Types of samples and the factors/methods to determine its size

Population

How is a population delimited?

Populations should be clearly placed around their
characteristics of content, place and time.

A population is the set of all the cases that agree with a series of specifications (Selltiz et al., 1980)

The delimitation of the characteristics of the
population depends not only on the objectives of the research, but also on other practical reasons

Sample

Errors that should be avoided when selecting the sample

1) dismiss or not choose cases that should be part of the sample (participants that should be and were not selected).
2) include cases that should not be because they are not part of the population.
3) select cases that are truly ineligible
(Mertens, 2005 )

The sample is, in essence, a subgroup of the population. It is a subset of elements that belong to that set defined in its characteristics that we call population.

Types of samples

Probabilistic samples

Subgroup of the population in which all elements of it have the same possibility of being elected.

Non-probabilistic samples

How are they obtained?

you getdefining the characteristics of the population and the size of the sample, and by means of a random or mechanical selection of the units of analysis

Non-probabilistic or directed sample Subgroup of the population in which the choice of the elements does not depend on the probability but on the characteristics of the investigation.

How are they obtained?

the procedure is not mechanical or based on probability formulas, but depends on the decision-making process of a researcher or a group of researchers and, of course, the selected samples obey other research criteria

Procedures for making a probabilistic sample

  1. calculate a sample size that is representative of the population
  2. select the sample elements (cases) so that at the beginning everyone has the same possibility of being chosen.

Calculation of the sample size

ask: since a population is N size, 3
What is the smallest number of sample units (people, organizations, telenovelas chapters, etc.) that I need to make a sample (n) that assures me a certain level of standard error, say less than 0.01?

Questions

Size of the universe?
Maximum acceptable error?
Estimated percentage of the sample?
Desired level of confidence?

Define the maximum acceptable standard error (probability), the estimated percentage of the sample and the level of confidence

Calculate a sample size that is representative of the population

Select the sample elements so that at the beginning everyone has the same possibility of being chosen.

Choosing the right size for a sample

Variables

  1. The margin of error that the researcher will tolerate.
  1. The level of confidence desired.
  1. The variability or dispersion of the population studied.

The maximum admissible error, designated E, is the magnitude that is added to and subtracted from the sample mean (or sample proportion) to determine the extreme points of the confidence interval.

When working with a confidence interval, logically, relatively high confidence levels like 95 and 99%, which are the most common, will be chosen. To calculate the size of the sample, a z statistic will be needed that corresponds to the chosen level of confidence

is the standard deviation of the population. If the population is very dispersed, a large sample is required.
On the contrary, if it is concentrated (homogeneous), the sample size that will be required will be smaller.

suggestions for determining an estimator of the standard deviation

Conduct a pilot study
Conduct a comparative study
Use an interval-based approach

Sample size to calculate a population mean

Formula

image

image

n: it is the size of the sample.
z: is the standard normal value corresponding to the desired confidence level.
the standard deviation of the population.
E: is the maximum permissible error.