Chapter 10 - Final Flashcards

1
Q

Exploratory studies

A
  • “What”
  • Can be qualitative or quantitative
    Example: What is the relationship between affect (mood) and neurotransmitters?
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Descriptive Studies

A
  • “How”
  • Generally Qualitative
    Example: How do individuals deal with depression deal with their disorder?
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Explanatory Studies

A
  • “Why”
  • Can be qualitative or quantitative
    Example: Why do SSRIs alleviate depression?
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Cross-sectional Studies

A

Data is only collected once

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Longitudinal Studies

A
  • Data is collected at different points in time
  • Panel studies (panel = same participants)
  • Trend studies (different participants of the same population)
  • Cohort Studies (usually based on the same event)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Idiographic Research

A
  • In-depth as they apply to one person
  • Often qualitative using case studies and interviews
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Nomothetic Research

A
  • Often quantitative
  • Requires establishing 3 criteria:
    i. correlation
    ii. time order
    iii. non-spuriousness
  • Casualty in research
    i. ability to determine causality = Internal Validity
    ii. prerequisite for determining causality:
    a. Know that behavior changed along with the treatment
    [covariation, correlation]
    b. Know that treatment came before the behavior change
    [Temporal sequence, time order (temporal contiguity, precedence)
    c. Know that everything but the treatment stayed the same
    [elimination of confounds, control, and non-spuriousness]
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Experimental Designs

A
  • Randomized One-Group Posttest-Only Design
  • Randomized Posttest-Only Control Group Design
  • Non-Random posttest only control group design
  • Non-Random pretest posttest control group design
  • Non-Random one-group pretest posttest design
  • Solomon Four-Group Experimental Design
  • Non-Experimental or Pre-Experimental Designs
  • Quasi-Experimental Designs
  • Non-Random cross-sectional survey design
  • Longitudinal Cohort
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Anatomy of a design name

A
  • Is there Random Assignment to conditions?
    a. Yes = Randomized
    b. No = Non-randomized
  • Is there a pretest? Posttest?
    a. Pretest only = pretest-only
    b. Posttest only = posttest-only
    c. Both = pretest posttest
  • Is there a control group?
    a. Yes = Control Group
    b. No = One-group
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Non-Experimental Designs

A
  • Quasi-Experiments
  • Non-experimental Designs
  • Small-N (Single-N) Designs
    *Purpose – To minimize threats to Internal Validity in situations where treatments can’t be randomize
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Solomon four-group experimental design

A
  • 4 control groups (randomly assigned)
    a. 2 pretests (2 groups do it and 2 groups do not) TREATMENT
    b. no experiment
    c. 4 posttest
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Why Quasi?

A
  • lack of control (comparison groups)
  • no random assignment
  • naturally occurring groups
  • ethnical reasons (can’t assign people to get sick)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Comparison across groups

A

Non-equivalent group designs

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Comparison across time

A

Time series designs

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Comparison across groups and time

A
  • Time series with non-equivalent groups designs
  • One-Group Posttest-Only Design
  • Posttest-Only Design With Nonequivalent Groups
  • One-Group Pretest Posttest Design
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

One Group Posttest-Only Design

A
  • No use in determining an effect
  • Similar in some ways to a case study, except in a case study, a lot is known about one subject to help put results in perspective.
17
Q

Posttest-Only Design With Nonequivalent Groups

A
  • Differences could be due to treatment or selection
  • Conclusions can be helped with additional information to gage the equivalency of the two groups.
18
Q

One-Group Pretest Posttest Design

A
  • Improvement in terms of selection
  • No counterbalancing is used to take the two observations close together in a time like a within-subjects design.
19
Q

Quasi-Experimental Designs

A
  • Nonequivalent Control Group Design With Pretest and Posttest
  • Interrupted Time-Series Designs
  • Statistical Analysis of Quasi-Experiments
20
Q

Nonequivalent Control Group Design With Pretest and Posttest

A
  • Most common in Social Science
  • If O1 = O1 Groups are relatively equivalent
    a. Selection and Regression are minimized
  • If O1 = O2 for the control group
    b. History and Maturation are minimized
  • Allows for assessment of mortality
  • Selection interactions are still a risk, especially if pretest differences exist
21
Q

Nonequivalent Control Group Design With Pretest and Posttest Variations

A
  • Use of a proxy pretest
    a. A measure of variables that correlate with posttest
  • Separate pretest and posttest samples
    b. Used if testing can be a threat
  • Pretest observations at more than one time intervals
22
Q

Interrupted Time-Series Designs

A
  • No problems with selection or interactions with selection
  • Testing or regression issues should disappear with the multiple pretests
  • Maturity may show as a trend
23
Q

Interrupted Time-Series Design Variations

A
  • Addition of nonequivalent no-treatment control group time series
    a. Allows assessment of history as a threat
  • Interrupted time series with removed
    treatment
    b. Can be used if treatment effects are
    reversible
  • Can be repeated to produce multiple replications (similar to baseline designs)
  • Interrupted time series with switching replications can counter or help assess most threats to internal validity
24
Q

Statistical Analysis of Quasi-Experiment Advantages

A
  • Allows us to do research which may not be possible or ethical otherwise
25
Q

Statistical Analysis of Quasi-Experiment Disadvantages

A
  • When threats to internal validity are found, it may invalidate your results
  • More complex than an experiment