Experimental and Quasi-Experimental Design PDF

EXPERIMENTAL DESIGN Dr.(Mrs.) A.V.Raman Dean Omayal Achi College of Nursing Avadi, Chennai INTRODUCTION TO RESEARCH DESIGNS Quantitative research design Research design of a study spells out the basic strategies that researchers adopt to develop evidence that is...

EXPERIMENTAL DESIGN Dr.(Mrs.) A.V.Raman Dean Omayal Achi College of Nursing Avadi, Chennai INTRODUCTION TO RESEARCH DESIGNS Quantitative research design Research design of a study spells out the basic strategies that researchers adopt to develop evidence that is accurate and interpretable The research design incorporates some of the most important methodological decisions that the researchers make particularly in quantitative studies  ASPECTS OF QUANTITATIVE RESEARCH Intervention Comparisons Control of extraneous variables Timing of data collection Research sites and settings Communication with the subjects INTERVENTION In some studies nurse researchers want to test the effect of specific intervention on dependent variable (for e.g.. the effect of specific intervention on labor and neonatal outcome) COMPARISON The most common types of comparisons are 1. Comparisons between two or more groups (The hypothesis that the drug tamoxefin reduced the rate of breast cancer in high risk women could be tested by comparing women who received tamoxefin and those who did not) 2.In certain instances it is desirable to make comparisons for the same study participants (e.g. studying patients heart rate before and after intervention, Comparing low back pain for patients lying in two different positions) CONTROLS FOR EXTRANEOUS VARIABLES The complexity of relationships among human characteristics often makes it difficult to answer research questions unless efforts are made to isolate key research variables and control other factors extraneous to the research questions TIMING FOR DATA COLLECTION In most studies data are collected from subjects at a single point of time (for e.g. the subjects might be asked on a single occasion to describe the health promotion behavior) In some designs we contact the subjects in multiple time (e.g. to observe the growth pattern 3 monthly) Hence the researcher must decide on the number of data collection points needed to address the research question properly RESEARCH SITES AND SETTINGS sites are the overall location for research Settings are the more specific place where data collection will occur sites and settings to be selected so as to maximize the validity and reliability of th data in designing a study. The setting should be natural and avoid anxiety promoting or foreign to the subjects experiences COMMUNICATION WITH THE SUBJECTS  Decide how much information to be given to the participants  Full disclosure of the study?  Consent Oral or written?  Think about the reading and comprehension level of the subjects  Who will provide the information?  What additional questions from the subjects are expected?  Debriefing sessions?  EXPERIMENTAL DESIGN Science answers questions with experiments  EXPERIMENTAL DESIGN  An experiment is a scientific investigation in which observations are made and data are collected according to a set of well defined criteria. In an experiment, researchers are active agents not passive observers.  The procedures developed by physical scientists were profitably adopted by biologists during the 19 th century, resulting in many achievements in physiology and medicine.  DEFINITION Experimental method is an empirical research method used to examine a hypothesized causal relationship between independent and dependent variables.  DEFINE THE PROBLEM Begin by asking a question about your topic What is a good question for an experiment? One that is testable with the materials at hand  Now we need a hypothesis to guide our investigation. What is a hypothesis? Your best thinking about how the change you make might affect another factor. Tentative or trial solution to the question. An if ………… then ………… statement.  CHARACTERISTICS OF EXPERIMENTAL DESIGN A true experimental design is characterized by Manipulation, Control , and randomization  MANIPULATION Refers to the process by which the researcher manages the independent variable in order to study the effect on the dependent variable. EXAMPLE. Gentle massage is effective as a pain relief measure for elderly nursing home residents The Independent variable is gentle massage which could be manipulated by giving some patients the massage intervention and withholding it for others Then it is possible to compare the pain level (dependent variable) in the two group to see if differences in receipt of the intervention resulted in degree of pain levels  CONTROL The experiment must be organized so that the extraneous factors are prevented from operating and confusing the outcome which is to be appraised. Control is acquired by manipulating by randomizing, by careful preparation of the experimental protocols, and by the use of a comparison group or groups.  