HAN Exam 4- chapters 6-7

QUAL Research

• RQ explores why or how a phenomenon occurred

• develop theory

• describe an individual’s experience

• data collected over long period of time

Theoretical Unpinning

• guides researcher thru all 5 stages of research study

• almost all qual studies have

  • grounded theory is an exception

QUAL Research Designs

• Case Study

• Focused Ethnography

• Phenomenological (lived experience)

• Grounded Theory

Types of Case Study

• intrinsic

• instrumental

• collective

Case Study Design

• in-depth look of 1 person, event, or program to understand/illustrate changes over time

• answers how and why, with attention to context

Features of a Case Study

• sample always 1

• multiple types of data collected (triangulation) and analyzed

• can collect quan and qual data (not considered mixed methods)

Intrinsic Case Study

• exploring/explaining irl issues + how they’re unique

• only have 1 patient

Instrumental Case Study

• Select 1 case from a group of cases that when explored will help develop a better understanding of the irl issue

Collective Case Study

• combo of several similar cases to create a single case (fake case)

  • allows for in-depth understanding of issue

  • more transferable (“good fit”)

Focused Ethnography

• study specific health-related belief/issue/practice within a culture

• study all forms of communication

• understand day-to-day behavior

• data collection ranges from weeks to months

• has gatekeeper and key Informant

Gatekeeper

• used to gain entrance into group

• guides researcher and makes introduction to the group to build trust

Key Informant

• one with special knowledge/relationship of the group

• used by researcher to explain certain aspects of group’s behavior/communication

• helps researcher control bias and misinterpretation

Features Of Focused Ethnography

• sampling: gatekeeper

• data collection:

  • key informant

    • can’t use info from them as data, but can be used as clue

    • collect data thru triangulation (have to use triangulation if key informant is used)

• field notes: observational data

• data analysis:

  • any of the 29 data analysis methods

Phenomenological Design

• also called lives experiences

• gain in-depth understanding of participants’ lived experiences (purposefully selected)

Phenomenological Design Data Analysis

1. immersion

2. Reflective Bracketing

3. first cycle coding

4. memo keeping

5. second cycle coding

6. data saturation

7. member checking

• reflective bracketing

• member checking

• low inference data

Phenomenological Design Features

• sampling:

  • purposive

  • snowball

  • convenience

  • 5-25

• data collection:

  • relies heavily on interviews

  • data collected from first person perspective

Reflective Bracketing

• researcher intentionally put aside personal biases to let participants’ experiences guide the research process

Low Inference Data

• direct quotes from participant to explain codes and themes

Grounded Theory Design

• develop a theory about a irl “process, action, or interaction, shaped by participants’ pov

  • understand decision-making process

  • how situation unfolded

  • strategies used

  • explain why/how something occurred

• very little previous peer-reviewed lit

Grounded Theory Design Features

• doesn’t use theoretical underpinning

• sampling:

  • theoretical

  • 20-50+

• data collection:

  • use variety of methods

• data analysis:

  • uses first and second cycle coding

Experimental Design Types

• Solomon 4 group

• pretest/posttest

• posttest only

• within subject

• control group: time series

Quasi-Experimental Design Types

• simple time series

• pretest/posttest (nonequivalent)

• posttest only

• within subject

• control group: time series

Experimental Design

• variables are manipulated by investigator

  • identifies cause-effect relation

• uses probability sampling methods

• control group is needed

  • random assignment of control and experimental group

• most rigorous

• greatest degree of internal validity

• controls for selection bias + all threats to internal validity

Internal Validity

• how well a study controls outside factors so results are only due to intervention

Quasi-Experimental Design

• variables manipulated by investigator

  • seeks to identify if intervention is effective

• uses nonprobability sampling

• control group is needed

  • no random assignment to control or experimental group

• doesn’t control for selection bias or history but controls for all other internal validity threats

Research Notation “O”

• observation or measurement

• point where I’m collecting data

Number following “O”

• indicates the study having several different observation

Research Notation “R”

• random assignment

  • probability sampling + random assignment to control or experimental group

Number following “R”

• indicates 2 groups

  • one is experimental group

  • the other is control group

Research Notation “X”

• treatment/intervention

• experimental group

• no x means it’s a control group

Solomon Four-Group Design

• tests the efficacy of a treatment while controlling for pretest effects

  • highest level of controls for threats to internal validity

    • history + maturation

• most rigorous in quan research

• combines pretest-posttest design with posttest only design

Solomon Four-Group Design Visual Rep.

