Techniques of research control

Techniques of research control
Research control eliminates the influences of extraneous variables in the study so that true relationship between the independent and dependent variables can be understood. It excludes the contaminating factors that might affect the relationship between the variables. For example, the researcher wants to study depression among post partum women; the dependent variable is level of depression and independent variable is past partum women. There are other possibilities which affect the level of depression among post partum women such as age, socio economic status and marital status. If a researcher wants to study true relationship between independent and dependent variable, the extraneous variable has to be controlled. The extraneous factors can be divided into two;

• External factors
• Internal factors

Control of external factors
The external factors to be controlled in research are environment, time, culture and so on. To control the environmental influence on data collection, the researcher has to select the environment in natural or artificial setting according to the problem to be studied. Depending on the research topic, the relationship between dependent and independent variable may be influenced by the time of day, month and year. In these cases, the researcher should ensure the constancy of time during data collection.
Control of internal factors
Controlling internal factors such as gender, hereditary is doubtful in research. There are a few methods formulated o control subject characteristics to protect the research topic from the contamination of extraneous variables.
Methods are;

• Randomization
• Homogeneity
• Matching

Randomization
It means the selection of participants randomly, without any conscious decisions by the researcher. Randomization of study participants controls all possible sources of extraneous variables. For example, random assignment of subjects to experimental and control group would equalize the distribution of variables such as age, gender and neutralizes the influences of extraneous variables.
Homogeneity
When randomization is not feasible, the researcher can select the subjects who are homogenous with respect to extraneous variables are included in the study. Homogeneity means the researcher selects the study participants based on important extraneous variables such as specific age group, gender, education and so on. The main limitation of the homogeneity is that results of the study can be generalized only to the type of subjects participated.
Matching
During case control study, the researcher selects the study participants based on relevant extraneous variables and randomly divided into case control groups. This technique used to form comparable groups in case control design.

Classification of research design based on purpose

Classification of research design based on purpose
Survey research
The researcher collects the data regarding incidence, prevalence and distribution of variables of with in population through survey. The survey reports are based on self reports and respondents answer to the questions raised by the researcher. The common method for data collection is personal, face to face interview, interview by telephone and self administered questionnaire. The survey research is suitable for extensive data collection among large population and can focus on wide range of topics at a time.
Evaluation research
It helps to evaluate a process, program, treatment or policy. It can be cross sectional, longitudinal or experimental/non experimental. Based on the aim of evaluation, it can be classified into three;

1. Process analysis
2. Outcome analysis
3. Impact analysis

• Process analysis

A process analysis evaluates the process of implementing a new program or procedure and assesses the functioning in actual situation.

• Outcome analysis

An outcome analysis evaluates the final goals of the program, which are achieved after an implementation of program or process.

• Impact analysis

Impact analysis evaluates the net effects or impacts of the program after implementation of program or process.
Outcome research
Outcome research evaluates the effectiveness of health care services in an organization. This research conducts usually as part of global assessment. This types of assessment helps to appraise and document the quality of nursing care in an organization.

Classification of research design based on time dimension

Classification of research design based on time dimension
The designs can be classified based on when and how often data will be collected in a study. The data can be collected at a single point in time or more than one occasion. 

There are four situations in which it is a appropriate to design a study with multiple points of data collection.
1. Time related process- Depends on how phenomena evolved over time, for example, wound healing process.
2. Time sequenced phenomena- It is very important to ascertain the sequencing of phenomena. For example, the researcher has to make sure that pregnancy contributes to diabetes mellitus and not diabetes mellitus precedes pregnancy.
3. Comparative purposes- Here, multiple data points are used to compare phenomena over time.
4. Enhancement of research control- The researcher collects data at multiple points to enhance the interpretability of the results. For example, the collection of pre test data with demographic variables allows the researcher to detect for initial group difference. 

