A Questionnaire contains information is elicited using a series of questions. All the participants in the research or the sample in a population then fills these issues. The questions can either be open-ended, in which the respondent fills using his or her wording, or closed in which the respondent chooses from a set of responses provided for by the researcher. The closed-ended questionnaires can further be sub-classified into dichotomous questions of yes or no answers, multiple questions, cafeteria questions, rank-order questions or rating questions.
Questionnaires are convenient, as they require less time and skills to administer. They are especially preferred where the respondent includes a large elite group that can be able to read, comprehend and respond to the questions appropriately. Less time and energy is used to cover a large population or sample with least energy. The respondent is also able to think before supplying the response, as there is no pressure for the response as for the case of the interviews.
Delegate your assignment to our experts and they will do the rest.
Conversely, they cannot use questionnaires on illiterate people who cannot read; else a guide to reading and interpret the questions is required. Besides, the response leaves out some of the issues they choose thus failure to gather information. On the case of closed questionnaires, the responses from the population are usually limited to the researchers’ options of answers provided.
Interviews are another method of data collection that contain either structured or unstructured verbal questions to gather information used in a research study. Interviews are very useful in cases where the respondents cannot read for themselves or the elderly who may need some guidance. Besides, they are useful where respondents’ real response is required more so in the qualitative study. Interviews are however very cumbersome and conducting the in a large population of respondents may prove very hectic or impossible.
The third data collection method is the use of Records. When a researcher relies on the previously documented report or data to derive information required in research or study. Such documented information includes school performance and medical records. Record are useful in retrospective cohort studies or where primary data in not readily available. However, a mistake in the initial recording of the data may be replicated in the future subsequent studies. Observation can also be used.
Sampling methods
Sampling methods are classified broadly as either probabilistic or non-probabilistic. In probabilistic, every element in the population has a possibility of greater than zero of being picked in the sample. In non-probabilistic, some elements in the study do not have a chance of being picked in the sample.
Simple Random Sampling is probabilistic in which, all the elements are given an equal chance of selection into the sample by assigning a number to each element then using a lottery system to select. The advantages of simple random sampling are that the estimates are easy to calculate and easily inferred. Besides random simple sampling is always an EPS design. Not all EPS are however designed for random simple sampling. On the other hand, if sampling frame large, this method impracticable since the coding tends to be very cumbersome. Besides, minority subgroups of interest in population may not be present in sample in sufficient numbers for study
Systematic Sampling Involves arrangement of the elements into orderly schemes the choosing from the systems in regular order. Arrangement into the systems is normally random, but a selection of the projects is commonly systemic. The advantages of this method include but not limited to great ease in selecting the sample, ease of identifying the sampling frame and the even distribution of the sample in the entire population of reference. It is, however, tough from one survey to assess the precision of the estimate. Besides, if there is a hidden periodicity in the population which coincides with the interval of selection, bias may be introduced in the sample, which may prove very hard to detect.
The last probabilistic method is Stratified Random Sampling. The population is usually organized into some sub-populations, strata, from which individual elements are selected. The merit of this approach is that from the stratum, every element has a chance of selection. Use of the levels in sampling selection ensures proportionate selection and even representation of minority strata. Besides, since every layer is treated independently, different sampling techniques can be used in various strata. The greatest drawback of this method is that each stratum is separately prepared.
An example of Non-probabilistic methods is Convenience sample also known as accidental sampling, haphazard sampling or opportunity sampling. Selection is done based on availability and convenience of the element to the researcher or even the respondent
Inferential statistics
The inferential statistic is composed of procedures that are used to makes up conclusions, prediction, and decisions about the characteristics of a population. These features are obtained from a sample of the population. To achieve this, a series of the statistical test have to be done. Various assumptions are made to make a valid probability decisions using inferential test. To use inferential statistic the researcher makes the assumptions that the sample collected independently and randomly.
There are several steps in making an inferential conclusion. To begin with, a choice of hypothesis to the question is selected. Second, a researcher chooses alternative hypothesis which is to be accepted when the original hypothesis is rejected. A certain interval together with alpha values is set as rules to either reject or accept the null hypothesis. Finally, a decision is made after computation of appropriate random sample from a population.
The main inferential statistic can be used to either test group difference or test relationships. Statistical test for group difference includes the student T test and the Analysis of Variance (ANOVA). Relationships among characteristic of the population obtained from random sampling can be done using Chi Square, Correlation, and Regression model among others. The most widely used statistical test in solving business problems includes the Chi-square, ANOVA, and Linear regression.
Chi-square is used to test whether observed frequency fit the researchers experience. In One –way classification Chi-square, the researcher question whether the observed frequency match expected rate. There expected frequency may be known or computed. Two Way classification Chi-square tests whether there is an association of rates between two categories.
Analysis of Variance or ANOVA is used to determine the difference in means across two independent groups. For one to use ANOVA test, assumptions of the parametric test have to be met. They include homogeneity of variance, independence of observations, and equality of variance and lack of significant outliers.
Inferential statistics have largely been used in market research. In a particular research, a researcher was testing whether there was a difference in an average number of customers buying Mercedes model 220 from two different Chicago and Washington Dc cities. The researcher used Independent t-test to test the null hypothesis. Since the p-value was less than 0.005, there was no difference in buying between the two cities.
Descriptive statistic
The descriptive statistic is a single value that describes a sample or a population. The two category of descriptive statistic is the measure of central tendency and a measure of dispersion. A business problem may be described using mean, mode, median variance or standard deviation. Descriptive statistics describe certain phenomena using a sample of a population without drawing any conclusions, predictions or describing cause and effect. The descriptive statistic can then be used to show graphical representations of the characteristic of a sample using charts and graphs. The definitive test was employed in this study to compute the means that of customers buying Mercedes Benzes between the two cities.
In conclusion, the type and the choice of sampling techniques together with the statistic are most crucial in any business research. For any significant results, the researcher must adhere strictly to the study requirements, which will affect the accuracy the decision and conclusion made.
Reference
Films for the Humanities & Sciences (Firm), Films Media Group.,&Uniview Worldwide. (2006). Inferential statistics . New York, N.Y: Films Media Group.
Hayward, K., & Hobbs, D. (September 01, 2007). Beyond the binge in ‘booze Britain’: market- led liminalization and the spectacle of binge drinking. The British Journal of Sociology, 58, 3, 437-456.
Richard K. Miller & Associates. (2006). The ... retail business market research handbook . Loganville, GA: Richard K. Miller & Assoc.
