Chapter 2 2

Research Problem, Variables and Hypothesis

Chapter Two

1

What is a Research Problem? 





It is a problem that someone would like to investigate. It is considered a situation that needs to be changed or addressed. These problems involve:    

Areas of concern. Conditions to be improved. Difficulties to be eliminated. Questions seeking answers.

2

What is a Research Problem? 





It is a problem that someone would like to investigate. It is considered a situation that needs to be changed or addressed. These problems involve:    

Areas of concern. Conditions to be improved. Difficulties to be eliminated. Questions seeking answers.

2

Research Questions 





Usually a research problem is initially posed as a question, which serves as the focus of the researcher’s investigation. Proper research questions need collectable data to create possible answers. Research questions must have an empirical (based on observable evidence) referent(s).

3

Research Questions Example of question:  “Should philosophy be included in high school curriculum?”   The question is a question of value- it implies notions of right and wrong, proper and improper and therefore does not have any empirical (observable) referents.  There is no way to deal empirically with the verb should.  How can we empirically determine whether or not something “should” be done?  What data could we collect?  There is no way to proceed.  Thus, this is a not researchable question. 

4

Research Questions The question is modified:  “Do you think philosophy should be included in high school curriculum?”   The question can be researchable.  We can collect data to help us answer the question. 

5

Research Questions Example of question:  “Is God good?”   This is an example of metaphysical (beyond physical, transcendental) question, which does not have empirical research.  Therefore it is not researchable.  Change the question to researchable question:  “Do you believe God is good?”  

6

Characteristics of Good Research Questions 

Good research questions possess four essential characteristics: 







The question is feasible (can be done with available resources of time and money). The question is clear (most people would agree as to what the key words in the question mean). The question is significant (worth investigating because it will contribute important knowledge). The question is ethical (it will not bring physical or psychological harm to human being).

7

Characteristics of Good Research Questions 





Some of the terms in the questions are ambiguous and difficult to define because different people may have different meaning. Researcher has to be specific about the terms used in a research question, to define precisely what is to be studied. How the clarity of a research question be improved?

8

Key terms to define in a research study 



 



Terms necessary to ensure that the research question is sharply focused. Terms that individuals outside the field of study may not understand. Terms that have multiple meanings. Terms that are essential to understanding what the study is about. Terms to provide precision in specifications for instruments to be developed or located. 9

Defining Terms 

There are 3 ways to clarify important terms or meaning in a research question: 1) use of constitutive definition (the dictionary approach). 2) use of proper example(s). 3) use of operational definition (specifying operations used to measure or identify examples of the term).

10

Common Problems for Student Research 

Student justifications for a proposed study are likely to have two weaknesses: 1) The assumption is too great that everyone will agree that the stated problem is a ‘problem’. 

It is the researcher’s job to clarify and prove that the study or question is important.

2) Students often overstate the implications of a study. 

Evidence of the effectiveness of a certain treatment doesn’t mean that this will occur or result automatically when adopted or implemented.

11

Example 1 Research Problem: Malay speech recognition Define terms: 1) The terms are too general. 2) Speech can be adults or children speech. 3) Dependent on age/race of the subjects. 4) Types of speech: isolated, syllabus, phoneme… 5) Methods of speech recognizers: neural networks, Hidden Markov Models, Support Vector Machines… Modified research problem: Malay vowel speech recognition of Malay children using neural networks.

12

Example 2 Research Problem: Fundamental frequency analysis of Malay children Define terms: 1) Types of speech: sustained vowel, reading, conversation. 2) Age range of the children: 7-12 years old,.. 3) Males, females or both. 4) Normal children or children with disorders.

Modified research problem: Fundamental frequency analysis of Malay children for sustained Malay vowels between 7 and 12 years old 13

Problem Statement 









Selection of research problem does not necessarily mean that it is adequately stated. A research problem needs some working to get in into a suitable form. A research problem can be stated broadly and then systematically restricted through a review of the literature. Research problems can be stated in a declarative or descriptive manner or in question form. The most important characteristic of the problem statement is that it must provide adequate focus and direction for the research.

14

. 

Problem Statement Components/Aspects: the topic, the research problem, the justification, the deficiencies and the audience.

