Paul J. Dejillas, Ph.D.
Professor of Anthropology
What kind of research are you intending to?
There are two kinds of research: pure and applied. Pure research--- also called fundamental, basic or university research--- involves developing and testing theories that are intellectually stimulating to the researcher, but has no practical applications. The main purpose is mainly to contribute to the general body of knowledge in a given discipline. This kind of research is more common in mathematics, chemistry, physics and related sciences.
Applied research addresses practical problems. The purpose is to solve problem existing currently in the work setting (industry, firm or school). This is the most common kind of research nowadays. Applied research is frequently descriptive and its main strength, according to Duane (1996), is its immediate practical use. People in businesses, government, social services and higher educational institutions especially conduct applied research. The results of the study are made available only to the one commissioning the research or to a small number of policy advocates who are in the position to use or not use the results of the study.
What research design are you going to use?
Research can be experimental or non-experimental in design. Between the two, t he experimental research design is more common. Experimental research investigates causal relationships and manipulates/controls conditions or variables. It is most effective when evaluating or program or policy.
There are three primary types of experimental research design:
1.Pre-experimental research design, which lacks random assignment to the program group and the control group. This design illustrates some inherent weaknesses in terms of establishing internal validity. The better designs are the true experimental and quasi-experimental designs.
2.Randomized or true experimental design, which is “a design in which subjects are randomly assigned to program and control groups. With this technique, every member of the target population has an equal chance of being selected for the sample “ (Abrahams,2000). This, according to Trochim (1977), is the strongest design for “establishing a cause and effect relationship, as random assignment is used - and the groups involved are considered equivalent.”
3.Quasi-experimental designs, which employs multiple measures or a control group without randomly assigning participants to group. It differs from true experimental design by the omission of random assignment of subjects to a program or control group. The closer your nonequivalent group approximates a true experimental population the stronger the interval validity.
Non-experimental research (NER) also investigate causal relationships, but does not manipulate conditions or variables. NER is typical survey research, content analysis and ethnography. Survey research “examines data to determine relative incidence, distribution and relationships of sociological or psychological variables.” Content analysis uses criteria or indicators based on research question in studying content, for example, of documents and instructional materials. Ethnography describes or analyzes in depth an intact cultural setting , for example, a class or school. NER designs “do not employ multiple measure, do not use a control group and do not use any random assignment in its design. These are usually descriptive studies conducting using a simple survey instrument only once. While they are useful in their own right - they are weak in establishing cause/effect relationship” (Trochim,1977; Abrahams,2000).
What method(s) of research will you use?
Your data decide your method. If your data are words, you will be doing qualitative research; if your data are numbers, then you will be doing quantitative research. The method can also be a mixture of both qualitative and quantitative research. Qualitative research is exploratory. Studies on ethnography and phenomenology are examples of this type of research. In ethnography, the researcher “emphasizes the observation of details of everyday life as they naturally unfold in the real world.” In phenomenology, the researchers “try to understand those whom they observe from the subject’s perspective.” The method that can be used in qualitative research are numerous: participant observation, direct observation, unstructured interviewing and case study. There are several other types of research that can be used. They include evaluation research, action research, cross-cultural and longitudinal. We will discuss some of these below.
What is evaluation research?
This is the widely-used type of applied research. It addresses the question: “Did it work” Examples of evaluation research question are: “Does flexi-time increase the productivity of full-time faculty members?” “Does the participatory method of teaching improve the learning absorption of students over the lecture method?” It is frequently descriptive but can also be exploratory or explanatory. Evaluation research is most appropriate for assessing educational programs, curricula, course materials, teaching instruments, methods and processes of teaching and learning outcomes. It can be used also for evaluating the academic performance of students as well as the teaching performance of faculty members. There are many known models that can be used to evaluate a phenomenon. In relation to HEIs, these models can be adopted or modified to serve as the framework and model of the research study (refer to the “References” at the end for a discussion of these models):
· Objective-Base model
· Tyler’s Model
· Provus’ Discrepancy Evaluation Model
· Input-Output model
· Stufflebeam’s Contex-Input-Process-Product Model
· Stake/Glass Model
· Scriven’s Formative and Summative Evaluation Model
Faculty members intending to do an evaluation study may simply adopt one of the above models and apply it to their respective school setting.
What is descriptive research?
