Quiz 8

Question 1

Validity

Answer 1

Refers to the authenticity of the results of a study

Question 2

Internal validity

Answer 2

Results are directly related to the intervention . A study has internal validity when the experimental results are directly related to the intervention/treatment used in the study rather than a confounding variable.

Question 3

Confounding variable

Answer 3

Is one that the researcher is unaware is having an influence on the study;as such, it is not measured or observed.

Question 4

Quasi- experimental vs experimental

Answer 4

Similarities:

• Control Group
• An intervention

Differences:
-quasi-experimental design does not use probability sampling methods and does not randomly assign study participants to an experiment or control group (lower degree of external validity)

Question 5

external validity

Answer 5

results can be generalizable to a larger population.

Question 6

Experimental research- The terms treatment and intervention can be used interchangeably in medical research

Answer 6

-When the researcher is doing medical research (testing medical treatment plan, testing a new drug treatment, testing a new research device, etc) the researcher will use the words intervention and treatment interchangeably.

-Treatment example
clinical studies of drugs, clinical medical treatments, Testing a new drug, device, medical treatment plan

-Intervention
If the researcher is testing a complex program that has been designed to change health behaviors, attitudes, or environmental and social conditions that impact health outcomes. ex. such as programs designed to change behaviors, attitudes, etc., the term intervention is used

Question 7

Experimental research- Manipulation of Variables

Answer 7

Independent is purposely changed by the researcher (i.e., treatment, new intervention) to test the effect on the dependent.

Question 8

Experimental research-Control

Answer 8

• Used to prevent outside variables from influencing the outcome of the study.
• The experimental and control group are equivalent with the exception that the experimental group gets the intervention or treatment; the control group does not.

Question 9

Experimental research- Random Sample

Answer 9

A random sample uses of one of the probability sampling methods.
-A random sample is thought to be composed of participants that are representative of the population at large

Question 10

Experimental research- Random Assignment

Answer 10

Participants in the random sample have an equal chance of being assigned to the experimental or control group

Question 11

Research Notations

Answer 11

Are a shorthand used by researchers to illustrate a specific research design.
-Developed by Campbell and Stanley
-O = an observation (dependent variable); O is the outcome of the influence of the independent variable (X)
-R = indicates participants have been randomly assigned to the experimental or control group
-G = indicates that participants have bee grouped in a non-random way
-X = the treatment or intervention (independent variable) the experimental group is exposed to
*The effects of the treatment or intervention are
measured
* The control group is NOT exposed to the
treatment or intervention

Time: the left to right dimension sequential order of procedures in the experiment

Subscripts: Subscript numbers show difference. When the number follows the R, it shows that there are 2 groups. When the number follows the O, it shows that the study has several different observations (data collection points)

Question 12

Experimental Designs: Pretest-Posttest Control Group Design

Answer 12

-controls for many threats to internal validity
-Two groups, randomly sampled and randomly assigned to experimental (R1) or control (R2) group
-Random assignment to experimental or control group increases interval validity of the study
-Group R1 receives a treatment or intervention; Group R2 does not
-Both groups receive an observation at the same time prior to and after the intervention (for the experimental group)
Because the groups were randomly sampled and randomly assigned, any differences in the posttest results should be due to the intervention and not any differences between the two groups

A possible threat to the external validity of this design is the pretest effect
process of observing or assessing people before administering the experimental treatment (pretest) may influence how people respond to the treatment; possible threat to the external validity of this design is the pretest effect

Question 13

Experimental Designs: Posttest-Only Control Group

Answer 13

• sometimes you can not pretest
  
  • May not be able to find a suitable pretest
  • May be studying a life event in which it wasn’t possible to pretest individuals before the event

-controls for threats to internal validity

Example: study effectiveness of an intervention for PTSD following a hurricane
Random selection of participants from a community that had recently experienced a hurricane and experienced PTSD; random selection to experimental and control groups
X = PTSD intervention

Question 14

Experimental Designs: Soloman Four-Group Design

Answer 14

This design has the highest level of controls for threats to internal validity by controlling for pretest effects. Combines the pretest-posttest design with the posttest only design. (pg145)
Allows the researcher to determine if the results are valid-meaning, has the study been influenced by pretesting

ex: Consider the following narrative description of the scientific notation. A researcher is interested in studying the effect of computer assisted cognitive training on memory loss in a population of adults, between the ages of 50 and 60. Participants are randomly selected and then randomly assigned to groups R1, R2, R3 and R4. The table below depicts the observations (memory tests) and treatments (computer assisted cognitive training) received by each group.

• Controls for many threats to internal validity
• Examples of threats to internal validity: history and maturation (has a control group); comparison of R2 and R4 control for pre-test effects

History: an outside event occurred during the research study that can impact the results of the study
Control for this threat: select a research design that includes a control group

Maturation: time passed and the participants grew “older, wiser, stronger, or more experienced
Control for this threat: select a research design that includes a control group

if the researcher wanted to add another layer of control to the study internal validity, a placebo could be added

Question 15

Experimental Design: Within-Subjects Design (Repeated measures design)

Answer 15

The subjects participating in the research are exposed to each intervention, otherwise known as the independent variable. Each subjects performance is repeatedly measured.

