Introduction
Reliability and Validity
Every research carried out should be tested for reliability and validity. Reliability is defined as the extent to which a given or taken measure will give reliable results. Three different reliability tests can be taken on research. They include; test-retest reliability, alternative forms reliability, and internal consistency reliability. The test-retest reliability is used to test whether the research results are similar if the same research is carried out under similar conditions.
Pearson coefficient is used to assess the stability of the repeated measures. In alternative form reliability, repeated results from research carried out using different forms are tested to see the similarity of the results. Finally, the internal consistency reliability is used to test how an individual measure is converted into a composite measure. This type of reliability test can be assessed by applying correlation performance on the two halves of a test. When carrying out research, one can increase reliability by increasing the size of the sample. An example of reliability is seen when measuring the weight of an individual where one believes that the weighing machine gives the same value each time the person steps on it.
Validity in research is used to test whether the research carried out has measured what it was intended to. There are five types of validation tests, which can be done on research. They include content validation, criterion validation, construct validation, internal validation, and external validation. Content validation is used to check how the research contents relate to the variables that are going to be studied. It seeks to answer questions on how well the research questions are representative of the research variables.
Criterion validation is used to test if the research criterion being used is relative to other criteria. Construct validation is used to test for the underlying constructs that are being measured. It tests for convergent validity, discriminant validity, and nomological validity. Internal validation is used mostly in primary research designs and is used to check for the relationship between the dependent and independent variables.
External validation is used to check if the results obtained from experimental research can be generalized. An example of validity is seen in a study where a researcher measuring whether leisure activities have an adverse effect on the performance of secondary school students conducts a survey to determine how many undergraduates drive on their way to school and then comes up with a correlation between the two. Validity is used to imply reliability. This is because a valid measure must always be reliable. On the other hand, reliability does not always mean the same as validity because a reliable measure does not necessarily imply that it is valid.
Discussion
Know what makes an effective questionnaire. If I were to give you a questionnaire that needed work, could you revise it?
A good questionnaire should consist of several things if it going to be effective when carrying out research. The first thing a questionnaire should contain is a good introduction. When you are beginning a survey, you should have an introduction that details your objective in such a way it grasps the attention of the potential respondents. If you are doing an online survey, then to avoid respondents failing to complete the questionnaire you should have instructions on how to complete the questionnaire and an estimate of how much time it will require to complete the questionnaire.
The questionnaire should ask questions that are relevant to the survey and should always be objective. If the questionnaire asks too many personal questions, then the respondent will most probably decline to answer the questionnaire or give false information. It is best if the questionnaire asks the important questions first followed by the demographic questions. This is important especially if you are carrying out an online survey because the respondent may abandon answering the questionnaire.
The questions in the questionnaire should be organized into logical groups. This makes it easier for respondents to answer the questions and saves a lot of time. The language used in the questionnaire should be easy to understand. If the questionnaire uses unclear or ambiguous language then the respondent may give misleading results. The questionnaire should always try to avoid the usage of technical terms and jargon since this might frustrate the respondent and make them abandon the survey.
The responses in the survey should be randomized to avoid the survey running the risk of order bias in responses. As a rule, the questions in the questionnaire should be short, simple, and to the point. After the survey, it is always important to thank your respondent for taking part in your survey. This helps foster a good relationship between you and the respondent, as you may need them to complete other surveys in the future.
Explain the difference in Correlation and Causation
There is a lot of confusion between the usage of the words correlation and causation. In theory, the two can be easily distinguished. In causation, one incidence leads to another while in correlation; one incidence connects or links with another. An example of causation is whereby cigarette smoking leads to lung infection whereas an example of correlation is whereby smoking is connected with alcoholism. If one occurrence leads to another, then they are highly correlated but any two occurrences happening together may not be because of causation. Correlated incidents may be due to an ordinary cause.