EXAMPLE If a researcher wanted to supplement diet of the low birth baby with a particular nutrient for 2 weeks, the weight gain at the end of 2 weeks would tell us nothing about the treatment effectiveness Suppose an average 1 kg. weight gain is noted does this gain support the conclusion that nutrition supplement (IV) caused the weight gain (DV) The answer will be NO? because babies normally gain weight as they mature Without a control group- a group that does not receive the nutritional supplements it is impossible to separate the effects of maturation from those treatment Therefore the term CONTROL GROUP refers to a group of subjects whose performance on a dependent variable is used to evaluate the performance of the experimental group on the same dependent variable  RANDOMISATION The term random essentially means that every subject has an equal chance of being assigned to any group. Randomization is the process that first ensures every unit in the target population has an equal chance of being chosen for the study sample, and then ensures that each unit in the study sample has on equal chance of being assigned to either the experimental or the control group.  EXAMPLE The effectiveness of a contraceptive health education programme for multiparous women. Two groups of subjects were included.One group was given intervention and other was not. The women in the sample are likely to differ in age, education, and attitude This would affect the woman's diligence in practicing contraception Although randomization is a preferred scientific method for equalizing the groups, there is no guarantee that the group will in fact be equal Therefore we need a procedure to consciously control of those characteristics of subjects that are likely to affect the outcome is called MATCHING (e.g. if matching were used in contraceptive education, then the researcher might ensure that if there was a married 38 year old woman with six children in exp. Group, then there will also be 38 year old woman with six children in control group The antecedent event in a proposed causal sequence is called the “independent variable”. The measured effect in the causal sequence is called the “dependent variable”. Eight classes of extraneous variables may especially interfere with research on human subjects, history, maturation, testing, instrumentation, statistical regression selection, mortality and interaction among these. History, the specific events occurring between the first and second measurement in addition to the experimental variables 2. Maturation, process within the respondents operating as a function of the passage of time per se (not specific to the particular events), including growing older, growing hungrier, growing more tired and the like 3. Testing, the effects of taking a test upon the scores of a second testing 4. Instrumentation, in which changes in the calibration of a measuring instrument or changes in the observers or scores used may produce changes in the obtained measurements 5. Statistical regression, operating where groups have been selected on the basis of their extreme scores 6. Biaser resulting in differential selection of respondents for the comparison groups ? 7. Experimental mortality, or differential loss of respondents for the comparison groups 8. Selection-maturation interaction etc., which in certain of the multiple group quasi-experimental designs, might be mistaken for the effects of the experimental variable.  PRINCIPLE GUIDELINES OF EXPERIMENTS The MAX-MIN-CON. By Kerlinger 1986. The abbreviation stands for the following. MAX-Maximize experimental variable MIN- Minimize error variance. CON-Control extraneous variance.  MAJOR STEPS IN EXPERIMENTAL DESIGN 1. Delineate the population or universe to be studied (i.e. the set of subjects or objects that share a common observable characteristic) 2. Select a sample from the population by random sampling 3. By random assignment, sub-divide the sample into two sub-samples 4.. Specify one sub-sample, the experimental group and other the control group 5. Before introducing the independent variable, observe and record all important characteristics of the two groups 3. Introduce the independent variable into the experimental group but withholds it from the control group 7. After introducing the independent variable, observe the dependent variable in both experimental and control group 8. Compare the changes that occur in the experimental group with those that may have occurred in the control group 2. Record the difference 10.Compare these values with statistically computed values that judge the significance of the difference, and indicate whether or not the observed differences could have occurred by chance.  