Cross sectional design
The researchers collect the data at one point in time and describe the status of phenomena or relationships among phenomena. For example, the researchers assess the level of depression among post partum women and test the relationship between independent (age, socio economic status and level of education) and dependent variables (level of depression). Advantages of cross sectional designs are economical and easy to conduct the studies for the researcher. A major disadvantage of conducting a cross sectional study is that problems arise when society changes rapidly. The researcher can not generalize the cross sectional study findings to a rapid changing society.
Longitudinal design
In this design, the researcher collects the data over an extended time period, identifies changes over time and sequence phenomena like a chain, which is an essential to establish causality.
Longitudinal studies are three types;

1. Trend studies
2. Panel studies
3. Follow up studies

Trend studies
Trend studies help the researcher to collect the data by drawing samples from general population over time with respect to a phenomenon. The researcher draws different samples from same population at repeated intervals, examines the changes at present and makes predictions about future. Foe example,predicting nursing shortage in coming years.
Panel studies
The researcher collects the data from the same participants at different period of time. With the same participants, the researcher examines the changes and identify the reasons fro change. The major problem with the panel studies are subject attrition. Subject attrition is defined as dropping of study participants during the study.
Follow up studies
In follow up studies, the researcher collects the data subsequently, especially after an intervention. In non experimental studies, the sample may be followed periodically to assess the changes in the variables. For example, weight gain among premature infants.

Best Scientific Paper Award 2016

Best Scientific Paper Award

I have presented a scientific paper on "A qualitative study to explore social burden among the wives of alcoholics in tribal area" in a National Conference, conducted by Sree Balaji College of Nursing, Bharath University, Chennai, Tamilnadu, India on 27.06.2016 & 28.06.2016. I have received "Best Paper Award" for my presentation. Thank you very much for your help and support.

Non experimental research

Non experimental research
There are many research problems that can be conducted using experimental or quasi experimental designs. In non experimental research, the independent variable is non-manipulate or unethical to manipulate, however, gives answers for research questions. 

Types of non experimental research
Correlational or Expost facto research has been conducted to study the relationships among variables after variation in the independent variable has occurred. It is designed to assess causal relationship between variables. it is divided into two types; Prospective study and Retrospective study

• Prospective study

In prospective study, the researcher looks forward to assess the effect and study is started with an examination of presumed cause. For example, the researcher selects samples of chronic alcoholics and non alcoholics and continues assessment for a period of time and later assesses the effect.

• Retrospective study

In retrospective study, the researcher observed at the effect of present and looks on to past history of the samples to collect and compare antecedent factors that have caused it.
Another class of non experimental research is descriptive research. The purpose of descriptive research studies is to observe, describe and document aspects of a situation. The researcher uses percentages or proportions to present the results. In descriptive correlational study, the researcher describes relationships among variables, but not causal connection.
Disadvantages of non experimental research
The major disadvantage is its inability to reveal casual relationships conclusively. It may be having false interpretation because the researcher works with known groups.

Quasi experimental research

Quasi experimental research
It involves manipulation of an independent variable, lacks either the randomization or control group features.
Types
There are two types most commonly used in nursing research;
1. Non equivalent control group design
2. Time series design

• Non equivalent control group design

It involves a treatment and two or more groups of subjects observed before and after its implementation. This design is useful, where the researcher is not able to do randomization due to unavoidable ethical reasons. This design is weak without randomization. On the other hand, the design is strong due to collection of pretest data allows us to determine whether the groups had similar morale initially by collecting pretest data, if the two groups are comparable with regard to demographic variables, the researcher could be relatively confident that any post test difference is self reported morale was the result of experimental treatment.

• Time series design

It involves the collection of data over an extended time period and introduction of treatment during that period. The researcher is able to collect measures of outcome variables before the intervention, during the intervention, after a period of intervention and so on.
Advantages of quasi experimental research
It seems to be more practical than the experimental designs. Nursing researchers conduct studies in natural settings, where it is unethical to deliver an innovative treatment to a cluster of people through randomization.
Disadvantages of quasi experimental research
With these types of experiments, there are alternative explanations for observed results. A researcher wants to evaluate the effectiveness of weight gain among infants, if there is no comparison group to observe weight gain, one can easily ask question that: Is the weight gain among infants would have occurred without intervention? These types of questions called as rival hypothesis. If the answer is yes, inferences about treatment effectiveness are weakened.
In research reports, the researcher does not mention regarding random assignment or comparison group, probably it is safe to conclude that the design is quasi experimental and pre experimental.

Experimental Research

Experimental Research
In experimental research, the researcher is an active agent rather than a passive observer. To do an experimental research, the researcher has to follow these principles such as
1. Manipulation
2. Control
3. Randomization

• Manipulation

It means the researcher intervenes something in the group and analyze the outcome. In other words, the researcher varies the independent variable by administering an intervention or treatment and analyzes its effects on the dependent variable. For example, the researcher analyzes the level of knowledge (dependent variable) before and after video teaching programme (intervention).

