Research Instrument Validity and Reliability PDF

Research Instrument, more appropriate and accurate Validity and Reliability information. 4. Easily Tabulated - Since you will be Research Instruments are basic constructing an instrument for tools researchers...

Research Instrument, more appropriate and accurate Validity and Reliability information. 4. Easily Tabulated - Since you will be Research Instruments are basic constructing an instrument for tools researchers used to gather quantitative research, this factor data for specific research problems. should be considered. Hence, before Common instruments are crafting the instruments, the researcher performance tests, makes sure that the variable and questionnaires, interviews, and research questions are observation checklist. established. These will be an important basis for making items in the research The first two instruments are instruments. usually used in quantitative research, while the last two Ways in Developing Research instruments are often in qualitative Instrument research. However, interviews and 1. Adopting an instrument from the observation checklists can still be already utilized instruments from used in quantitative research once previous related studies. the information gathered is 2. Modifying an existing instrument translated into numerical data. when the available instruments do not yield the exact data that will Characteristics of a Good answer the research problem. Research Instrument 3. Researcher made his own 1. Concise - Have you tried instrument that corresponds to the answering a very long test, and variable and scope of his current because of its length, you just study. pick the answer without even reading it? A good research Common Scales Used in Quantitative instrument is concise in length Research yet can elicit the needed data. Likert Scale. This is the most common 2. Sequential - Questions or scale used in quantitative research. items must be arranged well. It Respondents were asked to rate or is recommended to arrange it rank statements according to the scale from simplest to the most provided. complex. In this way, the instrument will be more Semantic Differential. In this scale, a favorable to the respondents to series of bipolar adjectives will be answer. rated by the respondents. This scale 3. Valid and Reliable - the seems to be more advantageous since instrument should pass the tests it is more flexible and easy to of validity and reliability to get construct. Types of Validity of Instrument Reliability of Instrument Validity. A research instrument is Reliability refers to the consistency of considered valid if it measures the measures or results of the what it supposed to measure. instrument. 1. Face Validity. It is also known 1. Test-retest Reliability. It is as “logical validity.” It calls for achieved by giving the same test an initiative judgment of the to the same group of respondents instruments as it “appear.” Just twice. The consistency of the two by looking at the instrument, scores will be checked. the researcher decides if it is 2. Equivalent Forms Reliability. It valid. is established by administering 2. Content Validity. An two identical tests except for instrument that is judged with wordings to the same group of content validity meets the respondents. objectives of the study. It is 3. Internal Consistency Reliability. done by checking the It determines how well the items statements or questions if this measure the same construct. It is elicits the needed information. reasonable that when a Experts in the field of interest respondent gets a high score in can also provide specific one item, he will also get one in elements that should be similar items. There are three measured by the instrument. ways to measure the internal 3. Construct Validity. It refers to consistency; through the split-half the validity of instruments as it coefficient, Cronbach’s alpha, corresponds to the theoretical and Kuder-Richardson formula. construct of the study. It is concerning if a specific Instrumental Development measure relates to other measures. Another important step in preparing 4. Concurrent Validity. When the for your research is determining the instrument can predict results appropriate data collection method similar to those similar tests from which you should determining already validated, it has your research instrument. concurrent validity. If you are doing a quantitative 5. Predictive Validity. When the research, you may use test instrument is able to produce administration, observation and the results similar to those similar most common which is the tests that will be employed in questionnaire. the future, it has predictive validity. This is particularly useful for the aptitude test. Qualitative research usually use If you are including a survey of the observation, interview and the profile of your respondents, the first focus group discussion. If you part of the questionnaire should be are doing the mixed method, that easy for you. is, both qualitative and quantitative methods, then you Things might become complicated as have to use the collection methods you go to the second and third parts. proper to both. At this point, the theoretical framework and review of related The basic question in determining literature and studies will be of great the collection method is how you help. will be able to gather data to answer your specific questions as Data Collection and Analysis required by your study. Three Phases in Data Collection Data on the study of bullying would be best gathered with the Before