Population Sampling: Logic of Sampling

A Basic Understanding of Sampling

A basic understanding of sampling as a technique is necessary in order to understand sampling. These fundamental ideas include, among others, the population and the census sample. First of all, a population is a group of objects or people included in a particular study who must be relevant to the issue being investigated. Second, the term “census” refers to the procedure used to gauge the size of a population of interest. The subset of the population being examined is referred to as a sample. A population of 200 million people, for instance, is nearly hard to quantify, but a group chosen from this population might produce the needed results. Samples are deduced from two aspects; probability sampling where each participant has an equal probability of being included and the nonprobability approach where some units have a higher chance of inclusion than others.

The Logic of Sampling

The logic of sampling is understood from a case where the researcher wants to acquire information about an issue in the population. For instance, an issue touching on the policy preferences in the U.S. will attract interest from voters as the population for the study. Since the population is connected to a common aspect such as voting, a group from this population will be sufficient for the study. The selection would be made using the probability to avoid biases of information. Probability allows for the aspect of representativeness which indicates that the characteristic of the sample is closely related to those of the population.

In some rare cases, the researcher may choose to employ a nonprobability approach in sample collection. Firstly, this is done when the collection of data is a costly difficult procedure. Secondly, there are cases where the population of interest does not offer a random selection of members.

Sampling Designs

The Random simple sample is referred to subset group of the population containing units with an equal probability of inclusion. On the other hand, as outlined in this chapter, multistage random sampling deals with the consideration of random samples in stages. For instance, the consideration of sample from the country’s perspectives, to the counties and lastly to cities which ideally present three stages. Systematic sampling is a probability approach by separating units that have been selected randomly. Lastly, stratified sampling is an approach to selection when dealing with a heterogeneous population.

Chapter Fourteen: Introduction to Probability and Distribution

Probability

Probability shows the chances of occurrences of an event that ranges from Zero to one. Where zero represents the no chance for an event to occur and one represents 100% of the occurring of an event. Another relevant concept concerning probability is the notion of mutually exclusive. This represents a case where there is all the occurrence of all possible outcomes but does not occur at the same time.

Calculation of probability is performed in several ways. Firstly, the probability of an event is deduced by dividing the number of outcomes by the total number of trials made. Where “P” is the probability of x-events. Another concept that can be employed is the idea of independent events which asserts that the outcomes of an event do not influence the other. The independent events can also be deduced by the use of joint probability. This form of probability is the shows two specific outcomes over two independent trials.

Gaussian Distributions

The Gaussian distribution is a statistical approach that is used to generate the probability of outcomes from a known population, and it is commonly used for continuous data. The mean value, therefore, becomes a major consideration. Variation in this set of data results from natural randomness. In this distribution also the area under the curve generated is considered in calculating the probability. The normal in this case can be used to generate exact percentages by employing the concept of standard deviation.

Binomial Distribution

This form of distribution is used in the estimation of the probability of non-continuous data generating mutually exclusive and jointly exhaustive outcomes. The distribution employs Bernoulli process which uses the statistical concept of the combination. A generalized combination is Cnr pr in-1 where: “n” is the number of trials, “r” the number of the desired results, “p” is the probability of the desired result and “q” is 1-p.

Chapter Nineteen: Mixing Research Methods

The underlying logic of mixing different research methods in a single study is identified in cases where the researcher wants to develop approaches without relying on the strength and weaknesses. Ideally, all research methods have their weaknesses and strengths, therefore, combining them can allow a comprehensive scope of the study question. One of the advantages of this method revolves in the idea of triangulation which is the strength of one research method counterbalances the weakness of the other. Secondly, mixed methods allow for corroboration which is ideally the comparison of results of studies developed from different methods tackling the same research question. Lastly, mixed methods allow for the comprehensiveness which relies on the combination of all the results from various methods to outline the possible answers to the study question.

Designing Mixed Methods Research

The designing of mixed research methods is done to offset or mitigate the challenge encountered in the management of various streams. A system known as the notational scheme has been established to define the relationship between this streams and make the management easier. A stream, in this case, is a component of the mixed methods that includes a lone strategy and occurs in two forms; quantitative (quant) and qualitative (qual).

Identification of the type of stream leads to the development of a strategy on the sequences to be followed in applying these streams. Majorly this is done through the concurrent and sequential designs. A concurrent design is an approach that allows the research streams to happen simultaneously without reference to each other. On the other hand sequential design, the first chosen stream feeds information into the consecutive one.

References

Eller, W., Gerber, B., & Robinson, S. (2013). Public Administration Research Methods: Tools for Evaluation and Evidence-Based Practice. Routledge.