Descriptive research tells only what happened. Often, there are no formal hypotheses. The purpose of the research is merely to learn more about the subject or topic. It merely attempts to describe a phenomenon. Descriptive research also focuses on the “how” and “who” questions. “how did it happen?” Who were involved?
What is case-study research?
This type of research examines several features of a few cases over a duration of time. The subject of study may be individuals, groups or institutions. The data are mostly qualitative. The purpose is to look for patterns across individuals, groups, movements or institutions. They can also be events and geographic units.
What is survey research?
Survey research asks people questions in a written questionnaire or during an interview. The researcher then summarizes the answers in percentage, averages using tabular or graphic presentations. Surveys give the researcher a picture of what many people think or doing. Oftentimes, a sample is used, but generalizes results to apply to a larger group from which the smaller group is chosen. Survey techniques are often used in descriptive or explanatory research.
What is content-analytical research?
This is a type of research that gathers and analyzes the content of a text or document. The content may refer to words, meanings, pictures, symbols, ideas, themes or any message that can be communicated. The document can be a book, an article, a speech, an advertisement, a film or videotape, musical lyric, photograph or a work of art. The basic goal of content analysis is to take a non-qualitative document and transform it into quantitative data. Thus, the results of content are generally presented in tables containing frequencies or percentages. Content analysis is used for exploratory and explanatory research, but is often used in descriptive research.
What is action research?
Action research treats knowledge as a form of power. It seeks to empower those who are being studied. Thus, it allows the respondents to participate in the research process. Action research usually seeks to raise the consciousness or awareness of the respondents on some social or political issues. The end result of action research is political action, i.e., mobilizing respondents or those affected to do something to correct a given situation.
What is historical-comparative research?
This examines aspects of social life in the past. Historical research may focus on one or more historical period (longitudinal), compare different cultures (cross-cultural), or mix historical periods and cultures (historical-cross cultural). It combines theory with data collection. It uses a mix of evidence, e.g., existing statistics, documents, observations and interviews.
What research instruments will you use to gather data?
Research instruments are devices used to collect and consolidate data. The most common instruments include the questionnaires and interview schedules. There are advantages and disadvantages in using a questionnaire or an interview. In the interview, the researcher needs interviewers, which may have to be trained and paid for their work. But the rate of return of the interview schedules can be 100 percent. Questionnaires, on the other hand, may require some mailing expense and the rate of return may be less than 50 percent.
How do you ensure the validity and reliability of your study?
Validity refers to “the strength of our conclusions, inferences or propositions.” It refers to the “truthfulness” of the study. Are your findings right? There are four methods of checking validity, construct validity and conclusion validity.
1. Content validity can be checked by asking expert in the fields. Because of their expertise and experience, experts are aware of nuances of the topic you are studying.
2. Criterion validity relates to another measure of the same variable. There are two types of criterion validity: predictive and concurrent validity.
· Predictive validity refers to the predictive ability of the findings of the study.
· Concurrent validity refers to the “relationship between scores on the instrument and some external criterion.”
3. Construct validity becomes “a concern only when the investigator is examining how well the instrument measures some psychological construct (such as anxiety), and frequently is measured by a factor analysis.
4. Conclusion validity relates “to the statistical power of your study, and is the first type of validity to deal with in conducting any study – as it addresses the attempt to determine a relationship between your variables and your outcome (Trochim, 1977; Rymarchyk, 2000; Hedden, 2000).
Meanwhile, reliability refers to “the consistency of the measurement or the degree to which an instrument measures the same way each time it is used under the the same condition with the same subjects.” It is “the repeatability of your measurement; a measure is considerd reliable if a person’s score on the same test given twice is similar (Trochim. 1977). There are many ways that reliability is usually tested: testing and re-testing, comparing opinions of two or more observers, checking for internal consistency and giving two equivalent forms of research instruments.
Who are the subjects of your study?
The subject of your research may be persons or documents. The document can be a book, a historical record, an article or speech, work of art, symbols and the like. As subjects of research, “persons” are the individual human beings from whom the researcher gathers the data and information needed to respond to the problems raised in the study. In actual research, persons can refer to faculty members, students, public officials or even the entire population. When the subject of research involves dealing with thousands or millions of individuals, the researchers may have to employ sampling, especially if he/she is working on a very limited budget
What is sampling?
Sampling is the process of selecting a representative part of a population for purpose of determining the characteristics of the whole population (Mugo, 2000). When dealing with people, a sample is a set of respondents (people) selected from a larger population for the purpose of a survey. There are two types of sampling: random and non-random sampling.