Example Research problem: F0 and perturbation measures of Malay children between 7 and 12 years old Problem Statement: Fundamental frequency (F0)has been reported for various languages such as American English, British English, Mandarin, Japanese and so on. So far, the acoustic characteristics are not well established for sustained vowels among Malay children. We need the data for clinical diagnosis of patients with pitch disorders as well as to describe the vocal acoustic characteristic differences globally.

16

What is a Variable? 





A variable is a variation within a class of objects (e.g., eye color). A constant is a characteristic with no variations within a class of objects (e.g., grade level of ninth graders). Researchers choose certain variables to study since they are suspected to be related to a possible relationship to be discovered.

17

Quantitative vs. Categorical Variables 



Variables are classified as either quantitative or categorical. A quantitative variable exits in some degree (rather than all or none) along a continuum from less to more (e.g., height). 



Can be subdivided into smaller units (e.g. miles, yards, feet, inches).

A categorical variable does not vary in degree, amount, or quantity, but in quality (e.g. gender). 

There is no middle ground or in-between measurement. 18

Quantitative vs. Categorical Variables (Figure 3.1)

19

Independent vs. Dependent Variables 

The independent variable is what the researcher studies to see its relationship or effects. 



The dependent variable is what is being influenced or affected by the independent variable. 



Presumed or possible cause.

Presumed results

Independent variables may be either manipulated  or selected. 



A manipulated variable is a changed condition the researcher creates during a study, also known as an experimental or treatment variable. A selected variable is an independent variable that already exists. 20

The Relationship between Independent and Dependent Variables

21

Other Types of Variables 

Outcome Variable  





A dependent variable that is quantitative. Most studies demonstrating this would have a categorical independent variable. The investigator is interested in the effects that might lead to differences in methods = outcomes.

Moderator Variable 



Special type of independent variable/secondary independent variable. Selected to determine if it affects (modifies) the basic relationship between the primary independent variable and the dependent variable. 22

Other Types of Variables (cont.) 

Extraneous Variable 

 



Independent variables that have not been controlled. Considered a basic problem in research. The task is to control these variables by reducing or eliminating their effects. A method to control extraneous variables is to hold them constant.

23

Example 1 Research problem: Investigation of the effect of the signal length on the Malay vowel speech recognition of Malay children using neural networks. Independent variable: signal length of the speech sounds. Dependent variable: results of speech recognition. Selected independent variables: different genders, different age groups, type of speech, types of neural networks. Extraneous variables: gender, age group, type of speech, types of neural networks.

24

Example 2 Research problem: Fundamental frequency analysis of Malay children for sustained Malay vowels between 7 and 12 years old.

Independent variable: age of the children. Dependent variable: fundamental frequency of Malay children. Selected independent variables: different genders, different age groups. Moderator variables: effect of different vowels on the fundamental frequency. Extraneous variables: gender, age group, type of speech. 25

What is a Hypothesis? 



A hypothesis is a prediction regarding the possible outcome of a study. Advantages of stating hypotheses include: 







Forces us to think more deeply and specifically about the possible outcomes of the study. Enables us to make specific predictions based on prior evidence or theoretical argument. Helps to clarify whether we are or aren’t investigating a relationship.

Disadvantages of stating hypotheses include:  



May lead to a bias on the part of the researcher. In some studies, it would be presumptuous to predict what findings would be. Focusing on the hypothesis could prevent the researcher from seeing other phenomena that might be important to the study. 26

Directional vs. Non-directional Hypotheses 

A directional hypothesis is one in which the researcher indicates the specific direction that he or she expects will emerge in a relationship in the study. 

The direction is based on what the researcher has found from:   



Literature. Personal experiences. Experience from others.

A non-directional hypothesis is when there is no specific prediction about what direction the outcome of a study will take. 

Sometimes it is difficult to make specific predictions upon a study 

A study may state that it would point to non-specific directions vs. distinct possibilities.

27

Example 1 Research problem: Investigation of the effect of the signal length on the Malay vowel speech recognition of Malay children using neural networks. Hypothesis: 1) Signal length does have effect on the accuracy of Malay vowel speech recognition. 2) Different types of neural networks may have different performance of the Malay vowel speech recognition.

28

Example 2 Research problem: Fundamental frequency analysis of Malay children for sustained Malay vowels between 7 and 12 years old 1) 2) 3) 4)

Different age of the children has different fundamental frequency. Younger children has higher fundamental frequency compared to older children. Females have higher fundamental frequency than males. Fundamental frequency of children varies in different vowels.

29