GR615 Elements of Research

1
GR-615Elements of Research

• Edward M. Levinson, Ed.D
• Professor of Educational and School Psychology
• 242 Stouffer Hall
• emlevins_at_iup.edu
• http//www.coe.iup.edu/emlevins

2

• Scientific Research - The systematic, controlled,
  empirical and critical investigation of the
  hypothetical propositions about the presumed
  relations among natural phenomena

3
Important Characteristics

• Systematic logical, rational, orderly, not
  random
• Controlled - attempts to eliminate bias and
  error is objective rather than subjective
• Empirical - information and data based
• Critical - allows for review, rebuttal, and close
  scrutiny via replication

4
Purposes of Research

• To study relationships among variables
• To explain variance

5
Research vs. Commonsense

• Many think that psychology is commonsense and
  they use commonsense to solve problems and to
  make important decisions

6
Test Your Commonsense

• Which approach would you use to ask for a salary
  increase and why?
• Straight-forward approach
• Gamesmanship

7
Research vs. Commonsense

• What makes sense to one person does not make
  sense to another....
• Commonsense is unreliable, subjective, and biased
• Research is reliable and objective

8
Go with the odds
9
Research vs. Commonsense

• To the maximum extent possible, rely on research
  rather than commonsense when making decisions and
  solving problems

10
Those who fall in love with practice without
science are like a sailor without a helm or a
compass, and who never can be certain wither he
is going- Leonardo Da Vinci

11
Teaching methods based on research in the
cognitive sciences are the educational
equivalents of vaccine and penicillin. Yet few
outside of the educational research community are
aware of these breakthroughs or understand the
research that makes them possibleBruer, J.T.
(1993). The minds journey from novice to expert.
American Educator 17 (2), 6-15
12
Types of Research
Nomothetic vs. Idiographic
Experimental vs. Correlational
13
Nomothetic Research

• Large, Diverse Groups
• Varied Settings
• Results are Widely Generalizable
• Yields general laws/truths

14
Idiographic Research

• Small, singular groups
• Homogenous settings
• Results are person or setting specific, not
  generalizable

15
Research Continuum
Nomothetic
Idiographic
16
Experimental Research

• Compares two or more groups of individuals
• Assigns individuals to groups in a random manner
• Allows researcher to infer cause-effect

17
Quasi-Experimental Research

• Compares two or more groups of individuals
• Attempts to infer cause and effect but
  limitations often prevent researcher from doing
  so
• Uses intact groups rather than random assignment
  of participants to groups

18
Correlational Research

• Compares groups of individuals on one or more
  variables
• Establishes relationships between/among
  variables
• Does not allow cause-effect inferences

19
Correlation Coefficients

• Range from -1.00 to 1.00
• Plus/Minus sign indicates the nature of the
  relationship
• Positive correlations mean high scores are
  associated with high scores and low scores are
  associated with low scores

20
Correlation Coefficients

• Negative correlations mean high scores are
  associated with low scores and vice-versa
• Absolute value of the coefficient indicates the
  strength or magnitude of the relationship
• So, -.92 gt.65

21
(No Transcript)
22
Descriptive Research

• Typically uses surveys or interviews
• Is simply designed to describe current status
• Used extensively by developmental psychologists

23
Action Research

• A form of descriptive research
• Conducted by educators in their own classrooms or
  schools
• Often lacks objectivity because of dual role
  (teacher and researcher)

24
Steps in the Research Process

• Selecting the Problem
• Accumulating pertinent knowledge and information
• Latent period (reasoning)
• Idea or hypothesis formation
• Designing the test of the hypothesis
• Critical analysis and evaluation of observations
  made
• Rejection or acceptance of the original hypothesis

25
Consolidated Listing of Steps

• Definition and development of the problem
  including a survey of the related literature and
  formulation of the working hypothesis
• Selection or creation of appropriate data
  gathering techniques and actual collection of
  data
• Classification and analysis of data
• Conclusions, generalizations, and applications

26

• How are the steps in the research process
  embodied within the different sections of a
  journal article, or chapters in a dissertation?

