Webnr 2 Validity in Research

Question 1

What is internal validity

Answer 1

A causal relationship exists between the independent and dependent variables.

Question for interval validity is “ whether the treatment(s) caused the observed changes in the independent variable”.
In other words, “are there other (extraneous) factors that may be responsible for that change?”.

Question 2

THREATS TO INTERNAL VALIDITY

Answer 2

History
Maturation
Testing
Instrumentation
Statistical regression to the mean
Assignment (subject selection)
Subject attrition

Question 3

Give of an example of “history” that would change the internal validity say like the one subject was receiving PT during the course of the study while others weren’t

Answer 3

If event is unrelated to the treatment of interest occur during the course of the study and may plausibly change the dependent variable.
Any changes in the clinic would affect.
Examples:
The hypothetical historical events which are not under researcher’s control:
Less treatment session for two weeks because of snow.
A new therapist who works manually more.
……..
They would add, change, affect the result of the study.

Question 4

To minimize the effects of history:

Answer 4

Planning,
Use of a randomly selected control group,
Description of unavoidable historical events.

In the single subject designs:
Establishing pre-treatment baseline,
Withdrawing and reinstating treatment,
Rapidly alternating between two treatments, or between treatment and non-treatment.

Question 5

Give example of Maturation aka old age that would affect the internal validity

Answer 5

Changes within a participant caused by the passage of time.
Participants get older, more experienced, stronger, or bored, some patients get better, or get worse by time.

Examples:
Balance deficits will change with immobilization duration.
Swelling will decrease by time if the patient rests appropriately without any intervention

Question 6

To minimize effect of maturation effect wjhat are somethings to get before starting the baseline measurement

Answer 6

Use of a control group with random assignment.
To take multiple baseline measures of participants pre-treatment.
Stable baseline.
Treatment withdrawal or reversal.
Rapid treatment alternation.
If it is applicable, the researcher may design a study so that the total time span is not overly long

Question 7

Repeated Testing is also a threat to internal validity why????

Answer 7

Repeated testing itself is likely to result in changes in the dependent variable.
Improvement measurements on next test because of familiarization with the testing procedure.
Examples:
Patients can get tired because of repetitive MMT without rest.
Patients can answer some questions wrong if they are exposed the survey first time because maybe they don’t understand what some questions mean.

Question 8

How to get rid ot the “testing” threat don’t offer it as much duh!!! that way the patients can’t bullshit the answers.

Answer 8

Randomly selected experimental and control groups
Use of posttest-only design ???
Use of familiarization sessions with testing personnel or equipment, a pilot study for stable performance

Question 9

How is “instrumentation” a threat to the internal validity. Think like what if the instrument wears down.

Answer 9

When changes in measuring tools themselves are responsible for observed changes in the dependent variable.
Many tools requires calibration.
Apply a mathematical correction factor.
Researchers can be measuring tool.
Examples:
Hand-held dynamometer calibration after every 10 or 20 subjects?
Untrained evaluator.

Question 10

Statistical regression how is this a threat to internal validity.

Answer 10

It is associated with reliability of a test. If participants are selected based on extreme scores on a single administration of a test.
This events occurs even in the absence of intervention.
Extremely high scores tend to decrease, extremely low scores tend to increase, and scores that fall around the average tend to stay the same.

Example:
Differential effects of a specific exercise on weaker and stronger patients.

Question 11

Assignment (Subject selection) how to make better subjects the only thing I got out of this slide is that the subjects should be randomized whoop de fucking doo

Answer 11

Groups of participants are different from one another on some variable that is related to the dependent variable.
Control of assignment threats is most effectively accomplished through random assignment to groups within a study.

Examples:
Different ratio of literacy for survey study.
Two groups with different mean age on balance
study.

Question 12

How is subject attrition aka “drop out” rates affect the internal validity????

Answer 12

The participants are lost from the different study groups at different rates or for different reasons.

Examples:
Starting with 5 patients for two groups and ending with 5 and 2 participants because of less motivated participants were in the second group .

Question 13

To minimize “subject attrition” how would you does this think maybe test subjects only on one day.

Answer 13

Planning to minimize possible attrition.
Collecting information about the lost participants and about the reasons for the loss of participants.
Testing and treating on a single day.
Making adherence as easy as possible.
No problem if dropouts are equal in both groups

Question 14

The Hawthorne Effect what is it and how to deal with it. I would just not tell the patients what they are being measured on!!!! hidden cameras to see how they act when they are not being observed.

