Unit #2 – Flashcards
52 test answers
Unlock all answers in this set
Unlock answers 52question
In summary, research design serves to: (2)
answer
1. Provide valid answers to the research question or hypothesis 2. Maximize control over factors that might increase variance
Unlock the answer
question
Research designs can be explained by symbols such that:
answer
- O= observation or measurement of the dependent variable - X= manipulation of the independent variable - R= random assignment of subjects to groups
Unlock the answer
question
O....X....O
answer
- Indicates a one-group design w/ an observation (or measurement) followed by a treatment and then another observation
Unlock the answer
question
O....X....O R O..........O
answer
- Indicates a two-group design with only one group receiving a treatment and random assignment of subjects to both groups
Unlock the answer
question
What are the 2 criteria used to assess the validity of research findings?
answer
1. Internal validity 2. External validity
Unlock the answer
question
What question does internal validity ask? What are these things called?
answer
- Did the "treatment" or "intervention' that constitutes the independent variable produce the effect seen in the dependent variable or as the effect due to some other factor - These other factors are called "threats to validity"
Unlock the answer
question
What are the most common threats to internal validity? (6)
answer
- History - Maturation - Testing - Instrumentation - Mortality - Selection bias
Unlock the answer
question
In regards to validity threats, what is history?
answer
- Refers to any specific event that occurs between the pretest and posttest and is not part of the planned study, but may affect the dependent variable
Unlock the answer
question
In regards to validity threats, what is maturation?
answer
- Refers to psychological, biological or any internal maturation that occurs within individuals over tim, and is not part of the planned study.
Unlock the answer
question
In regards to validity threats, what is testing?
answer
- Refers to the potential effect or influence that taking a test or filling out a scale can have when being retested, that is, the effect of testing at time 1 on time 2.
Unlock the answer
question
What is the "Hawthorne Effect"?
answer
- When behavior of subjects change when they know they are participating in a study - may be trying to please the "researcher" - This change in behavior is not due to the research treatment
Unlock the answer
question
In regards to validity threats, what is instrumentation?
answer
- Refers to changes in measuring instruments that may occur during the course of a study
Unlock the answer
question
In regards to validity threats, what is mortality?
answer
- Refers to subjects who drop out of a study before it is completed - Becomes a threat to internal validity if those who drop out are different in relation to the dependent variable than those who continue in the study
Unlock the answer
question
In regards to validity threats, what is Selection Bias?
answer
Refers to the possibility that 2 groups of research subjects may be different prior to introduction of the manipulation or treatment
Unlock the answer
question
After reading a research report, a reader might ask:
answer
Can these findings be generalized to other: - Populations - Settings - Conditions
Unlock the answer
question
The threats to external validity include, but are not limited to: (2)
answer
- How the sample was selected, AND - How representative the sample is
Unlock the answer
question
The 3 types of Research Design
answer
1. Experimental 2. Quasi-Experimental 3. Non-Experimental
Unlock the answer
question
What is the purpose of Experimental Design?
answer
- The purpose is to test cause and effect relationships between variables - That is to determine if manipulation of the independent variable has an effect on the dependent variable.
Unlock the answer
question
What is the most powerful design for testing hypotheses of cause and effect relationships?
answer
Experimental Design
Unlock the answer
question
What is required in experimental designs?
answer
Hypotheses
Unlock the answer
question
3 criteria that must be met for a cause & effect relationship.
answer
1. The cause must precede the effect in time 2. There must be evidence that the independent and dependent variables are associated 3. There must be evidence ruling out other factors as possibly determining or influencing the dependent variable
Unlock the answer
question
What are 3 requirements for experimental design? Explain each.
answer
1. Manipulation of the independent variable - manipulation can be anything the researcher can control 2. Control - an experimental design requires there must be 2 or more groups of subjects 3. Random assignment of subjects to either the experimental or control group
Unlock the answer
question
What is one problem with Random Assignment into groups?
answer
- Theoretically, random assignment to groups eliminates any sytematic bias that may exist in groups and that, therefore, may affect the dependent variable. (See example in notes)
Unlock the answer
question
What are 3 examples of Experimental Designs?
answer
1. Pretest/Posttest Control-Group 2. Posttest Only Control-Group 3. Solomon Four-Group
Unlock the answer
question
Give Schematic Diagram of Pretest/Posttest Control-Group. What is it called if no random assignment?
answer
Experimental Group R O₁...X...O₂ Control Group R O₁........O₂ -Called a "nonequivalent control-group quasi-experimental design"
Unlock the answer
question
Give Schematic Diagram of Posttest Only Control-Group Design. Theoretically, why isnt this design a threat to validity?
answer
Experimental Group R X...O Control Group R O - Since only posttest and no pretest done, not certain two groups were equivalent at beginning. - Theoretically, randomization of groups should eliminate this threat to validity.
Unlock the answer
question
Give Schematic Design of Solomon Four-Group Design. What is advantageous about this design?
answer
Experimental Group 1 R O₁...X...O₂ Experimental Group 2 R X...O₂ Control Group 1 R O₁........O₂ Control Group 2 R O₂ - You can potentially make multiple comparisons with this design.