For instance, the idea that red hair is linked with blue eyes originates from a common hereditary specification. A correlation may also occur due to the presence of causality. This is seen where smoking is not only linked with lung cancer but in reality, causes it. Causality is established through a controlled study where two sets of persons comparable in nearly every aspect are given two distinct sets of incidents and the result is compared.
If the two sets of persons have considerably distinct outcomes, then the distinct outcomes may have led to distinct results. One can typically create correlation unless the results are tremendously remarkable. In general, people should be cautious of their own prejudice. The media winds up a causal association among correlated occurrences when causality is not even reflected by the study. It is therefore clear that when there are no apparent reasons for agreeing to causality.
People should only recognize the correlation. Two occurrences happening in close immediacy do not mean that one led to the other even if this appears to make ideal sense. Generally, it is very difficult to determine causality between two associated events. In disparity, there are several statistical tools to determine a correlation that is significant statistically.
If I give you a company and its traits (product, audience/customers, etc), you should be able to tell me how to collect data from their audience/customers (i.e. through internet survey, through phone calls, etc.)
Collecting data from a company’s audience and customers can be done via internet surveys, individual phone calls, and conferences. The technology for the internet survey is young and developing. Conducting an internet survey was in the past perceived as a long process that needed familiarity with internet programs. Today, software packages and internet survey services have made internet survey research much simpler and quicker.
The advantages of internet surveys include the ability to contact individuals in faraway locations, the capacity to reach partakers who are difficult to access, and the expediency of having automatic data collection which decreases the researchers’ time. The limitations of the internet survey include doubts over the strength of the data collection and sampling concerns, issues surrounding the research design, execution, and assessment of the internet survey (Wiid & Diggines, 2009).
An individual phone call is another method of collecting data from a company’s audience and customers. The utmost advantage of this method is its speed and minimal costs. In constructive circumstances, about five personal phone calls in a company can be done in a day via the telephone. However, there are certain circumstances where individual phone calls may not be appropriate methods of data collection.
For instance, if the respondents require several factors to test their opinions, it is difficult for them to keep in mind five or more issues. In addition, the absence of personal contact hinders the company from evaluating the respondents and it is thus not able to achieve an additional feel behind the respondent’s reply. Despite these disadvantages, the pros of individual phone calls in data collection are substantial and the technique is likely to persist in making crossroads against other methods (Aaker, 2010).
Conferences as a method of data collection involve inviting the company customers for a talk and compensating them for their time. Improvements in information technology and methods of data collection in a company have resulted in the accessibility of big data sets in companies. Through conferences, the company has an exceptional opportunity to examine the data collected and obtain intelligent and helpful information. They provide an opportunity for the presentation of the current company proceeds including software and other systems. However, this method is time-consuming and may lead to bias.
In the research, what do measurement scales and attitude measurement mean? Support your thought and give examples
Measurement scales refer to the nature of the postulations made concerning a particular variable and its properties. The measurement scales mostly used in marketing studies can be divided into comparative and non-comparative scales. In Comparative scales, the respondent indicates differences between two or more manufacturers, services, trade names, and other stimuli. Examples of comparative scales include paired assessment, dollar metric, and scales used in line marking.
Non-comparative scaling involves the valuation of a single product. This valuation is free of the other variables that the researcher is interested in. This scaling may also be described as monadic since it is more extensively used in business marketing research. Non-comparative scales include constant rating scales and Likert scales (Wiid & Diggines, 2009).
Attitude measurement refers to the measurement of a consumer’s predilection concerning a certain good or service. If it is positive, then the customer is likely to buy the good or service. Attitude measurement is composed of elements such as beliefs, expressive feelings, and customers’ willingness to react. Beliefs involve measurement of the products potency and economy. Expressive feelings involve the likes and dislikes concerning a product. These three factors when merged together result to an image of the good or service in the consumers’ brain. Several techniques such as Disguised, Non-disguised, Structured, and Non-structured are employed in attitude measurement.
References
Aaker, D. (2010). Marketing research. Chichester: John Wiley.
Wiid, J. & Diggines, W. (2009). Marketing Research. Cape Town: Juta.