In research methods, validity refers to the ability of findings of a research study to truly represent the phenomenon that a researcher is measuring. Kirk and Miller (2009) define validity as the extent to which a research study measures the subject that it is supposed to measure. Validity is an important element of a research because it helps in ensuring accurate application and interpretation of results of the study.
Research validity can be either internal or external. Internal validity refers to the ability of independent variables in a test to produce accurate results as stated in the hypothesis. Internal validity also determines the relationship between the dependent variable and independent variables. On other hand, external validity refers to extent to which the findings of a research study can be generalized to a larger group or the entire population. Findings of a research study that lacks external cannot be applied to other subjects or variables of a given population, for example, the findings of a research study on specific behaviors of men may not be applied to women. According to Nagel (2010), a research is externally invalid if its findings do not truly represent the general characteristics of the population or phenomenon. This means that the findings of the research cannot be applied to other contexts outside the environment in which the research was carried out. Thus, external validity is achieved when the results of test samples can be applied to the entire population.
Validity can also be categorized into content validity, criterion-related validity and construct validity. Research validity can be affected by various factors such as varied characteristics of the population, improper methods or techniques of collecting data, biasness from the researcher, confounding and negative impacts of the research environment. From my part, validity helps in making the research study more legitimate and convincing.
Reliability refers to the consistency of results of a research study as well as accurate representation of the research population. It is the ability of a research study to reproduce same results when carried out under similar conditions or environments. Bloor and Wood (2011) define reliability is the extent to which a research study would produce similar results when different researchers perform the same experiments under similar conditions. In most cases, researchers do repeat experiments so as to increase the reliability of their findings. In my opinion, the reliability of a research can be tested by replicating the tests or experiments under similar condition with different samples of the population.
The reliability of a test can be improved by deploying appropriate techniques of collecting data. Reliability can also be achieved through performing accurate experiments, using larger samples, random selection of samples, use of appropriate measuring instruments during experiments as well as detailed or critical analysis of statistical data collected during the tests. Nagel (2010) states that the reliability of a research study can be determine by various methods such as test-retest, internal consistency test, alternative form and split-halves. Reliability is important in a research study because it indicates that dependable and can be trusted.
Generalizability is the ability to extend the results of a research to other subjects outside the original experiments or samples. For example, a researcher may conclude that college students are exorbitant spenders after conducting a research study on spending behaviors of 1000 students from different colleges. Generalizability also refers to ability to predict future occurrences of a particular phenomenon based on repeated occurrences in the past. For example, a researcher may generalize his or her findings on a phenomenon to similar circumstances. However, Kirk and Miller (2009) argue that generalization may not be conclusive enough because it is based on probability of occurrences. In my opinion, generalization requires carrying out rigorous research studies for it to be considered exhaustive. Generalizability of research is important because it shows that the samples that were used during the study truly represent the total population.
In my view, validity, reliability and generalization of results of a study are the main elements that determine whether the scientific community will accept the hypothesis of the research or not. Finally, all research studies are bound to produce relevant results by ensuring that the basic elements of the study such as validity, reliability and generalization are not compromised.
How Probability-Sampling Techniques Helps in providing Samples that effectively represent the Target Population
Probability sampling refers to a sampling technique in which subsets of individuals or subjects in a given population are gathered together so that each individual has an equal chance of selection for the samples. Sproull (2007) defines probability sampling as the process by which a researcher gives each member of a population an equal chance for selection to represent other members of the population in the sample. In probability sampling, the researcher ensures that every individual in the entire population has an equal opportunity for selection into a sample representing the population. Probability sampling is done through randomization. Probability sampling also entails selecting subjects from strata samples, thus ensuring that representation of the members of the entire population is highly increased.
On the other hand, simple random sampling refers to a sampling technique in which a researcher gathers all members of the population and then randomly selects the desired number of subjects. For example, the researcher would make a list of all criminals in the United States of America then randomly select the desired number of criminal that he or she would like to use in the study, say one hundred criminals.
