Quantitative Research

Quantitative Research

Traditional, positivist, scientific method which refers to a general set of orderly, disciplined procedures to acquire information. Uses deductive reasoning to generate predictions and Gathers empirical evidence.

experimental or non-experimental.

Quantitative Research can be classified as

Selection Bias, Maturation, History, Instrumentation Change, Mortality (Attrition), Testing Effect

Threats to Internal Validity

Selection Bias

Occurs when participants are not randomly selected, causing subjectivity and limiting objectivity in results.

Maturation

Natural physical, emotional, or psychological changes in participants over time may influence outcomes, especially in long-duration studies.

History

Unexpected external events occurring during the study may affect participants' behavior and distort results.

Instrumentation Change

Using different measurement tools or altering instruments mid-study can produce inconsistent or invalid data.

Mortality (Attrition)

Participants dropping out, transferring, or dying during the study may affect the comparability of results.

Testing Effect

Exposure to a pretest may influence participants' posttest performance, as they might remember previous answers or become sensitized to the test.

Experimenter effect, Hawthorne effect, Measurement effect

Threats to External Validity

Experimenter Effect

Characteristics or behavior of the researcher influence participants' responses, resulting in unnatural or biased data.

Hawthorne Effect

Participants alter their behavior because they know they are being observed as part of a study.

Measurement Effect (Reactive Pretest Effect)

The pretest itself sensitizes participants, influencing how they respond to the treatment and affecting generalizability.

Experimental Designs

These involve active manipulation of the independent variable and often include random assignment to groups.

True Experimental Designs

Characterized by randomization, control group, and manipulating an independent variable (the cause) and measuring its effect on a dependent variable (the effect)

investigate cause-and-effect relationships.

Primary goal of experimental designs

"method of difference" principle

True experimental designs operates on this principle where the effect of a single variable can be isolated and assessed.

Intervening/Extraneous Variables

These are other variables that can influence the results and are considered limitations. They are known as threats to internal and external validity.

Pretest-Posttest Control Group Design, Posttest Only Control Group Design, Solomon Four-Group Design, Factorial Design, Randomized Block Design, Crossover Design

Basic Experimental Design:

Pretest-Posttest Control Group Design

Both groups receive pretest → experimental group receives intervention → both groups receive posttest.

Posttest Only Control Group Design

No pretest; groups are randomly assigned → intervention is given only to the experimental group → posttest is done.

Solomon Four-Group Design

Combines pretest-posttest and posttest-only designs to control for testing effects. Subjects are randomly assigned to 4 groups. 2 of the groups (one experimental and one control group) are pretested while the other two are not. A posttest is given to all four groups

Factorial Design

An experimental design that manipulates two or more independent variables (IVs) simultaneously. This allows researchers to test not only the main effect of each IV but also the interaction effects between them.

Randomized Block Design

A design that uses a blocking variable (a pre-existing characteristic not manipulated by the researcher) to form homogenous groups ("blocks") before random assignment. This controls for the influence of a known extraneous variable. There are two factors, but only one is experimentally manipulated; the other is a stratifying variable. First, subjects are divided into blocks based on the blocking variable (e.g., gender, disease severity). Then, within each block, subjects are randomly assigned to the experimental conditions.

Crossover Design

A within-subjects design where the same participants are exposed to all different experimental treatments, but in a randomized order. Achieves a high degree of equivalence between groups because the same subjects are compared against themselves under different conditions. High statistical power and requires fewer subjects. Risk of Carryover Effects.

Risk of Carryover Effects

The effects of the first treatment influence the response to the second treatment.

Counterbalancing

Used in crossover design to rule out ordering effects by systematically varying the sequence of treatments across subjects.

Quasi-Experimental Designs

used when an experiment is desirable, but random assignment is not possible for practical or ethical reasons. Lack full randomization but still involve manipulation of the independent variable. An intervention is present, but random assignment is absent. Researchers use other methods to compensate for this lack. Balances internal validity with external validity (applicability to real-world settings).

Nonequivalent Control Group (NECG) Before-and-After Design, and Nonequivalent Control Group (NECG) After-Only Design

Common Quasi-Experimental Designs:

Nonequivalent Control Group (NECG) Before-and-After Design

the strongest and most common quasi-design. Both an experimental and a non-equivalent comparison group are observed before (O1) and after (O2) the intervention. The pre-test allows researchers to check how similar the groups were at the start.

