RM - Lecture Eleven of transportation.pptx

Lecture

1. Research Methodology Dr. Nermin M. Gohar

2. Scientific Research Instruments: Samples

3. Basic concepts

4. Why use samples? Homogeneity of the original research community. High cost, time, and effort: if the study population is large and geographically far apart. Poor control and supervision: when the study population is large. It is not possible to enumerate the entire individuals of the original research population. It is not possible to conduct the study on the entire individuals of the original research community

5. Phases of sample selection Identify the original population of the study: The researcher must determine from the outset the objective of the study, its type, and the individuals included and not included in the study. Prepare a list of members of the original population of the research. Selection of a representative sample Selection of a sufficient number of individuals in the sample.

6. Samples Types

7. Probability (random) samples

8. The simple random sampling Used when two conditions are met: All members of the research community should be known . To have homogeneity between these individuals . The simple random sample is selected according to the following methods: Draw style Table of random numbers Random choice of retrospection can be from a limited population, and without retrospection. However, this method of retrospection is impractical and rarely used in social research.

9. Systematic Random Sampling A regular sample is usually selected by listing the individuals of the original study population and then each individual is given a sequential number. Then the number of individuals of the research population is divided by the size of the required sample and produces the number that will be separated

10. Stratified Random sampling This type of sampling is used in heterogeneous populations whose vocabulary varies according to certain characteristics, such as educational level. Selection steps: The division of society into homogeneous categories or groups according to a certain characteristic. Determine the number of individuals of the total sample. Determine the ratio of each level in the selected sample to the total size of the original population. Determine the number of individuals per level in the selected sample

11. If we assume that there is a population consisting of three classes, the upper class of 1,000, the middle class of 4,000, and the lower class of 5,000, it is required to select a random stratified sample of 100 people using the proportional distribution method. The answer is in the following table: Stratified Random sampling, example: Sample Percentage No. Classes 10 10% 1000 Upper 40 40% 4000 Middle 50 50% 5000 Lower 100 100% 10000 Total

12. Cluster sampling Example: If we want to study the annual household income in Gaza City, we may choose a cluster sample in two stages as follows: In the first stage, we consider clusters as city neighborhoods, and we may divide the city into neighborhoods and take a sample of them in an appropriate size with the size of the neighborhood. We divide each of the selected neighborhoods into condominiums and choose from each of them an appropriate number of apartments and then choose the income of the families who live in these selected apartments. This gives us a two-stage cluster sample.

13. Non-Probability Samples

14. Purposive Sample It is known as a purposeful, intentional, or judgmental sample. Individuals with particular characteristics are selected. For example, if a researcher wanted to study consumer reviews about a type of instant coffee (Nescafe).

15. Quota Sampling The difference between quota sampling and the stratified sample is the method of selecting the members of each class, as the random method is not used in the selection in the quota sample, but rather the method of chance and intention is used. This type of sample is used in the study of public opinion and in educational and social studies.

16. Accidental Sample The sample consists of individuals whom the researcher meets by chance. If the researcher wants to measure the public opinion about an issue, he chooses a number of people he meets by chance, whether on the street or on the bus. This sample is only self-represented, but it is easy to use and gives an idea of what individuals think about the issue in question quickly. The larger the sample size, the more accurate the results.

17. Sample size and representativeness of the study population There is no given percentage of the size of the study population that can be applied to all cases. There are a group of factors that affect the size of the study sample, which are the following: 1. The degree of accuracy and confidence to be achieved: The degree of accuracy is the proximity of the sample results to the actual reality. Confidence degree is the extent to which the results of the study are likely to not match the actual results. 2. The homogeneity of the study population 3. The size of the study population: There is a direct relationship between the sample size and the size of the study population. 4- The degree of generalization sought by the researcher. 5. The used research method: Does the researcher want to use the survey or experimental method? And what kind of experimental method will he use? Surveys require a representative and sufficient sample, and some experimental designs require multiple experimental and control groups, which means the need to choose a large sample size ( Obeidat , Adas and Abdelhak , 1998).

18. Sample size and representativeness of the study population The following table shows the appropriate sample size at different levels of the original study population : Appropriate sample size Indigenous community size Appropriate sample size Indigenous community size 226 550 10 10 242 650 28 30 269 900 59 70 322 2000 118 170

19. Sources of errors in samples

20. Sampling error, coincidence error, or random error This error is due to the nature of the random selection of the sample members, so we find that the results of the sample differ from the results of the original population 0 Selecting samples with the best sampling methods do not guarantee that the selected sample is representative of the population 0 It is not possible to obtain a sample whose composition matches the composition of the community completely, so this type of error occurs when the values of the real community features diverge as a result of random sampling from the values we obtained from the sample 0

21. Bias error It is due to the researcher, in which there is a tendency to prefer units with certain characteristics over others to join the sample, and this causes the basic characteristics of the original community to not be represented 0

22. If the population of secondary school students consists of (50% males and 50% females) and the researcher obtained a sample of males (20% and females 80%), this may result in biased results. Then, the sample will not be accurately represented, and thus the type of individual in the sample becomes an influential variable. The bias error often occurs as a result of poor planning when selecting sample 0 this error is due to the following reasons: The inefficiency of researchers in calculating estimates, the ambiguity of questions, inaccurate responses of examiners, failure to collect data from some individuals or collect data more than once from the same individuals, and lack of a sound framework when selecting sample 0. Example

23. How to minimize selection bias errors Randomly select all sample units using one of the random selection methods Do not replace any selected unit with another Complete answers to all questions Conduct empirical research (survey sample) to detect intentional and unintentional bias Train researchers well to collect data and adhere to instructions

24. Bias Estimation Error It is the mistake we make which is related to the method of estimation or appropriate methods of analysis Bias error caused by misidentification of the sampling module When we define the sampling unit, it must be clearly defined in a way that minimizes bias errors that result if the unit is not clearly defined Other common errors in samples Unresponsive errors (attributions to not updating the frame). Categorization and data processing errors. Printing errors. Errors in interpreting results on the number from the correctness of estimation methods and analysis methods

25. In the light of what I studied Through groups select the type and sample size in your research with the logical justification for it

26. Thank You