TYPES OF EXPERIMENTAL DESIGN The investigator planning an experiment has many experimental design option to choose. Experimental designs fall into three major categories. b)True or classical experimental design c)Quasi-experimental design d)Pre-experimental design. TRUE OR CLASSICAL EXPERIMENTAL DESIGN TRUE OR CLASSICAL EXPERIMENTAL DESIGN True or classical experimental design includes six major designs. They are Pre-test___ post - test control group design Solomon four – group design After/ post - test __ only experimental design Factorial design Randomized block design Cross over design or repeated measures design.  T he most commonly used in nur sing studies, ar e discussed as nomenclatur e used and definitions developed by Campbell and Stanley ( 1963).  The standard notations that are helpful in under standing alternative experimental design are -  ‘x’ To denote the experimental manipulation. (Treatment or intervention)  ‘O’ To denote observations or measurement  ‘R’ To denote randomization  PRE-TEST POST-TEST CONTROL GROUP DESIGN Experimental Experimenta Post Pretest l Treatment group test Random assignme nt Control group Pretest Post test  EXAMPLE Effectiveness of antenatal nutrition education on pregnancy and labor outcome among primi para mothers attending out patient department of selected hospitals Madurai 2007-2009  2. Solomon four group design is a complex particularly useful in studies of developmental phenomena and permits the investigator to differentiate many effects. Experimental Experimental Pretest Post test Treatment group Random Control group Pretest Post test assignment Experimental Experimental Post test Treatment group Control group Post test  EXAMPLE Effectiveness of workshop on management of HIV to improve the nurses attitude toward patients with AIDS admitted in selected hospitals , Madurai 2006-2008 3. Post- test only control groups design this design, which is sometimes called after only control group designs, is composed of two randomly assigned groups, but neither of which is pre tested or pre measured in the before period of time. Post Experimental Experimental -test group treatment Random assignment Control Post group -test  EXAMPLE Effects of holding the newborns on paternal bonding behavior The researcher randomly assigned 36 first time fathers attending uncomplicated deliveries of normal infants to experimental and control group (Those who held and did not hold their infants at delivery)12-36 hours after the babies were born bonding behavior frequencies were recorded and observed the outcome 4. Factorial design In this design testing of multiple hypotheses is done in a single experiment In factorial experiments subjects are assigned at random a specific combination of conditions e.g. Does auditory stimulations have a more beneficial effect on the development of premature infants than tactile stimulation? (effect is weight gain and cardiac responsiveness)  EXAMPLE Type of stimulation AUDITORY TACTILE A1 A2 Daily 15 Min A1 B1 A2 B1 Exposure B1 30 Min. A1 B2 A2 B2 B2 45 Min. A1 B3 A2 B3 B3 1. Randomized Block Design There are two factors (independent variables) but one factor is not experimentally manipulated. e.g. Compare the effects of tactile versus auditory stimulation for female vs male infants. It can be structured as 2x2 experiment, with type of stimulation as one factor and gender as the other factor  Cross over design (or Repeated measures design) It involves the exposure of the same subjects to more than one experimental treatment e.g. Effectiveness of two alternative back rest position ( flat, horizontal vs. 30o elevation) on intracranial and cerebral perfusion pressures in brain injured adults PRE-EXPERIMENTAL DESIGNS The one –shot case study or single case study → The study design has a total absence of control, it is considered to be little value as an experiment. 2.One-GroupPretest-posttest design Only one group is observed before and after the independent variable is introduced. Loss of the control group decreases the usefulness of the study but may be necessary in cases where it is not possible or feasible to have control groups  EXAMPLE Effectiveness of selected nursing intervention on knowledge and level of satisfaction among clients with stroke  ADVANTAGES OF EXPERIMENTAL DESIGN 1.True experiments are the most powerful method for testing hypothesis of cause and effect relationship between variable because of the rigours, precision and control properties of experimental design. 2. This ‘if then’ type of relationship is important to nursing and medical researchers because of its implications for predication and explanation. 