Research validity refers to the ability of a given research study to measure exactly what it is intended to measure. For example, if a correctional officer or police chiefs carry out a research study claiming to measure how capital punishment has helped in reducing criminal activities, the correctional officer or police chief must understand how data should be gathered to help in establishing whether the research study really captures all the relevant information relating to the number of capital punishments imposed and the rate of decrease in the number of crimes performed during a particular period. The validity of a particular research is determined by its ability to measure what it is supposed to measure. Kirk and Miller (2008) define validity as the ability of a research study to establish that the results obtained during the actual study meet all the requirements of the research method.
According to Brinberg and Kidder (2010), and Bailin and Grafstein (2010), the impacts of the results of a research study are dependent on two major factors, namely, the validity and reliability of the statistical data collected and the assessment of variables of the research. Basically, the validity of research data determines whether the actual measurement or assessment of variables of the research is done according to the purpose of the research study. Kirk and Miller (2008) assert that the validity of a research study encompasses two essential parts, namely internal validity and external validity. Internal validity determines whether the results of the study are genuine and justifiable in relation to selection of sample populations, collection and recording of data and statistical analysis of the collected data. For instance, a research study may lack strong internal validity if testing of the research variables was not conducted in a similar manner to the way control groups were treated. Internal validity of a research may also be affected if confounding variables are not accounted for or accommodated in the study during the research design. On the other hand, external validity deals with the ability of the research study to provide reasonable results that can be easily transferred to other groups within the entire population represented by the sample groups. For example, in the research proposal to determine whether capital punishment deters crimes in the United States, results of a research study conducted in New York should be transferable to other states such as California, Illinois and Minnesota.
In my view, internal validity often dictates the structure of the research study and steps to be followed during the research study whereas external validity concerns examination of the final results and any causal relationships between the dependent and independent variables. According to Suinn and Oskamp (2009), external validity also deals with generalization of the research findings to populations and other measurement variables while internal validity deals with design of the research study in relation to the cause-and-effect principle. Moreover, validity can be further categorized into content validity, criterion-related validity and construct validity. Jupp (2011) also explained that validity of a research study determined whether it truly measured that which it has been intended to measure in addition to how truthful the results were presented.
From my part, high levels of internal and external validity can only be achieved through designing a proper research study. Moreover, a strict protocol should also be followed to ensure that there is minimal deviation from the objects of the research study. It is also important to note that external validity cannot be achieved before achieving internal validity.
Threats to Research Validity. The validity of a research study may be affected by various factors such as reliability, randomization, varying sizes of the samples and biasness of the researchers or participants amongst others.
Reliability. One of the most common threats to the research validity is reliability. Reliability refers to the ability of a research study to produce similar results or outcomes after repeated tests. Given (2008) proposes that the reliability of a research study is determined or achieved if different researchers can perform exactly the same experiments under similar conditions and be able to generate the same results. Reliability usually depends on the consistency of the results of the research study. For this reason, a retest or post test of a research study should provide the same results subject to the absence of changes in the initial conditions for the research sturdy.
According to Ellis, Hartley and Walsh (2010), the reliability of a research study may be affected if the assessments are taken over time or after changes in the initial conditions, or when the assessment is undertaken by different persons. For example, if a correctional officer determined that the use of capital punishment in the judicial system has helped in deterring crime in New York, the research study may produce different results or findings when carried out in a different state, say California, due to the factors such as geographical settings, socio-economic differences and religious or cultural believes of the communities. Moreover, as compared to the research carried out in 1940, a different research carried out in 2012 would give different results due to the time lag and changes in various research components such as behaviors of the new generations.
The reliability of a research study may also be affected if the assessment techniques used are highly subjective. For example, behavioral conduct of an individual may not be used because different researchers would have different interpretations of the same conduct. In my opinion, a researcher must ensure that reliability errors are highly minimized so that any differences that may be seen in the final research data can only be attributed to interventions and not sloppy or shoddy measurements. For example, when carrying out a research study to determine whether capital punishment deters crimes in the United States, collection of data through oral interviews may compromise the reliability of the study. This is because human judgment over certain issues often changes depending on several factors such as time, mood and attitude or place of the interview. Moreover, the participants might also not be willing to give reliable information or the researcher may distort the information collected in order to favor his opinion. Hence, adequate statistical data should be gathered to support such a hypothesis. Research studies with high levels of reliability and validity often receive quick acceptance by the wider scientific community.
