Docstoc

Introduction+to+Quantitative+Research

Document Sample
Introduction+to+Quantitative+Research Powered By Docstoc
					            Nursing 503 – Session 3

           Introduction to Quantitative
                    Research

            September 21, 2005




09-21-05          Dr. Kathy Kovacs Burns, Nursing   1
           Contents of presentation
 Research Designs – what, why, and
  characteristics
 Terminology in Research Designs
 Part 1: Characteristics of sound quantitative
  research designs
 Part 2: Validity and Reliability in
  quantitative research

09-21-05          Dr. Kathy Kovacs Burns, Nursing   2
       Research Design: How to
               Decide

What is the Research Design?

What is the purpose of research design?

What are the characteristics of the research
 design?
09-21-05        Dr. Kathy Kovacs Burns, Nursing   3
     Research Designs: What &
               Why
The Research Design
   is a structural framework within which the study is
   planned & implemented; a blueprint for conducting
   the study.
Its purpose is to
 Provide the best way to answer the research question
 Maximize control over factors which could interfere
   with the desired outcomes
 Instruct the researcher to gather & analyze data in
   certain ways, controlling who & what are studied.
09-21-05           Dr. Kathy Kovacs Burns, Nursing       4
      Types of Research Design
Descriptive – results in description of data through
  words, pictures, charts, tables and perhaps
  statistical & described relationships (Level I and II
  research questions)

Experimental – results in inferences based on the
  data which explain the relationship between or
  among variables (Level III)

Looking for “Goodness of Fit”
09-21-05          Dr. Kathy Kovacs Burns, Nursing     5
    Characteristics of Research
             Designs
 The setting for the study – laboratory vs. field
 Timing of Data Collection:
     – Cross-sectional: data collected one time only from a
       cross-section of the population at a given moment in
       time
     – Longitudinal: prospective or the future; over a period of
       time to see changes over time
     – Historical, retrospective, ex post facto – looking at
       events that have occurred in the past

09-21-05              Dr. Kathy Kovacs Burns, Nursing          6
                          Continued
 Type   & Method of Data to be Collected – Qualitative
  (interview, focus groups, observation) or Quantitative
  (testing, using instruments, measurements)
 Sample Selection, specifically the randomness of the
  sample for the study
   –        With random sampling every member of the population
            being studied has an equal chance of being selected
   –        Randomness is directly associated with generalizability
            across the entire population

 09-21-05                 Dr. Kathy Kovacs Burns, Nursing         7
           Key Terms in Research
                  Designs
   VARIABLES:
   Independent Variable – stands alone, is the cause of
    potential change in „cause& effect‟
   Dependent Variable – is affected by the independent
    variable(s); is the effect
   Intervening Variable – comes between the independent and
    dependent variables, and may interfere in some way; can
    also mask the effect of the independent variable
   Extraneous Variable – are not direct interest in the study
    by could affect the variables measured.

09-21-05              Dr. Kathy Kovacs Burns, Nursing        8
                   Other Terms
   Control – refers to having control over variables
    which are part of the study as well as those that
    could affect the study
     – Allowing for no variation
     – Specifying the variation allowed
     – Distributing the variation equally across study


     Test for control is Validity – External and
      Internal

09-21-05              Dr. Kathy Kovacs Burns, Nursing    9
    Validity in Research Designs
Internal Validity                       External Validity
                                         Is concerned with the extent
 Extent to which the effects
                                          to which the study findings
   detected in the study are a            can be generalized beyond
   true reflection of reality,            the sample used in the study
   rather than the result of             Degree to which the sample
   extraneous variables                   represents the population;
 Determined by the way the               probability sample
   experimental and control              If study is not generalizable
   groups are formed                      based on nonprobable
 Attributed to the action of             sample, then you could be
   the independent variable and           accused of making a
                                          “quantum leap” from the
   not something else                     data to the conclusions.
  09-21-05             Dr. Kathy Kovacs Burns, Nursing              10
   Part 1: Characteristics of sound
    quantitative research designs
OVERVIEW:
 Critical elements of a sound quantitative
  study
 Appropriate sample – determining sample
  size & adequacy of sample size; influence
  of research method;
 Factors influencing the power of a statistical
  test
09-21-05        Dr. Kathy Kovacs Burns, Nursing   11
               Critical Elements of a sound
                     quantitative Study
   Quantitative studies are categorized as experimental, quasi-
    experimental or non-experimental designs
   Sound Quantitative studies account for
    – the intervention and specific testing of the effects of the
        intervention
    –   Comparisons between or within groups, of rankings of the
        variable, or comparison with other studies
    –   Controls over independent variable & for extraneous variables
    –   Timing of data collection (cross sectional, longitudinal)
    –   Research sites and settings
    –   Communication with the subjects

    09-21-05             Dr. Kathy Kovacs Burns, Nursing            12
  Characteristics of True Quantitative
             Experiments
 Manipulation – The researcher does something to
  some of the subjects (intervention or treatment);
  alters/varies the independent variable to see what
  effect it has on the dependent variable
 Control – controls are introduced by the researcher
  over the experimental situation, including the use
  of a control group to be compared to a
  experimental group.
 Randomization – the researcher assigns subjects to
  a control or experimental group on a random basis

09-21-05          Dr. Kathy Kovacs Burns, Nursing   13
 Gold Standard in Quantitative
          Research
Experimental studies
 The randomized trial
     – Set out a purpose
     – Establish one or more hypotheses to test (hypothesis is
       a prediction of the relationship between two or more
       variables)
     – Statistical inference states that the hypothesis is framed
       to indicate no relationship between variables – this null
       hypothesis is assumed true unless evidence suggests the
       contrary

09-21-05               Dr. Kathy Kovacs Burns, Nursing           14
      Research Method influencing
             Sample Size
 Experimental studies, such as clinical trials, need
  large numbers of subjects which are randomized
 Need for randomized groups of subjects which
  require larger numbers of subjects to increase
  equity and decrease bias
 Qualitative studies are less inclined to need large
  numbers of subjects, depending on quality of
  technique used or amount of data needed from
  each subject or from entire target group.