Nonequivalent Control Group (NECG) After-Only Design

A much weaker design where two groups are compared only after the intervention. There is no pre-test to establish initial group equivalence, making it very difficult to attribute any differences to the intervention itself. It is often considered pre-experimental.

True Experimental Design

This is the "gold standard" for establishing cause-and-effect relationships. Random assignment of participants to experimental and control groups. Tests clear, pre-defined hypotheses. The researcher has full control to create/manipulate the independent variable. Procedures and measures are tightly controlled

Pre-Experimental Designs

These designs are considered very weak and provide little evidence for cause-and-effect. An intervention is present, but there is no random assignment and no proper control group to rule out alternative explanations.

The researcher has very little control over the research environment.

Key Weakness of Pre-experimental design

One-shot Case Study, One-group Pretest-Posttest Design

Pre-Experimental Designs

Time Series Designs

This design involves collecting data at multiple points both before and after an intervention to establish a trend. Multiple observations are taken before and after an intervention to analyze the pattern of change.

To determine if an intervention causes a change that is significant compared to a pre-existing trend.

Primary Goal of time series designs

Basic Interrupted Time Series, Time Series with a Non-Equivalent Control Group, Time Series with Withdrawn and Reinstated Treatment, Time Series with Multiple or Intensified Treatment

Types of Time Series Designs:

Basic Interrupted Time Series

The simplest form of time series, with a series of pre-test and post-test observations separated by a single intervention.

Time Series with a Non-Equivalent Control Group

Time series design that Strengthens the basic design by adding a control group that is also measured over the same time period but does not receive the intervention.

Time Series with Withdrawn and Reinstated Treatment

The intervention is introduced, withdrawn, and then reintroduced. If the dependent variable improves during intervention phases and worsens during the withdrawal phase, it provides strong evidence for the intervention's effect.

Time Series with Multiple or Intensified Treatment

The intervention is applied multiple times or its intensity is increased over the study period to observe a cumulative or dose-response effect.

Single-Subject Experiments (N-of-1 Studies)

This design applies the logic of time series analysis to a single individual or a very small number of subjects, commonly used in clinical and behavioral studies. An individual serves as their own control. Data is collected repeatedly during a baseline phase (A) and an intervention phase (B).

AB Design, ABA Design (Withdrawal), ABAB Design (Reversal)

Common single-subject experiments

AB Design

The basic structure involving a baseline (A) and an intervention (B).

ABA Design (Withdrawal): Baseline)

Baseline (A) -> Intervention (B) -> Withdrawal/Return to Baseline (A). This strengthens the case for causality

Non-experimental

Used when variables cannot be manipulated because of ethical, practical, or biological constraints. Most nursing research falls under these because many human traits (e.g., age, ethnicity, birth weight) cannot be experimentally altered. This designs are observational, meaning the researcher does not intervene. This design applies the logic of time series analysis to a single individual or a very small number of subjects, commonly used in clinical and behavioral studies.

Analysis Studies, Causal-Comparative Research (Ex Post Facto), Descriptive Research, Relationship Studies / Correlational Cause-Probing Research, Prediction Studies, Historical Research, Natural Experiments, Path Analytic Studies

Types of Non-Experimental Studies

Analysis Studies

Focus on examining existing processes, activities, or documents. examples are Case analysis, Job and activity analysis, Document analysis

Causal-Comparative Research (Ex Post Facto)

Seeks to identify causes or antecedents of an already existing condition. The researcher does not manipulate variables but compares groups based on a naturally occurring difference. Begins with an observed outcome and looks backward for possible influencing factors. Begins with an observed outcome and looks backward for possible influencing factors.

Descriptive Research

Aims to describe, observe, and document aspects of a situation as they naturally occur. Identifies existing conditions, practices, attitudes, or relationships. Often used as the foundation for developing hypotheses or theories. Does not establish cause and effect, only "what exists.

Relationship Studies / Correlational Cause-Probing Research

Focuses on describing the relationship between variables without inferring cause. Appropriate when studying potential causes that cannot be manipulated ethically or practically.

Correlation

an association or relationship between two variables. Variation in one variable is related to variation in the other.

Correlational studies cannot establish causality because the researcher does not control or manipulate the independent variable. Many external factors might influence outcomes.

