Variables, Hypotheses and Classification of Measurements

1
Variables, Hypotheses and Classification of
Measurements
2
What Factors Determine How Fast a Car can go?

• How much gas you give it
• Whether of not the break is engaged
• Number of cylinders in the engine
• The terrain over which one is driving
• The absence or presence of a wind
• Type of tires on the car and their inflation
  level
• Performance of the fuel injection system or
  sparkplugs

3
Variables

• In research we call the factors that can
  influence something else (such as the speed of
  the car) a variable.
• But more specifically, these are called
  Independent Variables.

4
What is an Independent Variable?

• A variable that precedes, influences or predicts
  the outcome of the study.
• The researcher manipulates this variable, or in
  the case of preexisting variables that cannot be
  manipulated (sex, race, etc), selects the
  variable for inclusion in the study.

5
How to Remember

• An Independent Variable starts with the letter I.
• So, I, the researcher can choose which variables
  to manipulate or include in the study.
• I have control of the selection or inclusion of
  the Independent variables.
• I I

6
Dependent Variable

• In the car example, we are interested in
  measuring the speed of the car.
• The thing that we are measuring (speed) is called
  the dependent variable.

7
What is a Dependent Variable?

• The variable that the researcher measures it is
  influenced by the independent variable.
• Sometimes called the outcome variable--measures
  the effect of whatever experiment you've
  performed. For example, if you're studying the
  effects of fertilizer on plant growth, the
  dependent variable might be measured as the
  height of the plant. (The amount of fertilizer
  you gave would be the independent variable.)

8
Independent Dependent Variable Examples

• Independent Variable
• Tire Pressure
• Time spent studying
• Family Income
• Salary
• SAT scores
• Fertilizer Amount
• Age
• Hours employed

• Dependent Variable
• Speed of the car
• Grade on a test
• 4-H Congress Attendance
• Job Satisfaction
• College Admission
• Crop Yield
• Church Attendance
• GPA

9
In Practice

• Typically researchers will study one dependent
  variable per study but they might look at several
  different independent variables
• Is job satisfaction of extension agents
  influenced by salary, hours worked, facilities or
  leadership style of the County Director?

10
In Practice

• The researcher determines what he/she
• Wants to measure the dependent variable
• Wants to vary or include in the study the
  independent variable(s)
• Often, the title of a research study will contain
  the dependent and independent variables
• The Effect of Computer Aided Instruction on Math
  Achievement of Algebra I students.
• Math Achievement is the Dependent Variable
• Type of Instruction is the Independent Variable

11
A Word of Caution

• In identifying independent variables students
  often get confused with distinguishing between
  the independent variable and the levels of the
  independent variable.
• A student may claim there are three independent
  variables, when there is only one independent
  variable with three levels
• See the next slide

12
Levels of Independent Variables

• A teacher is doing a study of Instructional
  Methods.
• One class was taught only with lecture and no
  visuals
• Another class was taught using the book and
  worksheets
• A third class was taught only using PowerPoint
• In this example there is only one independent
  variable Instructional Methods but there are
  three levels of that variable.

13
One more example

• A researcher wanted to identify how to increase
  response rates to surveys.
• 100 people received a survey from him with no
  incentive to respond
• Another 100 received a survey with a dollar bill
  inside
• Another 100 received a survey with instructions
  that all respondents would be entered in a
  drawing for 100
• Another 100 received a survey with a coupon for a
  free Chick-fil-a sandwich inside.
• The research compared the response rates

See the questions about this study on the next
slide!
14
Questions about the Last Slide

• Is there one or more independent variables?
• There is only one independent variable
• What is the independent variable?
• Type of incentive received
• How many levels of the independent variable are
  there?
• Four (none, 1, money drawing, food coupon)
• What is the dependent variable?
• Response rate

15
What are Confounding Variables?

• A variable, other than those being explored in a
  given study, that affects the research outcomes.
• There are three types
• Intervening other variables such as motivation
  or intelligence are at play (our text calls this
  a moderator variable)
• Organismic physical traits such as poor
  eyesight or hearing are at play
• Extraneous Fatigue, room temperature,
  distractions, etc. are at play

16
Measurement Scales

• In research we collect data.
• These data can be classified into 4 categories.
• Nominal
• Ordinal
• Interval
• Ratio
• The type of data we have dictates what statistics
  are used. More about that later

Note The word data is plural. Never write that
data is, it is data are
17
What is Nominal Data?

• Categorical data where the order of the
  categories is arbitrary and has no true meaning
  other than classification
• Social Security number
• Drivers License number
• Male or Female
• Ethnicity
• Political Party Affiliation
• Colors of marbles
• The text identifies this as Categorical Data

18
Nominal Data

• The data we collect often has to be converted to
  numbers for statistical or tabulation purposes.
• So when we have nominal categorical data we often
  arbitrarily assign a numerical value for
  tabulation purposes such as
• Male 1 and Female 2
• Democrats 1, Republicans 2 and Other 3
• Green 1, Blue 2, Red 3

19
Time to Think

• See if you can come up with 5 examples of nominal
  data!

