Research Design and Needs of research design Ch-4.pptx

Research Design

1. Chapter 4 Research Design

2. Points of Focus Introduction Needs of research design Selection a study design Sampling Design; Sample Size Sample Technique Probability Sampling Design Non Probability Sampling Design Summary

3. 4.1 Introduction Research design is the conceptual structure within which research is conducted; it constitutes the blueprint for collection, measurement and analysis of data. A   research design   is a strategy for answering your   research question   using empirical data. it is a master plan specifying the methods and procedures for collecting and analyzing the needed information.

4. Introduction….Cont’d The design decisions happen to be in respect of: What is the study about? Why is the study being made? Where will the study be carried out? What type of data is required? Where can the required data be found? What periods of time will the study include? What techniques of data collection will be used? How will the data be analyzed? In what style will the report be prepared?

5. Functions of a research design Identify or develop procdurse or logistical arrangments require to undertake study. Improving the quality of procedurse to ensure their validity, objectivity, and accuracy . To provide for the collection of relevant evidenc e with minimal expenditure of effort , time and money . Generally, the design which minimizes bias and maximizes the reliability of the data collected and analyzed is considered a good design

6. Cont’d A research design appropriate for a particular research problem, usually involves the consideration of the following factors: the means of obtaining information; the availability and skills of the researcher the objective of the problem to be studied; the nature of the problem to be studied; and the availability of time and money for the research work.

7. Furthermore, research design explains how the researcher wil l find answers to the research questions. It sets out the logic of the inquiry. But how all these can be achieved depends mainly on the research purpose. Research purposes may be grouped into four categories, viz., Exploration Description Explanation coorrelation

8. Needs of research design Reduce Mistake s: reduces errors and inconsistency Improves r eliability and efficiency : Eliminates bias and errors Reduces time waste: This helps test the hypothesis: Just as for better, e conomical and attractive construction advance planning of the methods to be adopted for collecting the relevant data and the techniques to be used in their analysis

9. Cont’d.... Clarity of research objectives: Improved data collection: Better data analysis Efficient use of resources:

10. 4.2 Selecting a study design there are difference types of study designs in case of q uantitative studies in case of qulalitative studies in case of mixed research designs. The various designs in quantitative studies have been classified by examining them from three different perspectives : The number of contacts with the study population The reference period of the study The nature of the investigation.

11. types of Study design types of design

12. 4.2.1. The number of contact Based on the number of contacts with the study population, designs can be classified into three groups:

13. A. Cross-sectional studies also known as one-shot or status studies, is the most commonly used design in the social sciences. Cross-sectional surveys collect data from a sample of participants at a single point in time. is best suited to studies aimed at finding out the prevalence of a phenomenon , situation , problem , attitude or issue , by taking a only one time /cross-section of the population. Such studies are cross-sectional with regard to both the study population and the time of investigation.

14. you collect data from a population at a specific point in time As these studies involve only one contact with the study population, they are comparatively cheap to undertake and easy to analyze. However, their biggest disadvantage is that they cannot measure change. b/c it is only one contact. Cross-sectional studies …

15. B. The before-and-after study design The main advantage of the before-and-after design (also known as the pretest/post-test design) it can measure change in a situation, phenomenon, issue, problem or attitude. It is the most appropriate design for measuring the impact or effectiveness of a program. can be described as two sets of cross -sectional observations on the same population to find out the change in the phenomenon or variable(s) between two points in time.

16. A before-and-after study is carried out by adopting the same process as a cross-sectional study except that it is comprised of two cross-sectio nal observations, the second being undertaken after a certain period. Depending upon how it is set up, a before-and-after study may be either an experiment or a non-experiment . It is one of the most commonly used designs is evaluation studies . The difference between the two sets of observations with respect to the dependent variable is considered to be the impact of the program. The before-and-after study design …

17. More expensive and more difficult to implement. It requires a longer time to complete, particularly if you are using an experimental design, and hence, need to wait until your intervention is completed. In some cases, the time lapse between the two contacts may result in attrition in the study population. It is possible that of those who participated in the pre-test , some may move out of the area or withdraw from the experiment for other reasons. The disadvantages are: The before-and-after study design…

18. 3. If the study population is very young and if there is a significant time lapse between the before-and-after observations, changes in the study population may be because it is maturing. The before-and-after study design …

19. c. Longitudinal study design Is used to determine the pattern of change in relation to time. visited a number of times at regular intervals, usually over a long period Longitudinal studies are also useful when you need to collect factual information on a continuing basis . A longitudinal study can be seen as a series of repetitive cross-sectional studies .