Importance of Correlational Studies

Descriptive Research

Aims to describe, observe, and document aspects of a situation as they naturally occur. Identifies existing conditions, practices, attitudes, or relationships. Often used as the foundation for developing hypotheses or theories. Does not establish cause and effect, only "what exists."

Univariate Descriptive Statistics

Focus on describing the frequency or distribution of a single variable (a behavior, condition, or characteristic). Even when multiple variables are measured, the goal is not to relate them, but to describe each variable individually.

Prevalence Studies, Incidence Studies, Relative Risk (RR)

Types of Univariate Descriptive Studies

Prevalence Studies

Measure existing cases of a condition at a given point in time.. Use cross-sectional designs. Useful for describing the burden of a disease in a population.

Incidence Studies

Measure new cases developing over a period of time. Require longitudinal designs. Focus on individuals free of the condition at baseline.

Relative Risk (RR)

Compares the risk of developing a condition between two groups. Shows how much more likely one group is to develop the condition than another.

Retrospective Studies

Studies that start with the outcome (DV) and look backward to identify potential causes (IVs) that occurred in the past. Often cross-sectional. Used when manipulation is impossible or unethical.

Prospective Non-Experimental Studies

Also called Cohort Studies in medical and epidemiological research. Begins with a presumed cause (independent variable) and follows participants forward in time to observe the presumed effect (dependent variable). Because there is no manipulation of variables, this design remains non-experimental, but it is stronger than retrospective studies because data are collected before the outcome occurs, reducing recall bias.

Prediction Studies

Conducted when there is existing knowledge or evidence that certain variables can forecast future behaviors or outcomes. Uses correlational techniques and statistical models. Assumes that some predictor variables influencing the outcome are already present and measurable at the time of prediction.

Historical Research

Used when answers to research questions are believed to lie in past events, records, or practices. Helps explain or shed light on current conditions, trends, or practices. Sources include archival documents, historical texts, old medical charts, oral histories, and eyewitness accounts.

Natural Experiments

Occur when a group is naturally exposed to a phenomenon with potential health or behavioral effects. Exposure is not controlled by the researcher; instead, circumstances assign exposure "at random."

Path Analytic Studies

A theory-driven method that uses advanced statistical modeling to test causal pathways among multiple variables. Part of structural equation modeling (SEM). Uses non-experimental data. To evaluate whether a set of relationships is consistent with a proposed causal model. Allows researchers to test complex interactions and mediating variables. May be done using cross-sectional studies and longitudinal studies, with the latter providing a stronger basis for casual interference because of the ability to sort out time sequence

Survey

gathers data from a large number of respondents at a particular point in time. It focuses on collecting self-reported information about attitudes, opinions, perceptions, characteristics, or behaviors.

Sample Survey

data collected from a subset of a population.

Group Survey

data collected from a specific group sharing characteristics.

Mass Survey

data collected from an entire population (e.g., national census).

Sample survey, Group survey, Mass survey

Survey Research based on whom data are collected

Telephone Surveys

Text Message Surveys

Snail Mail (Postal Mail)

Email Surveys

Face-to-Face Interviews

Survey Research based on Methods of Data Collection

Retrospective Surveys

Cross-Sectional Surveys

Longitudinal Surveys

Survey Research based on Time Orientation

Retrospective Surveys

The dependent variable (DV) is identified in the present. Respondents are asked to recall or report past events (independent variables). Useful when historical data are needed but not recorded.

Cross-Sectional Surveys

Data collected at a single point in time. Participants are at different stages of an experience or condition. Assumes that differences across groups approximate changes over time.

Longitudinal Surveys

Data are collected from the same respondents over multiple time points. Tracks changes, transitions, and developments within individuals. More time-consuming and costly; stronger for inferring time sequence.

Descriptive Survey, Comparative Survey, Correlational Survey, Evaluative Survey

Types of Survey Designs

Descriptive Survey

Describes characteristics, attitudes, behaviors, or conditions of a population. Provides a "snapshot" of what exists naturally. No manipulation of variables. Used for: Theory development, Identifying problems, Supporting clinical practice

Comparative Survey

Compares two or more groups regarding selected variables under natural conditions. Groups differ naturally

Correlational Survey

Examines the relationship (direction and strength) between variables. Explores how changes in one variable correspond to changes in another. Does not establish causal relationships.

Evaluative Survey

Assesses the value, effectiveness, or worth of a program, service, intervention, or policy. can be Formative (process evaluation - during implementation) or Summative (outcome evaluation - after completion)