In relation to research studies on criminal justice, a researcher would assign a number to each criminal or victim of crime and then use a computer to generate numbers that would identify the criminals or victims to include in the study (Maxfield, 2006). This would help in ensuring that all criminals or victims of crime have equal chances of being selected to research samples.
In my opinion, when probability sampling is carried out effectively, the samples of the study will consist of subjects that possess characteristics similar to other subjects in the entire population. Consequently, results obtained from the study would be similar to those which could be obtained from studying every member of the population. In addition, probability sampling would be preferable in conducting research studies on criminal justice because it would help in prevention or elimination of biasness of the researcher during the research.
Moreover, probability sampling would help in ensuring that research samples effectively represent the target population by providing all members of the population equal chances of selection. According to Stirzaker (2009) and Cochran (2005), the subjects to be used in research samples are systematically collected, thus adequately represent other subjects in the entire population. Finally yet importantly, given that random samples are drawn from other sets of data, such samples do not effectively reflect the reliability of the original findings because the samples do not represent the entire population but rather portions of the population.
Advantages and Disadvantages of a Survey Research
A survey research is a type of study that involves asking the participants questions; for example, a human resource manager may ask the employees how a training and development program would improve their performance or productivity at the workplace. Survey researches usually involve use of interviews and questionnaires to collect data from the participants. According to Sapsford (2010), survey researches are usually used when the researcher targets a large number of respondents. A large number of respondents refers to a group or groups of respondents exceeding one thousand people. In addition, the respondents must be spread over a large geographical area of more than 10 square kilometers such as a district, province or country (Sapsford, 2010). For instance, an educationist may ask students in all public universities in the United States of America to state how higher education have positively transformed their lives. Marsden and Wright (2010) postulate that survey researches are usually used to collect information on how people perceive things as well as their behaviors or actions. For example, a survey research can be effectively used to find out the effect of mass media in buying pattern of consumers, issues like insecurity that voters may perceive to be most important or level of interest of consumers to buy new products of a company.
One of the major advantages of conducting a survey research is the ability to collect large amounts of data. According to Maxfield (2006), survey researches usually allow the researcher to study a large group of people with varied backgrounds, hence providing a large amount of data or information. The researcher may also examine various characteristics of the subjects at the same time, for example, the lifestyles of working class people, attitudes of students towards their lecturers or preferences of consumers in relation to consumption to a particular type of product or service.
Secondly, the cost of conducting a survey research is relatively low as compared to the amount of data collected (Babbie, 2009). This is because data collection techniques used during survey researches such as questionnaires can be administered to all participants at the same time, hence reducing cost of travelling.
Thirdly, survey researches usually involve studying large number of participants, hence larger sample sizes. In my view, use of large sample sizes often increases the validity, reliability and generalizability of the study.
Fourthly, survey researches may be easily applied to a variety of fields such as marketing, international trade, global economics, social development and political sciences that require large numbers of participants.
On the other hand, a major disadvantage of survey researches is the inability to manipulate or control the independent variables. For example, a survey on the effect of mass media on buying patterns of consumers cannot be effectively measured because the effects of independent variables such as inclusion of graphic images in advertisements may not be controlled properly. Similarly, it does not give room for laboratory experiments that are easy to control.
Secondly, respondents in a survey research may give wrong information thus manipulating the effectiveness of the study, for example, a customer aged forty years may claimed that he is thirty years when interviewed through the telephone. Moreover, questionnaires that ask sensitive questions such as the net amount of salary earned by employees in an organization may not be answered in a truthful manner, leading to collection of wrong data.
Thirdly, the information collected through survey researches are based on personal judgment of the respondents. This further lowers the integrity and reliability of the study. Hence, the results of the survey are more prone to human errors.Rea and Parker (2007) also argue that inappropriate use of words during interviews and improper placement of questions in questionnaires during a survey research may lead to biased results.
Last but least, it is difficult to choose the right subjects or individuals to represent the total population in the research samples. For example, a survey on effectiveness of marketing strategies conducted within a given market segment may not truly represent the entire market due to geographical location of the chosen market segment.