Unlock the answer
question
What is a strength of Experimental Method? What threats to internal validity does it rule out? What threats does it not?
answer
- Most powerful design for testing cause and effect hypotheses - Rules out: - History - Testing - Selection Bias - Doesn't Rule Out: - Mortality - Instrumentation
Unlock the answer
question
What are the 3 limitations of Experimental Method?
answer
1. May not be possible or ethical to manipulate many of the variables of interest to clinical researchers, such as the withholding of basic patient care. 2. Many variables of interest are multidimensional and complex processes - important info may be missed when such an experience is "measured" by response to one variable or by response to one tool or instrument 3. Variables under the controlled conditions of an experimental study may not be the same when under "real world" conditions.
Unlock the answer
question
What is the purpose of quasi-experimental designs? How do they differ from experimental designs?
answer
- Has the same purpose -To test cause and effect relationships However, they do not have: - Random assignment of subjects to groups, or - A control group, or - Both of the above
Unlock the answer
question
What do quasi-experimental designs DO have? Why are they common?
answer
- They do have experimental treatments - Common because: - It may not be possible to randomly assign subjects to groups and/or - It may not be possible to have a control group
Unlock the answer
question
What are 4 examples of Quasi-experimental Designs?
answer
1. Pretest/Posttest Nonequivalent Control-Group 2. Posttest only Nonequivalent Control-Group 3. One-Group Pretest/Posttest 4. Interrupted Time Series
Unlock the answer
question
In quasi-experimental designs, what is the control group refereed to as? What does "Nonequivalent" mean? What is it represented by?
answer
- The "comparison group" - Groups were not selected randomly - Indicated by a dashed line seperating the 2 groups
Unlock the answer
question
Give Schematic of Pretest/Posttest Nonequivalent Control-Group Design. What is it common too and what is the difference? What does the difference cause?
answer
Experimental Group O₁...X...O₂ ---------------------------------- Comparison Group O₁........O₂ - Similar to pretest/posttest control-group design - Difference is lack of randomization - Makes design vulnerable to all the threats of internal validity
Unlock the answer
question
Give Schematic Diagram of Posttest only Nonequivalent Control-Group. Why is this design faulty?
answer
Experimental Group X...O ---------------------------------- Comparison Group O - Since no R and no pretest, virtually no way researcher can know if the 2 groups were equivalent at the beginning of the study in relation to the dependent variable
Unlock the answer
question
Give schematic of One-Group Pretest/Posttest Design. Explain if good/bad.
answer
Experimental Group O₁...X...O₂ - One of the least desirable quasi-experimental designs. - No comparison group, therefore no R. - Impossible to ascertain weher change between O₁ and O₂ would have occured w/o the experimental manipulation. - Open to all internal & external validity threats.
Unlock the answer
question
Give schematic of Interrupted time series design. What are 2 ways you could manipulate this design to make it stronger?
answer
Experimental Group O₁...O₂...O₃...X...O₄...O₅...O₆ 1. Addition of a comparison group -Makes a very good nonequivalent control-group 2. Random assignment of subjects to the 2 groups -Makes a very strong experimental design
Unlock the answer
question
Strengths of Quasi-Experimental Design?
answer
- Potentially more able to implement on the "real world" - May be the only way to study some variables. - The findings from such studies may contribute to the development of more controlled studies.
Unlock the answer
question
Three Limitations of Quasi-Experimental Design?
answer
1. At risk for most all threats to internal validity due to lack of "control" 2. W/o random assignment ot groups, all q-e designs are especially at risk for selection bias 3. Because of the above, there is limited ability to make cause and effect statements
Unlock the answer
question
What are 3 characteristics of Non-experimental design?
answer
1. The researcher isn't able to control or manipulate the independent variable. - Instead the variable of interest may be some natural occurring event 2. Subjects cannot be randomly assigned to groups - They are selected on the basis that they have experienced the variable of interest 3. The type of research question influences the choice of design
Unlock the answer
question
Give schematic of Non-experimental design.
answer
Group A O or Group A O ___ Group B O - No R and no X - Only observation of dependent variable
Unlock the answer
question
What is Survey Research?
answer
- The researcher gathers data from a portion of a target population for the purpose of identifying opinions (e.g., voting preferences), characteristics (e.g., one-parent families), or commonalities and differences (e.g., differences in procedures for skinning rodent specimens in taxonomy)
Unlock the answer
question
What are the strengths of Survey Research? (2)
answer
- Possible to gather data from large numbers of respondents - Can be used for many different topics and populations
Unlock the answer
question
What are the limitations of Survey Research? (3)
answer
- Does not address cause and effect relationships - Does not provide in depth data - May be costly in terms of resources - e.g. money and personal record keeping, especially in relation to quantity and quality of data obtained
Unlock the answer
question
What is descriptive research?
answer
- The researcher gathers data for the purpose of describing some phenomenon or the experience of a life event, usually about which little is known (widowhood, first-time fathers, mid-life crisis, etc.)
Unlock the answer
question
What is Correlational Research?
answer
- For the purpose of assessing and describing the relationship(s) among 2 or more variables -e.g. Exercise and dieting in wt. loss - In correlational research, no independent or dependent variables are specified.
Unlock the answer
question
What are 2 strengths of Descriptive/Correlational Research?
answer
- Provides rich descriptive or exploratory data upon which further research can be based - Useful for describing phenomena about which little is known for determining "new phenomena"
Unlock the answer
question
What are 2 limitations of Descriptive/Correlational Research?
answer
- Degree of generalizability of the data - There may be multiple interpretations of the data
Unlock the answer