20. Longitudinal studies have the same disadvantages as before-and-after studies, in some instances to an even greater degree. In addition, longitudinal studies can suffer from the conditioning effect . This describes a situation where, if the same respondents are contacted frequently, they begin to know what is expected of them and may respond to questions without thought , or they may lose interest in the inquiry, with the same result. Longitudinal Study…

21. The main advantage of a longitudinal study is that it allows the researches to measure the pattern of change and obtain factual information requiring collection on a regular or continuing basis. Longitudinal Study…

22. 4.2.2. Reference period The reference period refers to the time-frame in which a study is exploring a phenomenon, situation, event or problem. Studies within this perspective are thus categories as:

23. A. The retrospective study design. Investigate a phenomenon, situation, problem or issue that has happened in the past . They are usually conducted either on the basis of the data available for that period or on the basis of respondents ’ recall of the situation. For example; The living conditions of indigenous people in Ethiopia in the early twentieth century. The utilization of land before World war II in Ethiopia. A historical analysis of migratory movements in Eastern Europe between 1915 and 1945.

24. B. The prospective study design Refers to the likely prevalence of a phenomenon, situation, problem, attitude or outcome in the future . Such studies attempt to establish the outcome of an event or what is likely to happen. Experiments are usually classified as prospective studies as the researcher must wait for an intervention to register its effect on the study population.

25. The following examples are classified as prospective studies: To determine, under field conditions, the impact of maternal and child health services on the level of infant mortality. To establish the effects of a counseling service on the extent of marital problems. The prospective study design …

26. C. The retrospective-prospective study design Retrospective-prospective studies focus on past trends in a phenomenon and study it into the future . A study is classified under this category when you measure the impact of an intervention without having a control group . In fact, most before-and-after studies, if carried out without having a control-where the base-line is constructed from the same population before introducing the intervention-will be classified as retrospective-prospective studies.

27. 4.2..3. The nature of the investigation On the basis of the nature of the investigation, studies can be classified as:

28. Experimental Vs Non-experimental In experimental design, researchers plan to measure the response variable depending on the explanatory variable. The response variable is an outcome measure for predicting or forecasting purposes of a study. It is also called dependent variable or predicted variable Any variable that explains the response variable is called explanatory variable. It is also called independent variable or predictor A true experimental design is one in which study participants are randomly assigned to experimental and control groups

29. Experimental … Although randomization is typically described using examples such as rolling dice, flipping a coin, or picking a number out of a hat, most studies now rely on the use of random numbers tables to help them assign their research participants These typically include demographic variables such as age, gender, level of education, and any other variables that are measured or

30. Quasi-experimental Design •Although the researcher plans to measure the response variable depending on the explanatory variable, there is a lack of randomisation in the quasi-experimental design •It is a mixed design where random and non-random experiments are employed together

31. Characteristics of a Good Research Design Neutrality:   should be free from   research bias   and neutral Reliability Validity :The   questionnaire   developed from this design will then be valid. Generalization :   The outcome of your design should apply to a population and not just a restricted   sample. Objectivity:

32. How and why sampling?

33. How and why sampling? The world is large and full of people We wanted to find out things about people Sampling is a practical way of studying people and their activities, thoughts, attitudes, abilities, relationships in relation to business Note that sample must be representative of the population from which it is taken

34. Population: is the total set of units in which a researcher is interested; Can be finite or infinite population Examples: All employees of an organization to study the reasons of employee turnover Sampling: Important terms

35. Sampling: Important terms Element/case: a single member of the population. Census: includes all the elements in the population conditions are appropriate for census study: the population is small ( for populations under 50 it is usually more sensible to collect data) Sampling : is the process of selecting units into a sample from a larger set of the same units (Population)

36. Sampling: Important terms Sampling frame: a listing of all the elements in the population from which the sample is drawn For example the list of employees found in personnel department to get information on employee turnover

37. Unit of analysis: the type of object whose characteristics the researcher wants to measure and study. For example: If data are collected on Employees, the unit of analysis is employee. Is the object that the hypothesis describes. All variables in a hypothesis must be operationalized for the same unit of analysis. Sampling: Important terms

38. Sampling: Important terms Sampling unit: a unit or set of units considered for selection at a stage of sampling. Sampling unit may or may not be the same as a unit of analysis. It is possible to include several units of analysis. For example, if the researcher wants to interview senior managers in the public sector, the senior managers become the unit of analysis and the public organisations across the country become sampling unit.

