Chapter 10: Non-Experimental and Quasi Experimental Strategies

Question 1

T/F Non experimental and quasi experimental research designs can establish cause and effect relationships

Answer 1

false. There is not enough control and there is no manipulation of an independent variable in these types of studies. They cannot unequivocally establish a cause and effect relationship

Question 2

Difference between non experimental and quasi experimental studies

Answer 2

non experimental: makes no attempt to minimize threats to internal validity

quasi experimental: some attempt is made to minimize threats to internal validity.

Both involve comparing groups or conditions

Question 3

Both non and quasi experimental studies involve _____ groups or conditions

Answer 3

comparing groups or conditions

Question 4

What is a quasi-independent variable?

Answer 4

a variable that differentiates the group but is not manipulated by the researchers actively

Question 5

What is the dependent variable in non and quasi experimental research strategies?

Answer 5

the scores obtained

Question 6

T/F In non/quasi experimental designs, different group conditions are created by manipulating an independent variable

Answer 6

FALSE: different groups or conditions are not created by manipulating an independent variable, but are instead defined in terms of a PRE-EXISTING PARTICIPANT VARIABLE (ex/ college vs no college), or IN TERMS OF TIME

Question 7

2 Main categories of non and quasi experimental designs

Answer 7

1) between-subjects designs: non equivalent group designs

2) within-subject designs: pre-post designs

Question 8

general problem with between subjects designs in non/quasi experiments:

Answer 8

the problem with between subjects designs are INDIVIDUAL DIFFERENCES between participants that are often mitigated using RANDOM ASSIGNMENT OR MATCHING PARTICIPANTS or HOLDING A VARIABLE CONSTANT.

THIS MITIGATION DOES NOT HAPPEN in non or quasi experiments

Question 9

What is a non equivalent group design? purpose?

Answer 9

a design for non and quasi experimental designs. a study where groups of participants are already formed and thus the researcher cannot control assignment to groups. (groups therefore considered to be NON EQUIVALENT)

purpose: to show how the factor that the groups differ between each other is responsible for causing the differences between scores of the 2 groups.

Question 10

what is a major threat to validity for non equivalent group designs? example

Answer 10

Assignment bias: the pre-existing groups may have been produced with different participant characteristics.

ex/ studying 2 high school’s academic performance with the independent variable being their cell phone policy (allowed vs prohibited) the 2 schools may differ in terms of their IQ, socioeconomic levels etc. Potential confounding variables that provide alternative explanations for student performance

Question 11

What’re the 3 types of non-equivalent group designs and what main category does each fall into?

Answer 11

1) differential research design: a nonexperimental subjects design
2) post test only non-equivalent control group design: a non experimental subjects design
3) Pretest-posttest non-equivalent control group design: QUASI

Question 12

What is a differential research design and what is it’s goal? provide an example

Answer 12

differential research design is a non equivalent group design that utilizes BETWEEN-SUBJECTS aspects. the INDIVIDUAL DIFFERENCES are of primary INTEREST and it compares pre-existing groups where the participant characteristics define groups.

goal: to establish differences between pre-existing groups on the dependent variable

ex/ the effect of sex on self esteem: one group is boys, the other is girls (individual differences are of interest), and then you measure self esteem (dependent variable)

Question 13

differential research design is also known as ex post factor design because:

Answer 13

the research involves looking at differences after groups have already existed

Question 14

compare and contrast differential research designs with correlational studies

Answer 14

similarities: both compare 2 naturally occurring variables with no controls
- both establish the existence of a relationship but no causal inferences

differences:
differential research designs:
- participant differences DEFINE THE GROUP
- measurements made within each group
- GROUPS are compared via PROPORTIONS or MEANS

correlational studies:

• single group used
• measures 2 or more variables for each individual
• compares VARIABLES via pearson r values.

Question 15

What is a POST TEST ONLY non equivalent research design and what is it’s goal? provide an example

Answer 15

POST TEST ONLY non equivalent research design is a type of between subjects non-experimental design that uses PRE-EXISTING groups, one that serves as a CONDITION, and the other as a CONTROL

goal: to evaluate the effectiveness of a treatment administered, by comparing with scores of the nonequivalent group that has not received the treatment.
example: using 2 highschools are determining the effect of sex ed and pregnancy rates. One school gets sex ed and the other doesn’t (nonequivalent control group). then you measure the pregnancy rates at the school (post treatment)

Question 16

in a POST TEST ONLY non equivalent research design, the dependent variable/treatment is measured _____

Answer 16

post test

Question 17

Why is a POST TEST ONLY non equivalent research design also called a static group comparison design?