27
SELECTING A PROBLEM AND CONSTRUCTING HYPOTHESES
28
Characteristics of a problem

• 1. It should ask about a relationship between
  two or more variables.
• 2. It should be stated clearly and
  unambiguously, usually in question form.

29

• 3. It should be possible to collect data to
  answer the question(s) asked.
• 4. It should not represent a moral or ethical
  position.

30

• 5. It should be both practical and interesting.
• After selecting and stating the problem, a
  hypothesis (or hypotheses should be formulated.

31

• The hypothesis should
• Be a suggested answer to the problem.
• Be a conjecture upon a relationship between two
  or more variables.
• Be stated clearly and unambiguously in the form
  of a declarative sentence.
• Be testable, that is, it should be possible to
  restate it in an operational form.

32

• A hypothesis can be defined as an expectation
  about events based on generalizations of the
  assumed relationship between variables. . .

33

• . . . Such expectations are developed via a
  review of research in the problem area, and via
  use of inductive and deductive reasoning.

34

• Hypotheses are constructed on two levels
• 1. Operational level Events are defined in
  observable terms in order to operate with the
  reality necessary to do research. (Statistical
  reasoning is emphasized).

35

• 2. Conceptual level Events are defined in terms
  of underlying community (usually causal) with
  other events. (Logical/verbal reasoning is
  emphasized).

36
Types of Hypotheses

• Null
• Experimental
• Directional
• Non-Directional

37
Types of Variables

• Independent Measured or manipulated by
  researcher to determine its effect on a
  phenomenon presumed cause
• Dependent Observed and measured to determine the
  effect of the independent variable presumed
  effect

38
Types of Variables

• Moderator Measured to determine whether it
  modifies or changes the relationship between the
  independent and dependent variables
• Control Those factors controlled in order to
  cancel or neutralize any effect they might have
  on the DV

39
Types of Variables

• Intervening A factor that theoretically may
  affect the DV but which cannot be measured,
  manipulated or observed

40
Categories of Variables

• Active - variables which can be manipulated
• Attribute - variables which cannot be
  manipulated human characteristics
• Continuous - variables capable of taking on an
  ordered set of values within a certain range
• Categorical - variables consisting of specific
  categories that cannot be ordered

41
OPERATIONAL DEFINITIONS

• Operational definitions have the purpose of
  assigning meaning to a construct or variable by
  specifying the activities or operations
  necessary to measure it. It gives meaning to a
  variable by spelling out what the investigator
  must do to measure it. (Kerlinger, 1973)

42

• Kerlinger (1973) defines two types of operational
  definitions
• Measured A measured operational definitions
  describes how a variable will be measured.
• Experimental An experimental operational
  definition spells out the details (operations) of
  the investigators manipulation of a variable.

43
Benefits of Op. Defs.

• Increases accuracy in measurement
• Allows for replicability
• Allows discrepancies in research results to be
  explained

44

• SUMMARY OF CHARACTERISTICS AND EXAMPLES OF
  MEASUREMENT SCALES
• It is important to understand scales of
  measurement in order to determine appropriate use
  of statistical procedures
• Scale of measurement is influenced by how a
  variable is operationally defined and how it is
  coded by the researcher

45
NOMINAL

• Characteristics
• Objects are classified and classes are denoted by
  numbers. That the number for one class is
  greater or less than another number reflects
  nothing about the properties of the objects other
  that that they are different

46
Nominal Scale

• Different numbers simply reflect different things
• Numbers reflect qualitative but not quantitative
  differences

47

• Examples
• Racial origin, eye color, numbers on football
  jerseys, sex, clinical diagnoses, automobile
  license numbers, social security numbers.

48
ORDINAL

• Characteristics
• The relative sizes of the numbers assigned to the
  objects reflect the amounts of the attribute the
  objects possess. Equal differences between the
  numbers do not imply equal differences in the
  amounts of the attributes.

49
Ordinal Scale

• Different numbers simply reflect different things
• Numbers reflect qualitative and quantitative
  differences
• The things that are measured can be ordered along
  a continuum

50

• Examples
• Hardness of mineral, grades for achievement,
  ranking on a personality trait, military ranks.