39. Parameter : is a characteristics of the population about which researchers are interested to find out. Example: The average income of all families in a city or the age distribution of the city’s population . Sampling: Important terms

40. Sampling: Important terms Statistics: characteristics of a sample and is developed from information about the member of the sample, are used to make estimates of population parameters Example: The mean income computed from a sample or the age distribution of that sample are statistics.

41. Sampling errors: the difference between population parameter and the statistical estimate. sampling error can be expressed through the use of confidence levels and confidence intervals . Example: being 95% confident that the population mean is between + or – of the sample mean Standard error: the standard deviation of the means of the sampling distribution . Sampling: Important terms

42. Sample size: the number of elements selected for the sample to represent the population. Sample size determination is influenced by: The purpose of the study Population size Sampling: Important terms

43. The three criteria to determine the appropriate sample size: 1. the level of precision, sampling error that is mostly expressed in percentage point, example : ±5% 2. the level of confidence or risk: based on the Central limit theory that states when a population is repeatedly sampled, the average value of the attribute obtained by those samples is equal to the true population value. Sampling: Sample size

44. Sampling: Sample size 3. the degree of variability in the attributes being measured ( Miaoulis and Michener, 1976). The more heterogeneous a population, the larger the sample size required to obtain a given level of precision. The less variable (more homogeneous) a population, the smaller the sample size.

45. STRATEGIES FOR DETERMINING SAMPLE SIZE using a census for small populations (e.g., 50 or less). imitating a sample size of similar studies, using published tables, and applying formulas to calculate a sample size Sampling: Sample size

46. If population size is known and the precision level is estimated (Yamane 1967); Where n is the sample size, N is the population size, and e is the level of precision. (Source: Glenn D. Israel, University of Florida) Example: There are 2000 households in a certain Woreda to assess the satisfaction level on the Woreda Administration service with the precision level ±5%. Sampling: Sample size …

47. Sampling Technique

48. Is the set of procedures for selecting the units from the population that are to be in the sample. Two major types of sample design Probable sampling technique: Non-probable sampling technique: Sample Technique

49. There are four types of probability sampling Simple random sampling (SRS): Systematic random sampling (SRS) Stratified random sampling (SRS) Cluster/Area sampling I. Probable sampling

50. 1.1 Simple Random Sampling Each unit in the population has equal chance of being selected. Can be lottery method or a random number table It requires a complete list of the study population. The researcher assigns each member of sampling frame a number before selecting sample units

51. Helps to eliminate the inadvertent introduction of sample bias. Example: assume there are 150 employees (with BA degree and above) in the organization with the problem of high employee turnover. If the sample size is 35 employees. Use lottery and random number table. to select the sample elements . Simple random sampling....

52. Procedures: For both lottery and random table case Identify the population: All employees with BA degree and above in the organisation The sampling frame: The list of employees with BA degree and above, names are sequentially numbered from 001 to 150 Simple random sampling....

53. Requires the complete list of population, Reduces the amount of effort required to draw a sample and provides adequate results Applicable when the researcher wants to pick households in the sample from the population of consecutive households found along a street/road. 1.2. Systematic Random sampling

54. Systematic Random sampling … Procedures : Population has N units. Plan to sample n units and then The sampling interval/skip= N/n------K Line-up all N units and Randomly select a number between 1 and K Select the randomly selected unit and every k th unit after that

55. Example: the list contains 10,000 element and you want a sample of 1,000: Sampling interval = Population size/Sample size=10 Randomly select a number between 1 and 10. Assume the first element in the sample is number 7, then the selection of elements continue as 7, 17, 27 …, 9987, 9997 Disadvantage Does not result in a truly random sample or suffers from the problem of periodicity. Systematic Random sampling …

56. Involves a process of stratification or segregation, followed by random sample from each stratum. 1 st : divide or classify the population into strata, or groups, on the basis of some common characteristics such as sex, race, or institutional affiliation, level of management, or income, etc. 1.3. Stratified Random Sampling

57. Stratified Random sampling… Mutually exclusive groups : the classification should be done so that every member of the population is found in one and only one stratum. Separate samples are drawn from each stratum. (proportionately or disproportionately). It ensures homogeneity within each stratum, but heterogeneity between strata Stratified sampling can also be done at several stages, and then called multistage stratified sampling.