09-21-05          Dr. Kathy Kovacs Burns, Nursing       15
           Determining Appropriate
                  Sample
Sampling Concepts: target populations,
  elements, randomization, sample design,
  sampling frame, accessible population,
  representativeness.
Sample – a subset of elements or members of
  a population; employed in such a way as to
  suggest representation from the population
  from which it was drawn.
09-21-05         Dr. Kathy Kovacs Burns, Nursing   16
    How big should the sample be?
   No definitive statistical answer
   Bigger samples have less sampling error
   Smaller samples are easier to manage and are less
    expensive
   The larger the effect size in population the smaller
    the sample needed, and vice versa
   Pilot study can judge the implications of sample
    size for accuracy of final estimates
   Search literature for sampling suggestions in
    similar studies

09-21-05            Dr. Kathy Kovacs Burns, Nursing    17
     Purposes of Literature Review
       in Quantitative Research
1.    Clarify the research project        8.     Develop definitions of
2.    Clarify the research                       major variables
      problem                             9.     Identify limitations and
3.    Verify the significance of                 assumptions
      the research problem
                                          10.    Select a research design
4.    Specify the purpose of the
      study                               11.    Identify tools of
5.    Identify relevant studies                  measurement
6.    Identify relevant theories          12.    Direct data collection &
7.    Clarify research                           analysis
      subproblems                         13.    Interpret findings

     09-21-05            Dr. Kathy Kovacs Burns, Nursing               18
             Sampling error
   The bigger issue for sampling is sample
    distribution and relative error where the
    sample does not reflect the population being
    studied (drawing too many of one kind due
    to bias or small population to draw from, or
    homogeneity vs. heterogeneity of the
    population)

09-21-05         Dr. Kathy Kovacs Burns, Nursing   19
           A Sampling Distribution
Distribution patterns

Sampling Error




09-21-05         Dr. Kathy Kovacs Burns, Nursing   20
    Randomization of Sampling
   Use of a sampling frame (lists or set of elements
    from which sample can be selected)
   Random selection of subjects and placing them
    into groups at random
   Cluster randomization also works where groups of
    individuals can be randomly assigned
    interventions and compared
   Account for systematic bias in groups with respect
    to attributes that could affect dependent variable
   Matching of subjects for control and experimental
    groups
09-21-05           Dr. Kathy Kovacs Burns, Nursing   21
    Factors Influencing the Power of a
              Statistical Test
 Statistical significance is achieved by hypothesis testing –
  null and alternate
 The power of the statistical test is the probability of
  correctly rejecting the null hypothesis when it is false, or
  the ability of the test to identify correctly that there is a
  difference between groups in a trial
 Statistical power analysis exploits the relationships among
  four variables – sample size (N), significance criterion
  (),population size effect (SE), and statistical power.Each
  is a function of the other three.
 Probability of Type One error – In reality null hypothesis
  is true & results are significant, but null is rejected.
09-21-05             Dr. Kathy Kovacs Burns, Nursing          22
                  continued
 Effect size (ES) – prespecifying the magnitude
  of the difference between the two groups that
  can be regarded as clinically meaningful and
  important; measure of how wrong the null
  hypothesis is
 Probability (P value) – probability of obtaining
  the observed difference between the two groups
  if our null hypothesis is true; if the P value for
  the trial < , reject the null hypothesis, & vice
  versa

    09-21-05      Dr. Kathy Kovacs Burns, Nursing   23
Part 2: Validity & Reliability in
    Quantitative Research
 What does validity mean in quantitative
  research?
 What techniques enhance validity and
  reliability in quantitative research?
 How are validity and reliability evaluated in
  quantitative research?


09-21-05        Dr. Kathy Kovacs Burns, Nursing   24
        Validity in Quantitative Research
Definition of Validity:
   the extent to which any measuring instrument measures what it is
   intended to measure
Internal and External Validity
Construct Validity – examines the fit between the conceptual
   definitions & operational definitions of the variables
Content Validity – verifies that the method of measurement actually
   measures the expected outcomes.
Predictive Validity – determines the effectiveness of the instrument
   as a predictor of a future event
Statistical Conclusion Validity – concerned with whether the
   conclusions about relationships and/or differences drawn from
   statistical analysis are an accurate reflection of the real world

   09-21-05             Dr. Kathy Kovacs Burns, Nursing        25
       Enhancing Validity and Reliability
Reliability – associated with the methods used to measure research
  variables; refers to the accuracy and consistency of information
  obtained in a study; important in interpreting the results of
  statistical analyses; and refers to the probability that the same
  results would be obtained with different samples
  (generalizability)
Validity – also associated with the methods used to measure
  research variables; supports measure of generalizability

Enhance both through vigorous controls of research design.
  including the use of manipulation, randomization, and a control
  group
  09-21-05             Dr. Kathy Kovacs Burns, Nursing         26
           Evaluating Validity &
                Reliability
 Instruments used to collect data must be
  tested for accuracy and consistency in
  measuring what it is supposed to measure
 Pilot study to assess instruments, sample,
  and results obtained
 Results obtained from measurement should
  be true results & not due to error in
  instrument.

09-21-05        Dr. Kathy Kovacs Burns, Nursing   27
           Questions & Discussion
Questions??

Group Answers

Examples/Experiences



09-21-05         Dr. Kathy Kovacs Burns, Nursing   28

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:23
posted:5/23/2010
language:English
pages:28