51
INTERVAL

• Characteristics
• A unit of measurement exists by which the objects
  not only can be ordered but may also be assigned
  numbers so that equal differences between the
  numbers assigned to objects reflect equal
  differences in the amounts of the attribute
  measured. The zero point of the interval scale
  is arbitrary and does not reflect absence of the
  attribute.

52
Interval Scale

• Different numbers simply reflect different things
• Numbers reflect qualitative and quantitative
  differences
• The things that are measured can be ordered along
  a continuum
• Intervals between adjacent points on the scale
  are of equal value numbers now also reflect how
  much more or how much less

53

• Examples
• Calendar, time, Fahrenheit and centigrade
  temperature scales

54
RATIO

• Characteristics
• The numbers assigned to objects have all the
  properties of those of the interval scale, and in
  addition an absolute zero point exists on the
  scale. A measurement of 0 indicates absence of
  the property measured. Ratios of the numbers
  assigned in measurement reflect ratios in amounts
  of the property measured

55
Interval Scale

• Different numbers simply reflect different things
• Numbers reflect qualitative and quantitative
  differences
• The things that are measured can be ordered along
  a continuum
• Intervals between adjacent points on the scale
  are of equal value numbers now also reflect how
  much more or how much less
• There is an absolute zero point a score of zero
  implies the absence of a trait

56

• Examples
• Height, weight, numerosity, time, temperature on
  the Kelvin (absolute zero) scale.

57

• PROBABILITY,
• RANDOMNESS,
• AND
• SAMPLING

58

• PROBABILITY
• Probability is the ratio of the number of times
  an event occurs to the total number of trials. .
  .

59

• With this definition, one approaches probability
  empirically by performing a series of tests,
  counting the number of times a certain kind of
  event happens, and then calculating the ratio. .
  . .

60

• The result of the calculation is the probability
  of the certain kind of event.

61

• Probability ratios range from 0 to 1, and are
  always positive.

62

• POPULATION
• All the subjects or objects which are of
  interest in research.
• (Asher, 1976)

63

• SAMPLE
• A portion of the whole group (population)
  usually chosen in such a way so as to be
  representative of the whole group (population).
• (Asher, 1976)

64

• RANDOMNESS
• Events are random if there is
• no known law, capable of being expressed in
  language, that correctly explains events and
  their outcomes. . . .

65

• Events are random, when they cannot be predicted
  individually (although they may be predictable in
  the aggregate).

66
SAMPLING STRATEGIES

• PROBABILITY (Random) SAMPLES
• Use some form of random sampling in one or more
  of their stages.

67

• STRATIFIED SAMPLING
• The population is divided into strata, such as
  men and women black and white and the like,
  from which random samples are drawn.

68

• CLUSTER SAMPLING
• The most used method in surveys is the successive
  random sampling of units, or sets and subtests.
  In educational research, for example, school
  districts of a state or county can be randomly
  sampled, then schools, then classes, then pupils.

69

• SYSTEMATIC SAMPLING
• The first sample element is randomly chosen from
  numbers 1 through k, and subsequent elements are
  chosen at every Kth interval.

70

• NONPROBABILITY SAMPLES
• DO NOT USE RANDOM SAMPLING

71

• QUOTA SAMPLING
• Knowledge of strata of the population (sex, race,
  region, etc..) is used to select sample members
  that are representative typical and suitable
  for certain research purposes.

72

• ACCIDENTAL or CONVENIENCE SAMPLING
• Is the weakest form of sampling and probably the
  most frequent in which one takes available
  samples at hand.

73

• PURPOSEFUL SAMPLING
• Is characterized by the use of
• judgment and a deliberate effort to obtain a
  certain type of sample. It is often used as a
  strategy when one wants to learn something about
  select cases without needing to generalize to all
  cases.

74
Included are

• Sampling extreme or deviant cases
• Sampling typical cases
• Maximum variation sampling (picking cases that
  represent a range on some dimension)

75
(continued)

• Sampling critical cases
• sampling politically important and sensitive
  cases
• Convenience sampling--taking the easy ones

76
REPRESENTATIVE SAMPLE

• A sample which has approximately the
  characteristics of the population relevant to the
  research in question.