Poor Randomization. Randomization refers to the process of carrying out experimental trials at specific intervals. Randomization involves dividing the dependent variables of the research study into specific groups that are tested randomly. Through randomization, the researcher is able to create a group of dependable variables that have desirable characteristics. For example, in a research study to determine whether the capital punishment deters crimes, the participants should be potential criminals who have similar characteristics such as economic or financial stability, comparable behavioral characters or conducts and age must be from the same social class and nationality. Additionally, all participants must be selected randomly and be exposed to similar environments during experiments or tests.
According to Given (2008) and Ellis, Hartley and Walsh (2010), an ideal research study should cover the entire population. However, due to the high costs and near impossibility of carrying research using the entire population, sample populations are often used. A sample population is a segment of the entire population used to represent it. During a research study the sample populations are studied, adequate data is collected about the samples and conclusions are drawn based on the observed characteristics of the sample. These conclusions are then extrapolated to the entire population.
In most research studies the size of the sample population affects the validity of the research. For example, a sample population with a very small size often gives inadequate information that is inconclusive as well as may lack statistical significance. On the other hand, a very large sample population may also prove difficult to manage, hence leads to the wastage of resources such as time and money. Although large samples may be preferred due to precision and high statistical power, it is advisable to select a manageable size of a sample to ensure high validity of the research study. For example, in the proposed research study to determine whether death penalty may be used as a deterrence of crimes in the U.S., the appropriate samples of population would include the manageable number of criminals in prisons or randomly selected religious and political leaders.
In my view, the selected criminals, religious or political leaders would give their views and opinions on the research topic. Their opinions would then be used to represent opinions of the various populations they represent.
Bias during Research Studies. Biasness refers to the intention of a researcher to produce research findings or present facts that are untrue and would not be otherwise produced by other independent research studies.XXX defines biasness as a systematic error caused by improper design of a research study, improper collection of data, poor analysis of data as well as presentation of results that are not factual. During a research study, biasness may arise from poor identification of the dependent and independent variables, wrong measurements, incorrect estimates, wrong observations of variables and poor design or manipulation of the research structure.In addition, biasness in a research may also result from poor personality of the researcher, for example, due to his or her dishonesty or prejudice.Biasness often has stern negative effects of the validity of a research study. For example, wrong measurements often produce unreliable results.
In the research proposal to determine whether capital punishment deters crimes, strict adherence to the research protocols or procedures must be observed. Moreover, the researcher should not abuse his or her power in data collection as well as during interpretation of the results. For instance, if during the research study it is found that most people do not believe that capital punishment has significant effect in discouraging criminal activities, the researcher should not alter the results of the study so as to comply with his or her opinion and views.
From my part, biasness during the research study would be reduced by ensuring that there is no conflict of interest between a researcher and the topic of study. For example, a victim of a crime may not be selected as the chief researcher due to the high possibility of conflict of interest.
Other Threats to Validity of Research Studies. In addition to the above discussed threats, other factors that may affect the validity of a research study include the characteristics of populations, changes in the research environment, methods of collecting data and effect of time amongst others (Kraska & Neuman, 2011). Validity of a research may also be affected by social threats such as demoralization of the participants and uninformed consent, which can make the participants lose interest, confidence or trust in the study. Such social threats can be eliminated through helping the participants in understanding the purpose of the study and the various reasons for their engagement.
The above threats to the validity of a research study often exist almost throughout the research study; it is therefore important for the researcher to develop or construct an appropriate study design as well as formulate and implement various study protocols or procedures that would help in minimizing these threats to validity of the research. For example, the researcher should strengthen the research validity by using well-validated measurements, selecting manageable sizes of samples of equally approximation and adding adequate control groups to reduce or eliminate confounding of samples. Similarly, the researcher should perform re-tests and post-tests to verify the results of previous experiments, hence improving reliability of the findings. The measurement methods and procedures used must also be equivalent during the entire research.
In addition, the researcher must not be biased or have pre-informed opinion about the situation before conducting a research. He or she should also maintain appropriate balance between the various risks that are likely to be encountered during their study and their possible implications of the research results. The researcher should ensure an appropriate use of statistical techniques that have sufficient power, such as chi-square test, calculation of statistical inference either mean or median.
Adequate validity of the research study can also be achieved through random selection of the sample groups. Moreover, great care should be taken when allocating controls of the study. The control groups are usually used in minimizing problems that may arise from external validity.
Although some researchers may suggest that validity does not apply to their research studies, disregarding validity often leads to questionability of the trustworthiness of the research study as well as the truthfulness of its results. Lack of trust and truth in a research study often lowers the confidence of future users of such research findings.