Answer 17

because observations are only done post test, and one of the non equivalent groups acts as a control.

Question 18

T/F POST TEST ONLY non equivalent research design can establish causal effects

Answer 18

False. it may appear to show cause effect relationships because measurements are compared after the treatment conditions, but this type of experiment DOES NOT PROTECT against ASSIGNMENT BIAS

Question 19

What is a PRE TEST POST TEST non equivalent research design and what is it’s goal? provide an example

Answer 19

a PRE TEST POST TEST non equivalent research design is a between subjects QUASI experimental design in which researchers observe both groups, give a treatment to one group, and then observe both groups again.

Question 20

how do pretest observations in PRE TEST POST TEST non equivalent research design combat against assignment bias?

Answer 20

assignment bias is not eliminated but it may be reduced because you have a general “BASELINE” of difference between groups that you know is not because of treatment (due to the fact that you observed them before the treatment was applied)

Question 21

in a PRE TEST POST TEST non equivalent research design , if participants are similar before treatment but different after treatment, the researcher can be more confident that ____

Answer 21

the treatment has had an effect.

Question 22

T/F: PRE TEST POST TEST non equivalent research design eliminates time related threats to internal validity

Answer 22

FALSE; although PRE TEST POST TEST non equivalent research design minimize time related threats because both groups are observed at the same times and thus experience the SAME TIME RELATED FACTORS, these factors are NOT ELIMINATED BECAUSE HISTORY effects may differ from one group to another. (DIFFERENTIAL EFFECTS)

Question 23

What are differential effects

Answer 23

when history effects might differ from one group to another. ie/ if you are studying 2 different high schools performance and their phone policies, one high school may have a worse history effect (their football team lost every game in the season), and may contribute to depression and thus decrease in academic performance.

Question 24

Difference between a PRE TEST POST TEST non equivalent research design and a PRE-POST DESIGN

Answer 24

In a pre-post design, observations are made over time for ONE GROUP of participants (WITHIN SUBJECTS DESIGNS) before and after a treatment, rather than comparing the treatment group to a control group in a PRE TEST POST TEST non equivalent research design .

Question 25

How many control groups are in a Pre-post within subjects design ?

Answer 25

there are no control groups for the treatment group to be compared to. Only one group is used in this study

Question 26

Goals of a pre-post design

Answer 26

to evaluate the influence of the treatment or event by comparing the observations made before treatment to the observations made after treatment.

Question 27

Threats to internal validity for pre-post designs

Answer 27

1) time related factors: history, regression to the mean etc.
2) sometimes assignment bias: the group is pre-existing and may have been assembled with specific participants
3) CANNOT COUNTER BALANCE: you do not manipulate a pre-post design and therefore ORDER EFFECTS will occur

Question 28

What’re the 4 types of Pre Post studies and what main category of non-quasi experimental designs does each fall under?

Answer 28

1) one group pretest-post test design (non-experimental)
2) time series design (Quasi)
3) Interrupted time series design (quasi)
4) equivalent time series/samples design (quasi)

Question 29

why is a one group pretest post test design a non-experimental research method?

Answer 29

because there is no attempt to limit threats to internal validity. there is still assignment bias because one observation for each participant in the same group is made before and after treatment (OXO).
- no limiting of threats to internal validity= no cause effect relationship established.

Question 30

a ____ is a quasi-experimental within group design that involves repeated measurements or series of observations for each participants over time

Answer 30

a time-series design (OOOOXOOOO)

Question 31

name an example of a time-series design

Answer 31

determining well being by looking at number of trips to school nurses before and after a traumatic event for a period of months.

Question 32

Why is a time-series within subjects design a Quasi-experimental design and not a non-experimental design?

