Quantitative Research Designs

Quantitative Research Designs

1. Experimental Research

2. Non-Experimental Research

3. Quasi-Experiments

Decide if variables will be MANIPULATED.

Experimental Research

- Randomized Controlled Trial (RCT)

- Researchers are active agents.

- Controlled experiments → causes & effects.

Properties of RCT

- Manipulation

- Control

- Randomization

Manipulation

Manipulate the Independent Variable by administering a treatment or intervention to some people and withholding it from others

Control

- The researchers introduce controls over the research situation

- A control group does not receive the intervention

- To have a counterfactual approximation

Counterfactual

what would have happened to the same people exposed to a causal factor if they simultaneously were not exposed to the causal factor

Randominzation

The researcher assigns participants to a control or experimental condition on a random basis

Blinding/Masking

- Are used in RCTs to prevent biases stemming

from awareness

- Involves concealing information from participants, data collectors, care providers, intervention agents, or data analysts to enhance objectivity and minimize expectation bias.

Types of Bias

- Performance Bias

- Detection or Ascertainment Bias

Performance Bias

Logic behind this: best behavior is presented by the person who is being observed. He/she may do extra effort to keep their best foot forward while being the best version of themselves.

Performance Bias

Exercised by the researchers; researchers are keen to details because they are aware, they may do extra strategies so that the results will be of advantage.

Performance Bias

Example: During face-to-face classes, whenever a teacher is scheduled for a visitation from PAASCU to evaluate their performance, the teacher comes to class well-prepared. She instructs the students what to do. She is no longer behaving naturally or normally.

Detection Bias/Ascertainment Bias

- Exercised by the study subjects

- Example: A study participant who knows that they are receiving the placebo drug against hypertension will do extra measures to keep their blood pressure low and prove the researchers wrong (out-perform them). The study participant will outperform them by exercising and doing other non pharmacologic measures outside the variables included in the research.

Ascertainment

You have learned or discovered.

Types of Closed Study

1. Single-Blind Studies

2. Double-Blind Studies

ONLY applicable to experimental research, not in any other QN research.

Single-Blind Studies

- Study participants do not know if they belong to the experimental or controlled group.

- Consent form does not contain all information about the study because it will defeat the purpose of preventing the study subjects from having biases or ascertainment bias.

- Only the researchers know who the experimental and control group is.

Double-Blind Studies

- Both the researchers/data implementers and the study subjects do not know who the experimental group is to prevent biases.

- An extra step to avoid biases.

- Only the primary investigator and the sponsor have

full knowledge.

Types of True Experiments

1. Basic Experimental

2. Solomon Four-Group

3. Factorial

4. Randomized Block

5. Crossover

Basic Experimental

a. Posttest only BE

b. Pretest-Posttest BE

Posttest Only

- You have 2 groups, one being the experimental group, and the other the controlled group.

- Group 1: Manipulated with interventions

- Group 2: Control Group

Posttest Only

- Both groups are assigned with study subjects randomly and both observed for changes through a posttest. The changes manifested in the experimental group will be compared with the controlled group, if it is significant or not.

- The results taken from group 1 is compared to group 2

- A test or assessment is done after intervention has been made

Pretest-Posttest

2 groups where study participants are randomly assigned. But this time, an observation is done PRIOR to introduction of an intervention.

Pretest-Posttest

Example: A study on the effectiveness of a newly developed therapy for anxiety. Since you are conducting a pretest-posttest experiment, you will assess or determine first the study participants' anxiety levels (both under the experimental and control group) before introducing the intervention (the investigational therapy) to the experimental group. Then do a posttest to check for significant differences between the two groups.

Pretest-Posttest

2 Observations are performed, 1) prior to intervention, 2) after the intervention

Solomon Four-Group

Considered to be the most prestigious design because it minimizes threats to internal and external validity and controls the reactive effect of the pretest.

Solomon Four-Group

Example: Therapy towards a group of people with anxiety, if a pretest has been done to me (anxiety level has been assessed), then I may exhibit ascertainment bias because I am already aware of what the study is about. We are trying to control the effect of a pretest.

Solomon Four-Group

Allows us to triangulate or look at the effectiveness or changes brought about the intervention as manifested by the experimental group (group 1, 3) in different perspectives in comparison with other groups (group 2, 4).

Solomon Four-Group

- Drawbacks: you need a big number of participants to conduct this research.

