Вы находитесь на странице: 1из 19

How to Design

Discovery Research
Chapter 11
Discovery Paradigm
 Nature: One singular, objective reality
 Role of Knower: Discovered by any knower
 Role of Context: Detached and decontextualized
 Process: In precise, systematic, and repetitive ways
 Purpose: To accurately represent reality
 Goals: By classifying objects and identifying universal
rules or laws
The Process of Basic Science
 Select a topic of interest
 Select an appropriate theory/model/concept
 Read relevant research
 Derive hypotheses and research questions
 Design the study
 Test your hypotheses and research questions
Measurement
 Measurement: step by step procedure for making
observations and assigning numbers to them

 Quantification of data, prediction and power


Conceptualization
 Using theories and concepts to define other concepts
and link these concepts to our observations

 Constructs: agreed upon conceptual meaning


 Conceptual definition: what a construct is and is not
 Derived from literature
 Variables: constructs that have been defined in
measurable ways
Conceptualization
 All variables contain attributes
 Values/levels
 Sex
 Political Parties
 Types of relational maintenance behaviors
 Communication apprehension in different contexts
Operationalization
 Operationalization: the process of specifying
procedures used and instruments designed for
measuring communication constructs

 Operational definitions: every operation, procedure,


and instrument needed to measure a construct
 Describes a construct in terms of it’s observable and
measureable characteristics or behaviors by specifying
how the concept can be observed in actual practice
Example
 Verbal aggressiveness
 Conceptually defined as “any communication behavior
that attack’s a person self concept to inflict
psychological pain” which includes “character attacks,
competence attacks, teasing, insults, and ridicule”

 Operational: Verbal Aggressiveness Scale


Levels of Measurement
 To operationalize a concept, construct, or variable, you
will use one of four different levels of measurement
 Nominal
 Ordinal
 Interval
 Ratio

Remember these??
Research Design
 Research Design: the methodological framework you
use to test your ideas or theories

 Causal arguments allow for predictive control


 Like a blueprint for the research
 Research design is a plan that guides the decisions of:
 When and how often to collect data
 What data to gather and from whom
 How to analyze data
Causal Arguments
 Three criteria:
 Time order
 Covariation
 Controlling competing explanations
Experimental Research
Design Elements 1/3
 Comparison groups: groups of participants who are
exposed to manipulated levels of the independent
variable
 Control group: when participants are not exposed to
any variation of the independent variable
 Treatment/experimental groups: receive some
exposure to manipulation of the independent variable
 Placebo: false treatment – participants think they are
getting manipulated, but they are not
Example
The amount of sugar consumption increases
hyperactivity in children

Comparison group: children who get a lot of sugar

Control: children who get no sugar

Treatment/experimental: children who get ½ sugar

Placebo: children who think they are getting sugar but are
not (i.e. sugarless drink)
Experimental Research
Design Elements 2/3
 Random Assignment: refers to a procedure in which
the researchers select participants on a purely random
basis to participate in treatment or control conditions
 Not the same thing as selection or random sampling from
Chapter 5

 Help establish equivalent groups at the beginning of


experiment
Experimental Research
Design Elements 3/3
 Pretesting: assessing a baseline or naturally occurring
levels of the dependent variable before the IV is
manipulated

 Posttest: giving the pretest again after the


manipulation occurs

 Helps give a precise comparison


Types of Design –
Preexperimental Design
 One-shot case study: some change in the IV occurs
and the researcher claims a particular set of effects on
the DV

 One-group pretest/posttest design


 Strengthens the causal argument
 Helps understand change
Types of Design –
Quasi-experimental design
 Time series design: assess levels of the DV at several
points in time prior to and following the manipulation of
the IV
 Helps when other factors may occur
 Nonequivalent control group design: use a control
group not exposed to the IV manipulation
Types of Design –
True experimental
 Pretest/posttest control group
 Comparison groups
 Random assignment
 Pretest/posttest
 Posttest only control group
 No pretest
 Soloman four group design
 Two groups using pretest/posttest
 Two groups using posttest only
No Control Control

Pre-Exp Quasi Full/True


2 or more 2 or more Random
categories categories Assignment
OR OR
A pretest Pre/post test

Вам также может понравиться