Quantitative Research Design

Criteria for research using quantitative design

  • Clarity and relevance of purpose
  • Researchability of the problem
  • Adequacy and relevance of the literature review
  • Match between the purpose, design, and methods
  • Suitability of the sampling procedure and the sample
  • Correctness of the analytic procedure
  • Clarity of findings

Elements of quantitative research design

The beginning stage of any investigation is to first decide WHAT is being investigated; and then decide HOW it is being investigated...You examine the problem, hypothesis, research question, concern, etc. and determine how it has been set into a working plan or blue print for study.

  • The blueprint that has been established for investigating the WHAT is the Research Design.
  • Blueprints involve decisions and these decisions present several aspects for the research consumer to consider.
  • Methods used to prepare for data collection 
    • Sources of research problems
    • Delimiting the problem
    • Establishing the significance of the problem

Types of quantitative research design

Purpose - The first aspect to consider is the PURPOSE OF THE STUDY. Why do I care about WHAT the answer is? The nature or purpose of the study depends on the existing level of knowledge in the study area.

  • The purpose helps to decide if a study is exploratory, descriptive, and/or conducted to test a hypothesis.
  • Methodological rigor and sophistication should increase as a researcher moves from exploratory to hypothesis testing. This increase implies generally more cost to the researcher.
  • Purpose or nature of the study is a critical decision for the researcher, depending on the study objectives and available resources, and to the consumer for determining the ramifications for implementing the results.

 

  • EXPLORATORY STUDIES 
    • Done when a researcher does not know anything about the situation or subject at hand.
    • There is no information about how similar problems were solved in the past.
    • A large amount of preliminary work must be done to increase research's familiarity with facets of the situation and understand what is happening before more rigorous designs can be done

**The first time that a study is done in a specific organization, etc. it does not mean that it is exploratory. Exploratory only exists when little or no previous study has ever been done and knowledge is minimal in a certain field of study.

  • DESCRIPTIVE STUDIES 
    • Done to determine and/or describe the variables in a situation.
    • Examples range from demographics to perceptions.

Hypotheses/variables

  • HYPOTHESIS TESTING 
    • Studies can be analytical or predictive in nature.
    • Both are designed to establish "cause leads to effect" relationships. 
      • Analytical Studies want to know why or how things are associated with; or contribute to a situation. 
        • Goes beyond describing variables.
        • The interest is in investigating relationships among variables in a specific situation with the intention of problem solving.
      • Predictive Studies go beyond analytical. 
        • Analyzes not only the what, why, or how of something; but also the what, why, or how something could happen in several situations with respect to some problem.
        • Aimed at broader generalizability.
  • COMPARISONS 
    • Retrospective/Prospective
    • Between-subjects/Within-subjects
    • Pretest-posttest (Before - after).

Type of investigation - The SECOND aspect is the TYPE OF INVESTIGATION. Causal, Correlations, Group differences Is the researcher trying to determine reasons or not? Is the study causal or noncausal. Is there control over the independent variable? 

Will there be an intervention or is the gathered/obtained information about currently existing phenomena. (Experimental, quasi-experimental, nonexperimental).

  • Causal Studies 
    • Done if its necessary to establish a "cause leads to effect" relationship.
    • Want to delineate and/or identify one or more factors that are undoubtedly causing the situation.
    • Intent is that you want to state variable x causes variable y.
    • With causal some variables have to be manipulated and others controlled. This means that it can be done in nature, but only with specific analytical techniques or if data is collected at different points in time (Cross-lagged correlations, Path analysis)
    • Ex. Does smoking cause cancer? 
      • Experimental
      • Quasi-experimental
      • Ex Post Facto
      • Meta-analysis
  • Noncausal Studies 
    • Used to identify important factors "associated with" problems = CORRELATION.
    • Because most times in any given situation there are multiple factors that influence each other and a problem in a chain like fashion it might be more important to identify the crucial factors that are associated with the situation.
    • Correlation or noncausal can be conducted in nature without the researcher or study itself messing with the variables.
    • Types of studies - Find statistical significance among groups on the variable of interest, etc. (T-test, Compare means).
    • Ex. Are smoking, ETOH, and chewing tobacco associated with cancer? If so which occurs with the most variance in the dependent variable.
    • Ex. There is a positive correlation between teachers salaries and liquor sales.
    • Other examples/types include 
      • Case study
      • Survey
      • Evaluations
      • Needs assessments
      • Methodological
      • Pilot studies
  • Field research 
    • Field methods - data-collection strategies, including observation, interviewing, case studies, and document review, that rely on "firsthand knowing" under natural conditions.
    • Field work - mode of scientific inquiry that immerses the researcher in processes of day-to-day life. The "field" is the social-psychological area where data are gathered to find answers.
    • Features that are evident in this method are the following: 
      • Researcher, through face-to-face interviewing or participant observation, is the primary "instrument" for data collection.
      • Data analysis and collection go on in the natural setting.
      • Investigator tries to learn about how variables vary under usual and unusual conditions rather than trying to control all variables except for the few under scrutiny.
      • Field researchers must make particular accommodations to ethical considerations.
    • Traditional roots (cultural anthropology, physical anthropology, archaeology, ethnology) 
    • Characteristics of field research 
      • "field" is social-psychological arena where investigator gathers data to find answers to the central area of inquiry
      • researcher is primary investigator and the "instrument" for data-collection.
      • data collection and analysis go on in the natural setting
      • some contrasts in methodology with traditional
    • Ethical considerations - great potential for conflict of interest, deception, exploitation, invasion of privacy, inconvenience to subjects, and loss of confidentiality
    • Major methods 
      • Observation
      • Analysis of documents and records
      • Case study
      • Case history - purpose is to use theory in understanding human experience
      • Interviewing - depend on respondent answers to questions about their experiences, thoughts, perceptions, etc.
      • Field studies and scientific truth