Random sampling can be simple, systematic, stratified, cluster or multi-stage.
Simple random sampling gives every member of the population an equal chance of being selected. Selection may be done using the random number table found in many basic texts in statistics or sampling. Systematic random sampling includes only every kth (e.g. every fifth) person in the list. This is done if there is no available random number table or if using such a table is cumbersome. Stratified random sampling divides the total population into groups or strata based on some characteristics, e.g., income, education, etc. A separate simple random sample from each group is then selected. Cluster random sampling divides the total population into clusters (e.g., all public schools, a simple or systematic random sampling can then be applied (Mugo, 2000).
Non-random sampling can be purposive, i.e., dependent on the purpose of the study – often based on the research problems. It “selects information-rich cases for dept study” (ibid.) Patton (1990: 169-196) discusses 16 different types of purposeful sampling. These are: extreme and deviant case, intensity, maximum variation, homogenous, typical case, stratified, critical case, snowball or chain, criterion, theory-based or operational construct, confirming and disconfirming cases, opportunistic, random purposeful, sampling politically-important cases, convenience, and combination or mixed purposeful.
How does one estimate the sample size?
Cochran (1977) gives the following formula for estimating the sample size:
N where: n = required sample size
n = ----------------- N = total population
1 + N (e)2 e = standard error
Meanwhile, the Social Weather Station estimates the sample size by measuring the sampling error or S.E (measure of difference/deviation of sample value from the corresponding population value) using the following formula:
S.E = ----------
For example, an S.E of less than 10 percent gives an n of 1,200. Before estimating the sample size, it is important to define your study population, e.g., only full-time faculty members in the College of Arts and Sciences. Then, the researcher prepares comprehensive list of the names who belong to this category. Then, the sample size can be estimated based on this list. The question of how large a sample size should be is a difficult one. F. Mugo says sample size can be determined by various constraints, e.g., available finding. According to her, when research costs are fixed, a useful rule –of-thumb is to spend about one half of the total amount for data collection and the other half for data analysis. Some international research funding agencies are adopting this formula.
What are some of the basic tools for quantitative research?
There are several statistical tools that can be used to analyze quantitative data. These tools are given below.
1. Summary Statistics
· Measures of location or central tendency: mean, median, mode.
· Measures of dispersion, spread or variability: range, mean deviation, standard deviation, etc.
2. Correlation – seeks to show if a relationship between two quantitative data.
3. Multivariate Statistics
4. Multiple Linear Regression – used analyse and predict the influence of one or more independent variables on the dependent variables.
5. Econometric Models: quadratic, cubic, logarithmic, exponential.
6. Forecasting: time-series, ARIMA, neutral network system, seasonal/cyclical movement.
7. Linear programming
Unlike before, it is no longer difficult to use and apply the above statistical tools because of the availability of computers and statistical packages, e.g., SPSS, SAS, QM, etc. Nowadays, responses of the respondents that are coded in specifically-designed answers sheets can even be read by a scanning machine in a matter of a few minutes and made ready for any software applications. Thus, there is no more needed nowadays to manually code the data from the questionnaire to the computer. The reader is referred to the “References” or to some basic text in statistics for a discussion on the above quantitative tools. We will only deal here with multivariate techniques since these are getting common today, but are yet the least understood and least applied in HEI research.
What is multivariate statistics?
This is an array of techniques used for data analysis. There are several types of multivariate analytical techniques. These include multivariate ANOVA, multiple discriminant analysis, covariance structure analysis, which are used to analyze multivariate dependency structures or dependent-independent relationship. The other set of multivariate techniques are used when the focus of analysis is on a whole set of interdependent relationships and the interest of the researcher is to summarize the information in the set by using smaller sets. These techniques include factor analysis, cluster analysis, and multidimensional scaling (See Duane, 1996; Finn and Mattsson, 1978). Multivariate techniques are employed to develop taxonomies or systems of classification, investigate useful ways to conceptualize or group items, and to generate or test hypotheses (Trochim, 1977).
What are some of the basic tools for qualitative analysis?
There are several qualitative tools for analysis that are commonly used by researchers. Some of the most common include: meta-analysis, content validation, delphi techniques, case simulations, historical and phenomenological. Many basic texts on research methods deal with these analytical tools.