• R1: O X O

  • pretest posttest observation + intervention

• R2: O __ O

  • pretest posttest observation, no intervention

• R3: __ X O

  • intervention + posttest observation

• R4: __ __ O

  • only posttest observation

• R1 + R2 are pretest posttest control group design

• R3 + R4 are posttest-only control group

• R1 and R3 same score = pretest didn’t influence results

  • diff score =pretest impacted results

• R2 and R4 same score = pretest didn’t influence results

  • R2 better score than R4 = pretest influenced results

Pretest-Posttest Control Group Design

• made up of 2 randomly sampled and assigned groups

  • both groups receive observations before and after interventions

    • only experimental get intervention

• controls for many internal validity threats

• tests for efficacy of intervention

• any differences in the posttest should be due to intervention

• possible threat to external validity:

  • pretest effect

Pretest-Posttest Control Group Design Visual Rep.

• R1: O X O

• R2: O __ O

Posttest-Only Control Group

• simplest experimental design

• made of 2 randomly sampled and assigned groups

• goal:

  • if difference can be determined between the 2 groups in the posttest observation

  • assess cause-effect relation

  • tests for efficacy of intervention

• controls for many internal validity threats

Posttest-Only Control Group Visual Rep.

• R1: __ X O

• R2: __ __ O

Within Subjects Design

• also known as repeated measures design

• same subjects receives 1+ intervention

  • their performance is repeatedly measured

• can’t use this if effect of intervention last a long time

• compares efficacy between 2 treatments

Quasi-Experimental Within Subjects Design Visual Rep.

• Pretest-posttest:

  • G1: O X1 O

  • __ : O X2 O

• posttest-only:

  • G1: __ X1 O

  • __ : __ X2 O

Within Subjects Design Strengths

• doesn’t require large sample

• controls for internal threats:

  • history

  • maturation

• help reduce errors associated with individual differences between subjects

Within Subjects Design Weaknesses

• carry-over effect

• fatigue

• practice effects

Experimental Within Subjects Design Visual Rep.

• R1: X1 O1

• __ X2 O2

  or

• R1: O X1 O

  __ O X2 O

Carry-Over Effect

• when exposure to one intervention impacts the performance of participants for next intervention

Fatigue

• bored

• exhausted

• lack interest after multiple tests

Practice Effects

• improvements in one’s performance on a test due to repeated exposure

Nonequivalent (Pretest-Posttest) Control Group Design

• also called quasi-experimental pretest-posttest design

• have 2 groups where participants weren’t randomly assigned

Nonequivalent (Pretest-Posttest) Control Group Design Visual Rep.

• G1: O X O

  - - - - - - - - - - -

• G2: O __ O

• dotted line means groups weren’t randomly assigned

Nonequivalent Posttest-Only Control Group Design Visual Rep.

• G1: __ X O

  - - - - - - - - - - - -

• G2: __ __ O

Control Group Time Series Design

• inclusion of control group

  • increases internal validity

• take multiple observations at baseline and end of study to determine if intervention lasted over time

Control Group Time Series Design Visual Rep.

Time

——>

Baseline data (observations before intervention)

G1: O O O O O X O O O O O

- - - - - - - - - - - - - - - - - - - - - - - - -

G2: O O O O O __ O O O O O

OR

G1: O1 O2 O3 X O1 O2 O3

- - - - - - collecting diff types of data - - - - - -

G2: O1 O2 O3 __ O1 O2 O3

OR

R1: O O O X O O O

R2: O O O __ O O O

• collecting same type of data 3x

Single Group Time Series Design

• several observations of DV (O) are taken over time

• can only be quasi-experimental since no control group

• study impact of intervention on a single group

Single Group Time Series Design Visual Rep.

Time

——>

O O O O O X O O O O O

Experimental Researchers

• use combo of descriptive and inferential stats

• analysis determines if null hyp. can be rejected

• determine cause-effect relation

• results generalize back to pop.

Quasi-Experimental Researchers

• draw inferential conclusions about study

• limited conclusion

  • results can’t determine causality or generalizability