Control - The THIRD ISSUE is research control and concerns preventing/limiting influences/effects on the dependent variable to study the TRUE relationship between dependent and independent variables. This issue includes the extent to which the researcher would interfere with the normal flow of events. 

  • Research that does not deliberately attempt to change or manipulate variables; ie., work setting for a field study.
  • Data that is collected from several situations in a noninterfering manner = comparative study.
  • Mediating variables
  • External factors - data collection 
    • Minimize situational contaminants - constancy of conditions for data collection
    • Environmental control
    • Criterion variable may be influenced/affected by time
    • Protocols - treatment constancy (method(s) of interaction with subjects)
  • Controlling intrinsic factors (subjects) through extraneous variables 
    • Randomization
    • Homogeneity
    • Blocking
    • Matching
    • Analysis of covariance
  • Researcher interference 
    • Many studies can be conducted in their natural environment with minimal interference in the normal flow of events by the researcher.
    • When a researcher interferes deliberately or controls a normal flow of events in nature and manipulates certain variables to establish "cause leads to effect relationships = Experimental Design.
    • Ex. Deliberately increasing, decreasing, and maintaining pay to establish a cause/effect relationship between pay and performance.

Study setting- The FOURTH QUESTION involves deciding the setting for the study: Is it contrived or noncontrived?

  • Field studies are not generally contrived
  • It may not be possible to control contaminating factors in the work place; so the researcher makes a new setting away from the place where events normally occur (lab setting, lab experiment).
  • The researcher makes a decision on whether the study is done in the setting where events occur naturally; or, in an artificially contrived setting.

Sampling Unit of Analysis - The FIFTH CONCERN is deciding what the unit of analysis should be. Naturalistic versus laboratory

  • The unit of analysis refers to the level of aggregation or grouping of data during its subsequent analysis. (The population to be studied)Ex. In investigating the problem of how to raise the educational level of med/surg nurses. The interest is in the individual analysis and every employee response is treated as an individual data source.
  • If the interest is in studying groups of two's (like couples, twins), then dyads become the units of analysis. Ex. husband-wife interactions, supervisor-subordinate views.
  • If the study situation relates to group effectiveness then the level of analysis = group. Ex. ADN, DIP, BSN; compare different floors, depts, etc.

Time horizon- The SIXTH ISSUE raised is determining the time horizon for the study or the number of data collection points.

  • Cross-sectional study - A study is done that calls for data gathering at a single time (Do not get confused because most times that single data gathering extends over days weeks or months.
  • Longitudinal Study - A study of people or phenomena that occurs at several points in time to answer research questions. Ex. before and after
  • Identify various methods and mechanisms utilized in data collection. Approaches to research: Instruments, questionnaires, interviews, records, observation.

Data collection method - The SEVENTH ASPECT for consideration is the data collection method. This collection can occur from various settings and in many different ways (Sampling)

Validity

  • Threats to internal validity 
    • history
    • selection
    • maturation
    • mortality
    • Best designs - quasi-experimental, pre-experimental, correlational
  • External validity - generalizability

Quantitative research design

  • Working plan or blueprint for action after determining what to study, the question, purpose etc.
  • Begin design with overall picture and then narrow the focus to get bogged down in details before visualizing the end result.
  • Remember - there are always restraints on achievement of quality 
    • Need detailed attention to research design, can evaluate highest or lowest levels of quality possible
    • Assess study's realities of time, money, people, and other resource constraints