of the process
Strive to be objective Don’t worry about being objective
Less flexible
Must be technical expert
Lest vivid
No comprehensive observations Less cost
Easy to do
Less affected by variables
Must deal with people
More policy related
Less need to id subjects