Answer 14

Change in research participant behavior is due to the fact that they know they are being observed.
Example:
Working harder than usual because the researcher watches the subject
Giving exaggerated answers for pain since the subject think they can get more therapy session.
You can avoid this problem by using non-reactive measures. For example, measure the DV in such a way that participants do not know what’s being measured, or perhaps even that they are being observed (one way mirrors, hidden cameras, ..)

Question 15

What is construct validity?????? (think like does the instrument measure what it is really measuring???

Answer 15

Concerns with the meaning of variables within a study.
Is a issue for all research study.
Whether the researcher is studying a “construct labeled” or a “construct as implemented”.
Example:
Active shoulder range of measurement for shoulder function:
Construct as implemented: AROM
Construct as labeled: Function

Question 16

What is construct underrepresentation.

Answer 16

If constructs are not developed fully within a study.

Examples:
Taking one measurement at the end: It doesn’t represent changes during the intervention (when exact change happens 10 minutes, 20 minutes, 30 minutes later, ???)
Taking measurement after a single intervention: It is not realistic to see difference after 1 session

Question 17

What is experimenter expectancys. Think like this is what I want to see so subject does exctly what the experimenter wants him or her to see.

Answer 17

Example:
Physical therapist cannot decide that exercise two session per day is effective or not effective: If the patient does exercise in the clinic properly, but doesn’t do at home properly but he/she says that I did properly

Question 18

Interaction between treatment sessions. This one maily deals with medications think like?? maybe patient takes beta blockers before taking blood pressure.

Answer 18

If treatments other than the one of interest are administered to participants.
Taking medication during taking other treatments.
Examples:
Taking muscle relaxant or pain killer before provocative test or therapy.
Taking blood pressure during 6 minute walking test for the subject who had beta blocker before test.

Question 19

Interaction between testing and treatment explain

REVIEW construct validity one more time via PPTS a little confusing

Answer 19

Control of interaction between treatment is difficult because of ethical consideration
Researcher can document who receives additional treatment, and can analyze data by subgroups

If the test can be considered an additional treatment
Strength test familiarization: 9 times lifting as test can be 9 times training session
To minimize: comparing a treatment group who pretested, a treatment group who was not pretested

Question 20

External vailidity. ( basically is how the information study can be generalized to a certain place, or a certain time, or certain group of people how well the study can be translated to real life scenarios.

Answer 20

Is concerned with to whom, in what settings, and what times the results of research can be generalized.
One can generalize the groups, settings, or times similar to the one studied.
One can generalize across groups, settings, or times if one has studied multiple subgroups of people, settings, or times.

External validity is influenced by the design of the study as well as by research consumer who hopes to use the study findings

Question 21

Threats to external validity (3)

Answer 21

Selection
Setting
Time

Question 22

Selection

Answer 22

If selection process is biased
Volunteers can represent the whole population???: Because they are willing to do something about dependent variable.
If there is mandatory recruitment to the study, the subject’s results can be generalized???

Question 23

How to control for selection?????

Answer 23

Carefully considering the target population to whom they wish to generalize results.
Selecting participants accordingly.
Carefully writing their research conclusions to avoid making inferences to groups or across subgroups who have nit been studied.
Selection issues: racial, ethnic, age, and gender

Question 24

Setting how is setting a threat??? think if you studied aquatic PT and said this could be generalized to all PT clinics but, most PT clinics don’t have a pool so you see :) and how to control for it????

Answer 24

Peculiarities of the setting in which the research was conducted make it difficult to generalize the results to other settings.
Example:
Using special treatment table for specific mobilization and generalizing this mobilization for all settings.
To control the threat:
Researchers can simulate as closely as possible the setting to which they hope to generalize their results.
Researchers can make their research setting as simple as possible.

Question 25

How can “time” affect external validity and how to control for it???

Answer 25

If the result of a study are applicable to limited time frames.
It can be controlled by:
authors through timely submission of research results for publication.
description of known changes that make research results less applicable than when the data were collected.

Question 26

Statistical conclusion validity

Answer 26

Addresses the statistical relationship of the independent variables and dependent variables.
Evaluated by examining the study’s effect size, sample size (statistical power), the appropriate use of statistical tests, and measurement reliability and variance

Question 27

RELATIONSHIP AMONG TYPES OF VALIDITY

Cumlative vs Recipcroal

Answer 27

The relationship between the validity threats can be either cumulative or reciprocal.

CUMULATIVE: Influences occur in the same way for both threats .
Randomly assign the groups: controls maturity and also history
RECIPROCAL: Controlling one threats cause a threats .
Small-standardized experiment for internal validity, effects generalization (external validity).