aid of a questionnaire while data prepare the research instruments on the study on breastfeeding identify the authorities that will practices would be best gathered be involved and need to ask with an observation, with the aid of permission secondary data and a determine the samples size and questionnaire. corresponding respondents; per group if applicable. After determining the appropriate ask consent form (if respondents data collection method, it is time are 18 years old above) or for you to write the instrument, parent’s consent (if minor). considering the guidelines in the pilot test the research instrument preparation. if needed If you have chosen the During questionnaire, you have to ensure clear the instructions provided to that each specific question will be the respondents answered, and that no unnecessary administer the research question is included. instrument or implement the research intervention, if Therefore, it is advised that you applicable always go back to the specific collect or gather or take note of questions while writing the the responses instrument to check if each item has been represented in the questionnaire. After research sample. Frequency summarize the data gathered, distribution, measure of central in a tabular form tendencies (mean, median, mode), analyze the summarize data and standard deviation are the sets corresponding to the research of data from descriptive statistics. questions. Inferential Statistics is used when Data Analysis the research study focuses on finding Data analysis in research is a predictions; testing hypothesis; and process in which gathered finding interpretations, information are summarized in generalizations, and conclusions. Since such a manner that it will yield this statistical method is more answers to the research questions. complex and has more advanced During quantitative data analysis mathematical computations, you can gathered information were break use computer software to aid your down and ordered into categories analysis. in order to draw trends or patterns in a certain condition. In You also have to identify types of quantitative research, the statistical analysis of variable in your numerical data collected is not quantitative research. A univariate taken as a whole. In order to analysis means analysis of one understand it better, it is analyze variable. Analysis of two variables into components based on the such as independent and dependent chosen research variables and variables refers to bivariate analysis research questions you are going to while the multivariate analysis answer. involves analysis of the multiple These numerical data are usually relations between multiple variables. subject to statistical treatment depending on the nature of data and the type of research problem presented. The statistical treatment makes explicit the different statistical methods and formulas needed to analyze the research data. Test of Relationship between Two Variables Planning your Data Analysis Pearson’s r (parametric) Phi coefficient (non-parametric Descriptive Statistical Technique for nominal and dichotomous provides a summary of the ordered variables) or sequenced data from your Spearman’s rho (non-parametric for ordinal variable) Test of Difference between Two by telephone, or through e-mail, Data Sets from One Group messenger, social media post, or using T-test for dependent samples computer programs or forms (Barrot, (parametric) 2018, p 138). McNemar change test (non- parametric for nominal and In quantitative research, dichotomous variables) questionnaires use the following Wilcoxon signed-rank test (non- approaches: (1) scale (usually Likert parametric for ordinal scale); and (2) conversion of variable) responses into numerical values, e.g. strongly as 5, agree as 4, neutral as 3, Test of Difference between Two disagree as 2, and strongly disagree Data Sets from Two Different as 1. Groups T-test for independent samples The terms survey and questionnaire (parametric) have different meanings. A Two-way chi-square (non- questionnaire is an instrument used to parametric for nominal collect data while a survey is a variable) process of collecting, recording, and Mann-Whitney U test (non- analyzing data. Questionnaires can parametric for ordinal be structures, semi-structured, or variable) unstructured. Test More than Two Population There are three structures of making Means a questionnaire. The first, structured Analysis of Variance or questionnaires employ closed-ended ANOVA (parametric) questions. Unstructured questionnaires, on the other hand, Test the Strength of Relation or use open-ended questions in which the Effect or Impact research participants can freely Regression (parametric) answer and put his thoughts into it. Lastly, semi-structured questionnaires are combinations of Research Instruments both the structured and unstructured Questionnaires ones. Structured type is commonly A questionnaire consists of a series used in quantitative studies because it of questions about a research topic to is easier to code, interpret objectively, gather data from the participants. It and, most of all, easier to standardize. consists of indicators that is aligned to the research questions. Gathering of information can be carried out in the following methods: face to face, Advantages of Using Non-standardized test may not be Questionnaire scored uniformly. It is administered to Data can be quickly gathered from a certain set of people. a big number of participants. The participants are encouraged to Type of Test Questions be open to the