The strength of this design: Does not require a large number of study participants, as the same participants receive multiple interventions. Use of one group helps control for the internal threats to validity of history and maturation. Subjects serve as their own control for individual differences between subjects.

Weaknesses: Carry-over effect, fatigue, and practice effects.

Carry-over effect: exposure to 1 intervention impacts the performance of study participants for the next intervention.

Practice effect- an improvement in one’s performance on a test due to repeated exposure to the same test materials

This design can also fall under quasi-experimental study designs if non-probability sampling is used to obtain the participants for the study.

Question 16

Any of the experimental designs can fall under quasi-experimental categories based on how participants are sampled and assigned to experiment or control groups

Answer 16

true

Question 17

Quasi-experimental

Answer 17

Participants are not randomly assigned to the experimental and control groups, and this are considered nonequivalent.

limitation of this type of study: without randomization, there is no guarantee that the differences btwn the groups are due to the intervention, but rather the results might be due to chance.

The researcher uses 1 of the non-probability sampling methods, and then makes 1 group the experimental and 1 the control (no random assignment)

all quasi-experimental studies have less external validity than studies using an experimental design

Question 18

Quasi-experimental: Nonequivalent (Pretest and posttest) control group

Answer 18

Dotted line indicates groups were not randomly assigned; scientific notation varies across discipline

Two nonequivalent groups means:
Participants in study were not randomly assigned to the experimental or control group (as indicated by a dotted line)
Both groups were given a pretest and posttest; only one group received an intervention

Question 19

Quasi-experimental: Single Group Time Series Design

Answer 19

several measurements of the dependent variable over time. The measurements prior to intervention are called baseline data

Once there is a repeated measure of the dependent variable prior to the intervention, post-intervention measurements allow the researcher to assess what impact, if any, the intervention had on the variable

This design can only be quasi because no control group

could be influenced by history and maturation (threats to internal validity)

Interuption: the interrupted time series design could involve studying the change in a variable that the researcher did not manipulate

Question 20

Quasi-experimental: Control Group Time-series design

Answer 20

The only difference btwn this design and the single group time series design is the inclusion of a control group, which increases the internal validity of the experiment

Multiple observations at posttesting

controls for more threats to internal validity such as history and maturation because it has a control group

Question 21

Using non-probability sampling also reduces the external validity of the study

Answer 21

true

Question 22

Simple Random Sampling

Answer 22

Everyone in the population has an equal chance of being selected

ex:The study needs 500 participants. A computer randomly selects 500 participants from the master list of people receiving treatment and services for heart disease in the health system (the population

Question 23

Stratified Random Sampling

Answer 23

Divide the population into subgroups or strata, that are roughly equal in size, based on one or more variables of interest

Draw a simple random sample from each stratum

ex: The researcher needs 500 participants for a study and race/ethnicity is important for the study
Divide the master list into subgroups by race/ethnicity and randomly select an equal amount from each subgroup to get a representative sample of the population.

Question 24

Proportional stratified sampling

Answer 24

Strata are unequal in size
The sample size of each stratum is proportionate to the population size of the stratum

ex:The researcher needs 500 participants for the study but realizes that race and ethnic identity are not equally distributed across the population. Divide population into subgroups by race/ethnicity and randomly select the same percentages as found in the population to get a representative sample.

A probability sampling method in which different strata in a population are identified and in which the number of elements drawn from each stratum is proportionate to the relative number of elements in each stratum.

Question 25

Cluster sampling

Answer 25

Select an intact homogeneous group(s) from within the population

ex: : The researcher breaks down the master list by the various heart disease treatment clinics in the health system. The researcher sees that the patients in each of the 20 clinics are very similar and randomly selects 5 clinics to serve as the sample.

Question 26

Systematic sampling

Answer 26

The researcher selects participants based on a randomly chosen number

ex: Starting at a random point in the master list, the researcher would select every nth person.

Question 27

Non-probability sampling methods: Convenience

Answer 27

readily available participants.
Ease in selection, like a class.
SBU student data might not be generalizable to students from a small southern college.

Question 28

Non-probability sampling methods:Quota

Answer 28

selections based on the same proportions found in the general population.
Let’s say the number of females ages 36-40 are the same as those in the 41-43 age range.
10 females from each age range.
Not random but sample is stratified well.

gathering representative data from a group. Application of quota sampling ensures that sample group represents certain characteristics of the population chosen by the researcher.

Question 29

Non-probability sampling methods: Purposive

Answer 29

select participants for a particular purpose.
Knowledge of group is required before selection.
Risk of researcher selection bias.

Question 30

Non-probability sampling methods: Snowball

Answer 30

existing study subjects recruit future subjects from among their acquaintances. Thus the sample group is said to grow like a rolling snowball