Part 3 Flashcards
Research Design: An overview (EXPERIMENT)
Before conducting a research the resercher has to think in the best way to:
- Answer their research question
- Reach their objective
- Test their hypothesis
What is a research design?
The overall strategy that you choose is to integrate the different components of the study in a coherent and logical way.
Heildarstefnan sem þú velur til að samþætta mismunandi þætti námsins á samfelldan rökréttan hátt
Describe Research design?
A research design is a plan for answering a research question using empirical data.
- What type of research and how will you do it
Design research, decide on:
- Your overall aims and approach
- Type of design you will use
- How you select a participant
4 Your data collection method - The procedure you follow
- Your data analysis strategy
Quantitative and Qualitative Research Design
Quantitative research: specific behavior that can be easily quantified.
Qualitative research: people behaving in natural settings and describing their world in their own words.
Quantitative research
Objective - Sample - Data collection
Objective: Quantify data and generalize result from a sample to understand the population of interest
Sample: Large and broad, statistically projectable
Data Collection: Standardized instrument, operation of variables
Qualitative Research
Objective: Understanding underlying reason or motivations
- provide insight into the problem
Sample: small and narrow, not statically projectable
Data Collection: Adapt to the situation, variables not defended in advanced
Naturalistic observation (Tend to be qualitative)
Objective: provide a complete and accurate picture of what occurred in the settings, rather than test hypothesis formed prior to the study
Method:
The researcher must keep detailed field notes that are, written on a regular basis, everything has been happened (at least once each day)
Issue:
- Subjectivity
- Ethical Problem
- Time-consuming
Systematic Observation (can be quantitative)
Objective:
Careful observation of one or more specif behavior in a particular setting.
Method:
Coding system: decide which behavior are of interest, choose a setting in which the behavior can be observed, and how to codify.
Issue:
Equipment: not only pencel and paper
- Reactivity
- Reliability - At least two raters
Quantitative research - Experimental design: what is their aim?
Their goal of scientific psychology is to understand human behavior
- To accurately describe its casual underpinning
- To predict behavior
Quantitative research - Experimental Design.
Aim of the experimental design?
- Cause-effect relationship between variables
- Using predictive analytics
- Having high internal validity
Internal validity
To see which study establishes a trustworthy cause and effect relationship between a treatment and an outcome.
In a true experiment, we want to?
- We want to maximize the independent variance
- Minimize error variance
- Control external variable
In large samples, outcomes predicted by chance have
- normal (Gaussian) distribution
Described as “bell-shaped-curve” - The outcome is distributed around the means
- Probability is smaller and further from means
Variance is
The amount that scores vary around the mean score
(Variance should be similar between the groups) = Homogeneity
i
If variance is not similar between the groups
It might affect the validity of the outcome.
- Statistical significance is often based on how much the scores vary.
EXPERIMENTAL DESIGN (MAXIMISE AND MINIMIZE WHAT?
MAXIMIZE INDEPENDENT VARIANCE
AND MINIMIZE ERROR
Variance that is similar between groups is also called?
Homogeneity
Experimental control -> Minimize error variance (LARGE HROUP OF PARTICIPANTS)
- Large group of participants
- Suitable measuring instrument
- Rigorous research planning
Inductive
Inferring general principles or rules from specific fast.
YOU BEGIN WITH DATA
EXAMPLE: In the summer, there are ducks in our pond. Therefore, summer will bring ducks to our pond.
Experimental control - Controlling external variable
- Balancing design
- Remove outliers
- Assumption of normality
Design of experiments (Balanced design vs unbalanced design)
Balance design: has an equal number of observations for all possible level combinations
Unbalanced design: it has ab unequal number of observations.
Example: taste-test various cereal brands.
Balanced design: 30 boxes of each brand
Unbalanced design: 30 boxes of lucky charms, 28 boxes of cheerios, and 30 boxes of cinnamon crunch
Experimental control - Controlling external variable (ASSUMPTION OF NORMALITY)
The most common assumption for statistical analysis is the normality of data distribution (central limit theorem)
when there is NOT normal distribution, it is possible to fix it by applying statistical TRANSFORMATION. It changes the distribution by applying a mathematical function to each participant data value.
Defining the experimental method
Cross-section studies vs Longitudinal study vs Cohort study
Cross-section: Different groups of people are tested at the same time and their results are compared.
- Quick to carry out
- Easy to test the reliability of findings
Longitudinal study: the participant is studied over a long period of time
- Track development, - Monitor change over time
Cohort study: several groups are studied over a long period of time.
Defining the experimental method (Within-subject vs between-subject)
Within-subject: each participant takes part in all experimental method
Between-subject: each participant takes part in one experimental condition
Experimental design (Random assignment to experimental conditions)
Benefits of randomization:
- eliminated to a selection of bias
- Balance the group with respect to many known and unknown confounding or prognostic variables
- the randomized experiment is essential for testing for efficacy of treatment (tilraun er nauðsynileg til að prófa virkni meðferðar)
Quasi-Experimental Design
- When subjects cannot be randomly assigned to different treatment conditions
- When the independent variable cannot be manipulated
- no random assignment
- no manipulation of IV’s
- no control
Aim of the Quasi-Experimental Design
- Analytical predictive objective
- Impossibility to control confounding variables
- Reduce internal validity (because extraneous variables are not held constant)
Aim of the Quasi-Experimental Design
- Analytical predictive objective
- Impossibility to control confounding variables
- Reduce internal validity (because extraneous variables are not held constant)
E.g School A takes part in an after school tutoring program ("treatment" condition) and school B does not (control condition) Manipulated IV (tutoring program); no random assignment.
Quasi-Experimental Design is sometimes called?
Pre-Post intervention design is often used to evaluate the benefits of specific interventions.
Single-subject design
- Begin with systematic observations of an individual’s behavior for several sessions before treatment is introduced
- Known as baseline phase or pre-treatment phase
Purpose:
*Identify/describe the extent of a person’s problem (status of the target behavior in its naturally occurring existing state)
*Predict future behavior if intervention is NOT provided