77

• When we draw a random sample, we hope that it
  will be representative, that the relevant
  characteristics of the population will be present
  in approximately the same way they are present in
  the population. But we can never be sure There
  is no guarantee.

78

• The purpose of trying to achieve a representative
  sample is to allow the researcher to draw
  inferences (conclusions, generalizations) about a
  population from studying a sample taken from that
  population.

79

• Sample size influences the degree to which
  random sampling will result in a representative
  sample.

80
VARIANCE

• VARIANCE A statistical characteristic of an
  array of scores indicating the extent of
  differences or range of scores (Asher, 1976).

81

• V S X2
• n
• STANDARD DEVIATION SX2
• n
• VARIANCE (Standard Deviation)2

82

• Much of research has the goal of explaining
  variance That is, identifying factors which
  account for difference in scores

83

• For example, if we give a reading achievement
  test to a group of students we will find that
  their scores will vary (i.e.. not everyone will
  get the same score). . . .

84

• . . . We my be interested in what factors
  account for these different scores. That is, we
  wish to know what factors account for reading
  achievement. We are interested in explaining
  variance in reading achievement.

85

• GENERAL
• TYPES
• OF
• VARIANCE

86
SYSTEMATIC VARIANCE

• Variation in measures due to some known or
  unknown influences that cause the scores to
  lean in one direction more than another. Any
  natural or man-made influences that cause events
  to happen in a certain predictable way are
  systematic influences.

87
ERROR VARIANCE

• Variation in measures due to chance. Error
  variance is random variance. It is the variation
  in measures due to the usually small and
  self-compensating fluctuations of measures--now
  here, now there, now up, now down. (Kerlinger,
  1973)

88

• An example of error variance is getting two
  different weight readings from a bathroom scale
  in successive weighings (Your true weight didnt
  change)

89

• SPECIFIC
• TYPES
• OF
• VARIANCE

90

• BETWEEN GROUPS (EXPERIMENTAL) VARIANCE
• Variance that reflects systematic differences
  between groups of measures. (Kerlinger, 1973).

91

• WITHIN GROUPS VARIANCE
• Unexplainable variation in individual scores
  within a group presumably due to chance.
  Essentially, a measure of error variance.

92

• PRIMARY VARIANCE
• The observed, consistent variation in behavior
  related to the independent variable. (Matheson,
  Bruce Beauchamp, 1970).

93

• SECONDARY VARIANCE
• The observed, consistent variation in behavior
  related to factors other than the independent
  variable. Secondary variance is the product of
  an unrecognized variable.

94

• TOTAL VARIANCE
• PRIMARY VARIANCE
• SECONDARY VARIANCE
• ERROR VARIANCE

95
AN EFFICIENT, WELL DEVELOPED RESEARCH DESIGN
ATTEMPTS TO

• Maximize the primary (between groups,
  experimental) variance of the independent
  variable under study.

96

• Minimize the error variance, or random variance,
  including so-called errors of measurement

97

• Control the extraneous,
• secondary variance of
• unwanted variables that
• may have an effect on
• experimental outcomes, but
• in which the researcher is not
• interested.
• THIS IS KERLINGERS (1970) MAXMINCON PRINCIPLE!

98
(No Transcript)
99
CONTROLLING SECONDARY VARIANCE

• Eliminate secondary variable.
• Hold the secondary variable constant
• Match subjects on the secondary variable

100

• Use subjects as their own control
• Randomize
• Systematize the secondary variable (use the
  secondary variable as an independent variable).
• Conservative arrangement of the secondary variable

101
MINIMIZING ERROR VARIANCE

• To minimize error variance, we must reduce
  measurement error. We do this via use of
  sampling procedures and by insuring reliability
  and validity of our measures.

102

• COVARIANCE
• A measure of the relation between two sets of
  scores, covariance is expressed statistically as
  a correlation coefficient (r).

103

• COMMON FACTOR VARIANCE
• Variance shared by two or more variables. Common
  factor variance is expressed statistically as a
  squared correlation coefficient (r2).