3.Though the control imposed by manipulation comparison and randomization, alternative explanation to a cause interpretation can be ruled out or discredited. DISADVANTAGES OF EXPERIMENTAL DESIGN 1. Some variables are not feasible or ethical to manipulate e.g., assessing pregnant women to take a new drug found to be dangerous to fetal development would not be ethically possible. 2. Randomization and otherwise equal treatment of control and experimental group can occur in a laboratory, but these conditions do not resemble what goes under real world conditions, and experimental findings can therefore be based on rather artificial circumstance 3. Experimental design attempt to reduce variable to measurable terms. Many of the phenomena that are important to science in nursing are complex, multidimensional and holistic, and defy the reductionism that has worked reasonably in the physical or natural sciences. QUASI EXPERIMENTAL DESIGN  DEFINITION It is a research design in which the researcher initiates an experimental treatment but some characteristics of a true experiment is lacking.  PURPOSES  To provide alternative means of examining causality in situations not conducive to experimental group.  To facilitate the search for knowledge and examination of causality in situations in which complete control is not possible. RANDOM ASSIGNMENT It is a procedure used to assign subjects to treatment or control groups randomly.  COMPARISON GROUPS Comparison groups are not selected using random sampling and do not receive the experimental treatment. Types of comparisons groups: Groups that receive no treatment. Groups that receive placebo treatment Groups that receive usual treatment. Groups that receive second experimental treatment. TYPES OF QUASI EXPERIMENTAL DESIGN. QUASI EXPERIMENTAL DESIGN Non equivalent Interrupted control group times series design design NON EQUIVALENT COMPARISON GROUP DESIGNS 1. ONE GROUP POSTTEST ONLY DESIGN It is referred to as pre-experimental rather than quasi experimental where manipulation of Independent variable is done and posttest is conducted to measure the outcome. Manipulation of Measurement of Dependent variable Independent variable Post test Treatment Example Outcome of Antenatal Education on pregnancy and labor.  2. POSTTEST ONLY DESIGN WITH A COMPARISON GROUP In this design a comparison group is introduced to determine the efficacy of the treatment and after manipulation of independent variable only posttest is conducted. Experimental Group Treatment Post test Non equivalent control group Treatment Post test Introducing Computer education to improve staff morale  3. One Group Pretest-Posttest Design In this design the experimental group is expected to serve as a comparison group. Pretest and posttest scores are used to find the efficiency of the treatment. Experimental Group Pre-test Treatment Post test Effectiveness pelvic floor muscles exercises in control of incontinence of urine. Effectiveness of infection control policy to decrease infection in postnatal ward. 4. PRE-TEST & POSTTEST DESIGNS WITH A COMPARISON GROUP In this design the comparison group is not randomly selected. The analysis is made on examining the differences between comparison group and experimental group pretest, between pretest and post test, between comparison group and experimental group posttest. Experimental group Pre- test Treatment Post-test Non equivalent control group Pre-test Post-test Example: Effectiveness of education on partograph in assessing labour progress among nurses working in labour rooms INTURRUPTED TIME SERIES DESIGN It is similar to descriptive time series design except that a treatment is applied at s given point in time and the outcome is measured at different point of time In simple interrupted time series design there is no control or comparison group Use of multiple methods to measure the outcome greatly strengthen the design. Experimental group Pre-test T Post test O(1)O(2) O(3) o(1) o(2) o(3)  EXAMPLE Effectiveness of selected nutritional interventions on growth and development pattern of children attending under fives clinic at Omayal Achi community health centre at Tiruvalluvar district 2006- 2008  Interrupted Time series designs with a comparison group Comparison group allows examination of differences in trends between groups after the treatment and of the persistence of treatment effects overtime Experimental group Pre-test T Post test o(1) o(2) o(3) o(1) o(2) o(3) Comparison group Pre-test T Post test o(1)o(2) o(3) o(1) o(2) o(3) ADVANTAGES & DISADVANTAGES OF QUASI EXPERIMENTAL STUDIES Advantages: From the point of view of controls for internal validity quasi experimental designs are thought to be superior to pre experimental design. Disadvantages: Cannot test causal hypothesis Do little to ensure external generalizability.