researchers since Recall Questions. It requires their identity can be made participants to retrieve information anonymous. from memory (e.g. fill-in-the blank test, identification test, enumeration Disadvantages of Using test, etc.) Questionnaire Recognition Questions. It provides The questions can be interpreted respondents to select from given differently by the participants. choices the best or correct choice (e.g. Some participants will not be able multiple-choice test, true or false test, to complete the required responses. yes or no test, etc.) Some questionnaires will not be Open-ended Questions. It allows the returned on time. respondents more freedom in their Some questionnaires will be lost. responses, expressing their thoughts The answers from participants may and insights (e.g. essay writing tests lack depth. and other performance-based tests. Tests Type of test Tests are used for assessing various skills and types of behavior as well as for describing some characteristics. There are two types of test used in quantitative research: Standardized test and Non-standardized test. Standardized test is scored uniformly across different areas and groups. It is usually administered by institutions to assess a wide range of groups such as students and test-takers. It is considered as more reliable and valid. Examples are Achievement test, University Entrance Exam, Personality Tests, and the likes.  Observation Observation is another method of collecting data that is frequently used in qualitative research. However, it can be used in quantitative research when the observable characteristics are quantitative in nature (e.g. length, width, height, weight, volume, area, temperature, cost, level, age, time, and speed) Observation allows the researcher to observe the actual event or phenomenon. It has greater flexibility in the observation method. However, observation may lack participant validity and may be prone to the Hawthorne Interview effect phenomenon. Furthermore, it is A quantitative interview is a more exhausting and time-consuming method of collecting data about an especially when observations need to be individual’s behaviors, opinions, conducted for many years. values, emotions, and demographic characteristics using numerical Forms of Observation data. Controlled Observation. It us usually used in experimental research and is done under a standard procedure. It provides more reliable data (obtained through structured and well-defined process). The procedure can be replicated, and the data are easier to analyze. Lastly, the observer platforms a non-participant tole (i.e. does not interact with the participants). Natural Observation. It is carried observer is known and recognized by out in a non-controlled setting. It the participants as part of their group has greater ecological validity (i.e. or community but has limited flexibility of the findings to be interaction with them. generalized to real-life contexts). It Participant as observer - researcher also responds to other areas of does not primarily aim to capture inquiry not initially intended by the spontaneous or natural behavior. The researcher. Its major limitation is its observer is considered a researcher strength to establish a causal and not a member of the group or relationship due to the presence of community which is investigated. extraneous variables which can However, the observer is fully engaged affect the behavior of the with the participants. participants. Complete participant - researcher is to observe the spontaneity or natural Participant Observation. It allows behavior of the participant with as few the observer to become a member external factors affecting it as possible. of the group or community that the Taking this role raises ethical issues participants belong to. It can be about possible deception. He acts like performed covertly (i.e. an “undercover” member of the participants are not aware of the community. He fully involved with the purpose behind the observation. It participants who are unaware that they can be done also overtly, wherein are being observed. participants know the intention or objectives of the observation. Guidelines in Conducting an Observation Different Roles of a Researcher Before the Observation during a Participant Observation clarify the research objectives be ready with observation checklist Observational Role determine your time of observation Complete observer - Observe the and collection technique spontaneous natural behavior of record only relevant information the participants with as few prepare consent form for human external factors affecting it as observation possible. The observer is neither During the Observation seen nor noticed by the make sure to note the date, time, participants. The participants are and place where the observation is less prone to the Hawthorne effect. conducted Observer as participant - limit the person who will observe researcher does not prioritize explain the purpose of the recording spontaneous or natural observation to participants if behavior with as few external applicable factors affecting it as possible. The  avoid causing disruptions while observing make sure to guarantee anonymity of participants make use of multiple recording schemes such as taking notes, videos, audios, and photos but with proper ethical considerations After the Observation code or arrange the data gathered according to the variables of the study tabulate the data, use numerical value for quantitative research review the data gathered to determine the gaps before proceeding