- In some source, 48 in each group to yield valid and reliable results

- Feasible if you do your research in a longer period of time

Basic Experimental

Basic pretest-posttest

Solomon Four-Group

- Repeats the basic pretest-posttest experiment design but adds 2 groups that are NOT PRE-TESTED

- All 4 groups will undergo randomization in the assignment of subjects. Two will be pretested; and two will not receive the interventions but ALL FOUR will participate in a

posttest

Factorial

Used when you try to compare the effectivity between two interventions within groups with different parameters.

Time

Example: A research conducted on preterm infants to check if whether auditory or tactile stimulation is more effective in improving the health status or overall well being of the study subjects while being confined in the NICU or neonatal intensive care unit.

What is the parameter?

Auditory stimulation and tactile stimulation

Example: A research conducted on preterm infants to check if whether auditory or tactile stimulation is more effective in improving the health status or overall well being of the study subjects while being confined in the NICU or neonatal intensive care unit.

What are the interventions?

Factorial

In this type of design you will be needing more study subjects because each cell on this matrix or table pertains to 1 Group

Factorial

Example: 1 Group = A1B1, another group = A2B1, and so on) - meaning a group of infants will be exposed to auditory stimulation for 15 mins; another group to the same intervention, but 30 mins, and another but for 45 minutes. Same as another set of group receiving tactile stimulation in different intervals or duration

Factorial

- Being a true experiment, manipulation, randomization and control are always observed

- Doesn't have to be 4 groups.

Randomized Block

- It is done to compare sexes. Same treatment with factorial design but this time, the comparison is based on the study participants’ sex

- Only done if you want to compare significant changes or effects to sexes

Crossover

- "swap"

- Both groups will be exposed to both interventions

- Randomization is observed in the assigning of study subjects.

Crossover

- A pretest (initial observation) is done before an intervention is produced.

- On the first round of Manipulation, each group will be exposed to different interventions and a posttest will be done to check for changes. Then comes the crossover (swap of interventions), wherein the group who received intervention A will now be exposed to intervention B and vice versa.

- After some time, another posttest will be conducted to check for changes.

Crossover

Drawback:

1. Since a posttest has already been done prior to crossover, the study subjects are already aware of what is to be observed in posttest 2. Meaning a bias may occur.

2. Study subjects under group 1 may talk with subjects under group 2 and they compare experiences prior the crossover

Crossover

- Mitigate drawbacks: participants must not be allowed to talk with each other.

- 3 Observations:

1) Initial - Posttest one & two

2) Between the initial observation and Post one, there is an intervention

3) Between posttest one & two, there is a crossover or swap

Quasi-experiments

- Controlled trials without randomization

- Some may even lack a control group

- Intervention in the absence of randomization

- Use the term comparison group than control group

Quasi-experiments

- There is scarcity in test subjects

- You are not in full control of the environment

- Situations where an experiments is not ethically or practically feasible

- Can work with smaller sample size than randomized trial

Types of Quasi-Experimental Designs

1. Nonequivalent control group

2. Time series

Nonequivalent control group

meaning study participants of group 1 and 2 are not equal

Nonequivalent control group

a. Posttest only NE

b. Prestest-Posttest NE

Posttest only NE

- Randomization is absent

- Posttest has been done after an intervention has been given, only to experimental group, none for comparison group

- Group 1 or Experimental or treatment group will receive the intervention, the Comparison group will not. Posttest is done to both groups to observe for changes.

Pretest-Posttest NE

- Randomization is not observed in both groups and may even differ in the number of participants under each group

- Both groups will undergo two observations, the pre-test and posttest

- In between, an intervention is done towards the experiment group while the other does not receive any intervention or treatment.

Time Series

- As the name suggest, it is done in a series of time or a period

- you may have 1 group or 2 groups

1 group time series

5 observations are done prior to the giving of an intervention and another 5 are done after the intervention. This is to check for the long-term effects or lingering effects within a time period after the intervention has been done.

Time Series

Observations prior to the intervention: to check for any natural yet significant changes that may occur on the study subject. This natural changes should be identified since this may be treated as confounding variables that may affect the results of the study after the intervention has been introduced

Time Series

As researchers, we want the significant changes observed in the post test to be attributed solely to the intervention; and that confounding variables identified in the pretest be explicitly stipulated in the results of the study. It doesn't merit disqualification or withdrawal from the study because of the scarcity of study subjects.

2 group time series

we have an experimental group and a comparison group. One receiving the intervention and the other, not receiving any intervention or treatment at all.