58. Problem: The researcher wants to study about the satisfaction level of employees. Population: 800 employees sampling frame: all employees and the list of employees in personnel department. Sample size : 80 Procedure: stratify the population based on education level Stratified Sampling Example Educ level No. of empl Sampling fraction No. in sample PhDs 85 10% 9 MAs 150 10% 15 BA Degree 200 10% 20 Diploma Holders 130 10% 13 Highschool or below 235 10% 23 Total 800 10% 80

59. It involves division of elements of a population into groups-the groups are termed clusters Recommended when: it is necessary to study a large geographical area and It is difficult to identify the sampling frame The geographical distribution of the members is scattered 1.4 Cluster Sampling

60. Cluster sampling ... Stages in cluster sampling The sampling frame is the complete list of clusters rather than individuals Select a few clusters , normally using simple random sampling technique. then collect data from the cases within the selected clusters either using census or by taking sample .  Note: Cluster sampling can also be done at several stages, and then called multistage cluster sampling.

61. Example: The Problem: The AA city administration wants to assess the problem of transportation in AA Population: all households in Addis Ababa Sampling Frame: List of sub cities, list of Woredas , and list of Kebeles (Villages), List of households Cluster sampling …

62. Cluster sampling…. Procedure: construct a four -stage cluster sampling Randomly select a sub city from the lists of sub cities Obtain lists of Woredas from the selected subcity and randomly select a Woreda Obtain lists of kebeles / subworedas from the selected woreda and randomly select a kebele / subworeda Obtain lists of households from the selected kebeles / subworedas and obtain information from all households or by randomly selecting the sample elements. 

63. Different from stratified sampling because Every cluster is not sampled w here as every stratum is sampled in the case of stratified sampling. Saves time and money Disadvantage: it may require larger sample than other methods for the same level of accuracy Susceptible for the loss of key information as a result of random selection and re-selection process of groups. (better to use weight, based on the number of people living in the cluster, in the random selection process. Cluster sampling …

64. Four types Convenient sampling Quota sampling Purposive sampling (Expert sampling) Snow ball (referral Sampling) Non-probability sampling designs Can work well for exploratory studies Useful if it is not important to obtain accurate estimates of population characteristics The units are selected at the discretion of the researcher Cheaper and easier to carry out than probability designs II. Non probable Sampling

65. Non probable Sampling ... Some of the disadvantages of non-probability sampling : one cannot estimate parameters from sample statistics Such samples would not be a representative of the population : does not rely on random sampling

66. also called haphazard or Accidental sampling Involves collecting information from members of the population who are conveniently available to provide it. For example: collecting information from Volunteers Criteria : The availability/ the ease of obtaining/ and willingness to respond convenient and economical to sample employees in a nearby area During election times TV channels often present man-on-the-street interviews to reflect public opinion. 2.1. Convenience sampling

67. selecting a quota of individual units with defined characteristics in the same proportion as they exist in the population . address the issue of representativeness (gender: two categories: male, female; Class level: graduate and undergraduate, social-economic class: upper, middle, lower) 2. 2. Quota sampling

68. Quota sampling … Is called Dimensional sampling   If all dimensions of the population are considered in quota sample Example: A researcher is interested to assess the attitudes of employees towards working condition. male are 60 percent and female are 40% in the organizations : If Sample size is 30 employees, then 18 conveniently available male and 12 female workers will be sampled

69. 2.3. Expert /Purposive Sampling Expert sampling : involves selecting persons with known experience or expertise in an area. With purposive sampling the sample is ‘hand picked ’ for the research Example: A Local government uses purposive sampling when it seeks information from cities with a reputation for excellent administration (about their experiences, with outsourcing services, what performance measures they use, how they monitor citizen satisfaction).

70. Is judgmental/ deliberate sampling It invites the researcher to identify and target individuals who are believed to be typical of the population being studied. The researcher uses his own judgment about which respondents to choose, and picks only those best meet the purposes of the study. 2.3. Expert/Purposive sampling ….

71. Individuals are discovered initially, and then each individual is used to locate others (the names & addresses) who possess similar characteristics and who, in turn, identify others. 2. 4. Snowball sampling/referral sampling

72. Snowball sampling/referral sampling … Used when members of a population cannot be located easily by other methods and where the members of a population know each other. Example: we may want to sample very small populations who are not easily distinguishable from the general population or who do not want to be identified, example drug users, homeless people

73. Thank you!