Answer 32

because there is some effort to address potential threats to internal validity by using PRETEST OBSREVATIONS to note any TRENDS that ARE present in the sample BEFORE the treatment is introduced. If there are no trends that are noted in the pretest observations, than there is MORE CONFIDENCE that potential threats to internal validity are not influencing participants

Question 33

What is a main threat to internal validity in a time-series quasi-experimental within subjects design

Answer 33

history/time effects. History may be a threat to internal validity when event occurs at the same time as treatment–> did the treatment or did the outside event cause the increase in scores?

Question 34

What is an interrupted time-series design?

Answer 34

intervening event is not manipulated by the reserach. ex/ Natural disaster- may have to use archival data.

Question 35

What is an equivalent time-series design?

Answer 35

a design where a treatment is repeatedly administered and removed during a series of observations.
- series of events can be extending as long as the researcher wants (OOOXOOONOOOXOOO)

Question 36

What is an advantage os using an equivalent time series design compared to a time-series design?

Answer 36

it eliminates the threat of history as a threat to internal validity because if the treatment causes a change, A REGULAR PATTERN should be seen when you repeatedly administer the treatment. Even though there is a possibility of coincidental outside events (history) that may occur at the same time as the treatments, it is unlikely that it occurs every time.

IF A PATTERN OCCURS REPEATEDLY, THEN IT IS MORE CONFIDENT THAT TREATMENT IS RESPONSIBLE FOR CHANGE.

Question 37

IF A PATTERN OCCURS REPEATEDLY in an equivalent time series within subject-quasi experimental design, THEN I______ THAT TREATMENT IS RESPONSIBLE FOR CHANGE.

Answer 37

then there is more confidence that the treatment is responsible for change

Question 38

what is a single subject/single case design?

Answer 38

using a time-series pretest-post-test within subjects non experimental design and applying them to a single individual.

Question 39

What is a developmental research design? Goal?

Answer 39

study changes in behavior that relate to age (natural MATURATION PROCESS)
Goal- to describe relationship between age and other variables.

Question 40

3 types of developmental research designs

Answer 40

1) cross sectional
2) longitudinal within subjects design
3) cross sectional longitudinal design

Question 41

in developmental research designs, the independent/manipulated variable is:

Answer 41

the age of the participants (the age changes)

Question 42

What is a cross sectional developmental research design? is this an example of non experimental or quasi experimental design?

Answer 42

uses SEPARATE GROUPS (BETWEEN SUBJECTS) for each IV group (age group)
-NON EXPERIMENTAL DESIGN because there is no attempt to control internal threats to validity: the groups are NON EQUIVALENT- there are generational affects and individual differences between groups that are not just age.

Question 43

Advantages and Disadvantages of a cross sectional developmental research design

Answer 43

pros: 1) less time consuming (do not have to wait for groups to age to measure again)
2) does not require long term relationship between participant and reserarcher
3) less expensive
4) lower attrition rates

cons:
1) needs more participants than a longitudinal study
2) cannot determine how an individual develops over time (history and maturation)
3) cohort and generational effects

Question 44

what are generational/cohort effects

Answer 44

environmental factors that differentiate one age group from another- more problematic the more years between groups. this is a THREAT TO INTERNAL VALIDITY

ex/ the answers of gender identity questions in 60 year olds vs 20 year olds may be different because the mentality 60 years ago was that gender and sex is the same thing, and transgendered individuals were looked down upon. the 60 year olds may answer falsely.

Question 45

What is a Longitudinal developmental research design? i

Answer 45

a WITHIN SUBJECTS DESIGN that measures a DV in the same group of individuals over a period of time (REPEATED MEASURES)

Question 46

What is the treatment administered in a longitudinal study? (what is the independent variable?)

Answer 46

there is no treatment administered, the treatment is age. You invesigate the relationship between age and the variable of interest.

Question 47

Advantages and disadvantages of a longitudinal study

Answer 47

pros:
- need fewer participants
- no cohort effects
- able to determine and observe how an individuals behavior chnages over time.

cons:

• time consuming and thus expensive
• high attrition
• later scores may be partially effected by previous experiences (POTENTIAL ORDER EFFECTS) depending on length of time between treatments.
• changes of terminologies over time

Question 48

What is a cross secitonal longitudinal design? provide an example

Answer 48

1) compares results obtained from different samples (cross section) that were obtained at different times. (longitudinal design)

EXAMINES PHENOMENA OTHER THAN INDIVIDUAL AGING

ex/ how drug use and lifestyle have changed over the course of 30 years.