20
What is Ordinal Data?

• A scale that expresses data as rankings, rather
  than scores
• A, B, C, D, F
• Socio Economic Status
• Professorial ranks
• The distance between the categories is not equal
  the difference between a F and D is probably not
  the same as the difference between an A and a B.
• The text classifies ordinal data as categorical
  data.

21
Recording Ordinal Data

• After we collect data from a survey instrument,
  we then have to transfer the data to an Excel
  spreadsheet, Statistical Package for the Social
  Sciences (SPSS) or some other software program
  for tabulation purposes.
• The data have to be converted to numbers. So we
  might code
• Full Professor 3
• Associate Professor 2
• Assistant Professor 1

22
Fudging

• Many researchers treat ordinal data as they would
  interval or ratio data.
• That is they convert it to a numerical value with
  meaning
• For example they may assign an A a value of 4, a
  B a value of 3, etc. and then calculate a GPA for
  use in statistical analyses.
• Basically they are taking ordered data and
  turning it into interval data (more about
  interval data in a couple of slides)

23
Time to Think

• See if you can come up with 5 examples of ordinal
  data!

24
What is Interval Data?

• Measurement scales expressed in equal number
  units, but not having a true zero point
• IQ score (there is no such thing as zero
  intelligence)
• Scores on the Meyers-Briggs Scale (there is no
  zero)
• Temperature (there is a zero but it has meaning,
  it does not represent nothingness)
• Belt sizes
• The text calls this quantitative data
• When we record this data, we use the actual
  numbers

25
Time to Think

• See if you can come up with 2 examples of
  interval data!
• Why only 2 examples?
• It is hard to find interval data

26
What is Ratio Data?

• Measurement scales expressed in equal number
  units, but having a true zero point
• Test scores
• Salary
• Weight
• Distance
• For all practical purposes Interval data and
  Ratio data are treated the same statistically
• The text calls this quantitative data
• When we record this data, we use the actual
  numbers

27
Time to Think

• See if you can come up with 5 examples of ratio
  data!

28
What is a Hypothesis?

• A testable statement of a predicted relationship
  or difference among selected variables.

29
Why Use Hypotheses in Research?

• Forces us to think more deeply
• Forces us to read the related literature
• Clarifies our thinking
• Makes our research more scientific

30
Three Types of Hypotheses

• Directional
• Nondirectional
• Null

31
Directional Hypothesis

• A Directional Hypothesis is preferred in
  research.
• Basically is says if A happens then B will
  happen.
• If one pushes down on the accelerator pedal, the
  car will go faster.

32
Nondirectional Hypothesis

• A nondirectional hypothesis can be used if you
  think something will happen but your are not very
  certain what will happen.
• If students are taught a unit on
  multiculturalism, their attitude toward other
  cultures will change
• You dont know if the students will end up being
  more positive or less positive to other cultures
  as a result of the teaching unit.

33
Research Hypothesis

• Both nondirectional or directional hypotheses
  are often referred to as research hypotheses.

34
Null Hypothesis?

• A hypothesis that states that no differences or
  relationship exists among specified variables.
• Also called the statistical hypothesis.
• This is only used for statistical purposes.
• If a null hypothesis is found in a research
  study, it nearly always follows the research
  hypothesis.
• In the old days of research, researchers would
  nearly always state the null hypothesis.
• Today it is sort of assumed that one has tested
  the null hypothesis.

35
Hypothesis Shorthand

• In some texts you will see various notations (I
  call it shorthand) that is sometimes associated
  with hypotheses.
• Ha or H1 is the research hypothesis
• H0 or Ho is the null hypothesis

36
Characteristics of a Good Research Hypothesis

• Declarative sentence
• Is brief and clearly stated
• Identifies at least two variables (independent
  dependent)
• States an expected (predicted) relationship
  between at least one variable and at least one
  other variable
• States the nature of the relationship
• States the direction of the relationship
• Implies that the predicted relationship can be
  tested empirically
• Consistent with known facts, prior research, or
  theory

37
Bad Research Hypotheses

• The more supportive of political authorities a
  child is, the less likely the child will be to
  engage in political dissent as an adult
• Not testable
• Age determines whether a person is tolerant of
  social protest
• Nondirectional
• In comparing individuals, some people are more
  likely to donate money to political candidates
  than other people.
• No relationship is identified

38
Better Research Hypotheses

• If a prisoner is taught an in-demand skill
  during incarceration, he or she will be less
  likely to commit burglary upon release when
  compared to a similar prisoner without such
  training
• If elderly adults are taught to use the Internet
  (and provided free access to the necessary
  equipment and software), their feelings of
  loneliness will decrease

39
A Good Research Hypothesis

• Students in AEE 578 who can distinguish between
  independent and dependent variables will score
  higher on worksheet 3 than students who cannot
  distinguish between independent and dependent
  variables.

40
Time to Think

• See if you can write a good research hypothesis