• Using control group

Control group refers to a group of participants whose performance on dependent variable is used as a basis for evaluating the performance of the experimental group (a group which receives intervention) on the same dependent variable. The researcher keeps the control group (a group with no intervention or administers an alternative or false intervention such as placebo) parallel to experimental group. For example, to evaluate the effectiveness of nutritional supplements, the researcher divides the newborns into two groups with different mode of strategies.
Strategy I: A group of newborns receiving nutritional supplements daily (experimental group) and other group without any nutritional supplements (control group).
Strategy II: A group of newborns receiving nutritional supplements daily (experimental group) and other group with alternatives to nutritional supplements such as glucose powder called a group with placebo intervention (control group).

• Randomization

Randomization means the researcher randomly select the participants in the study without any bias. Through randomization, the every participant gets an equal chance to participate in the study. The methods of randomization may vary from flipping a coin to computerized randomization method.

Experimental designs
Basic Designs
The researcher assigns the subjects randomly into experimental and control group, intervenes the intervention in the experimental group and collects the data after intervention from experimental and control group. This is also called as after only or post test only design. In another way, the researcher collects the information from experimental and control group before and after intervention is called as before-after design or pretest-post test design.
Factorial design

Factorial Design

In factorial design, the researchers manipulate two or more variables together on the study subjects and evaluate the main effects and also interaction effects. Here, the subjects are assigned at random to a combination of treatments. The term cell represents in a schematic diagram as a box. The independent variables are referred to as factors. Each factor is exposing in different levels. For example, a study to assess the effectiveness of supplementary feeding versus normal feeding among toddlers on weight gain. Here, the researcher designs the study on 2×2 table with two factors such as type of feeding A (Supplementary feeding (A1) and normal feeding (A2) and frequency of feeding B one time/day (B1) and two times per day (B2). The researchers assign the toddlers randomly to one of the cells, which consist of combination of treatments. The results will show the best mode of feeding and also the interaction effects resulting from the combination of treatment methods.
Repeated measures design/ Cross over design
The researchers intervene two or more interventions among the same subjects called within subjects design or cross over design. It involves the exposure of the same study participants to more than one treatment. The study subjects are randomly assigned to receive different ordering of treatment. For example, to compare the effects of auditory and tactile stimulation on infants, some subjects would like to randomly assign to receive auditory stimulation first and others would receive tactile stimulation first with the subjects serving as their own control group.  It prevents participant bias by achieving possible equivalence among subjects exposed to different conditions. There is an expected disadvantage of carryover effects of first treatment among study participants. When the researcher introduces two or more treatments and intervenes the treatments in the randomized order is called as experimental repeated measures design.
Clinical trials
The researcher evaluates the effectiveness of new treatment through a randomized clinical trial. It involves the testing of a clinical treatment, random assignment of subjects to experimental and control groups and assesses the effectiveness of treatment. Clinical trials can be done using before – after or after only design.
Advantages of experiments
These are the most powerful designs for testing hypothesis of cause and effect relationships. It also meets the criteria of causality, by Lazarsfield (1955).
1. A cause must precede an effect in time. For example, to test the hypothesis- smoking causes lung cancer; subjects should not have developed cancer before exposure to smoking.
2. There must be an empirical relationship between the presumed cause and effect. For example, in case of smoking and lung cancer, the researcher should collect the evidences regarding higher incidence of lung cancer is directly related to occurrence of smoking.
3. The relationship of presumed cause and effect should not affect by a third variable. For example, a relationship between smoking and lung cancer might reflect an underlying causal relationship between prolonged dust exposure and lung cancer.
Through manipulation, presence of control groups and randomization, an alternative explanation to a causal interpretation can be ruled out.
Disadvantages of experiments
All variables can not be manipulated such as race, ethnicity and history of diseases.
There are many variables, which can not be manipulated ethically. For example, withdrawal of vaccines from control group to test the effectiveness of vaccines in experimental and control group.
Hawthorne effect- Awareness of being in a study may alter the participant’s behavior, which affects the study results.
Double hawthorne effect- Those who administer the treatment know the subjects who are in experimental and control group; alters the study results. In a double blind experiment, neither the subjects nor those administering the treatment know who is in the experimental or control group; are so powerful.