Communication Disorders: An Introduction for Community-Based Rehabilitation Workers

5. Principles of Research

5 Principles of Research Empirical research Types of research Research designs can be experimental or descriptive. Experimental designs will be outlined below. Descriptive research can be: 1. Comparative research. This involves studying the performance of two different groups on one measure, e.g., deaf children and hearing children. 2. Developmental research. This measures change over time (longitudinal research). The independent variable is maturation. 3. Correlational research. Here, the interest is not in the differences between groups, but how one or many (dependent) variables affect another, independent variable. 4. Survey research. These are designed to examine attitudes, prevalence of disease, or life-style or work-style practices, and commonly include questionnaires and/or interviews. These need to be very carefully designed and there are standard procedures and criteria that should be followed in survey design. S. Case study and retrospective research. These designs allow you to follow one or more patients over time to look at how the individual develops or responds to treatment; retrospective research allows you to go back through the records of previous patients to look for trends. A major problem with retrospective studies is that the research question is formulated after the patients' files have been compiled, so there are likely to be gaps in the data. 28 Communication Disorders Stages in research There are common steps In the development and implementation of empirical research: 1. A statement of the problem. This includes a review of relevant literature, why questions need to be answered, a statement of the hypothesis, and the expected contribution to existing knowledge. There must be a clear, concise statement of the problem. 2. Method of investigation. This includes subject selection, materials and procedures. These three topics cannot be covered here. An overview may be found in any introductory text on research methodology. Briefly, this stage includes: (a) subject selection, and protection of subjects' rights (b) materials used, including instruments and training materials (c) procedures used in data collection 3. Results. Results include methods of data reduction (statistical analysis) and data presentation (use of tables and figures). 4. Conclusions. The conclusions are drawn from the results, and reflect the literature review in the statement of the problem. The focus of this section of the book is on step 3: results. Research design 1. Variables. Empirical research concerns the relationship between variables: usually referred to as independent and dependent variables. The independent (antecedent) variable is the presumed cause of the dependent (consequent) variable. Example: Is the type of cleft palate related to speech ability? The independent variable is the type of cleft (say, bilateral, right unilateral and left unilateral) and the dependent variable is speech, e.g., number of errors on an articulation test. 2. Types ofdata. Data can be collected in different ways, which determines the subsequent statistical procedures used in data reduction: (a) nominal data is categorical, e.g., female = 1, male =2, hearing impaired = 1, non-hearing impaired = 2 (b) ordinal data has categories that are related to one another, e.g., degree of hearing impairment may be mild (0), moderate (l), or severe (2); tasks may be described as easy (0), moderate (l) or difficult (2). (c) interval data is continuous data, e.g., ternperature measures, IQ scores. (d) ratio data is the equivalence of ratios among the values, e.g., sound intensity. 3. Data presentation. Raw data is often presented in tabular form, rather than by point by point reporting in the text. Commonly used forms are frequency tables, histograms, frequency polygons, and cumulative frequency distributions. These allow easy comprehension of the raw data. Principles of Research 29 4. Statistical summary. Once the raw data has been presented, the results must be summarized meaningfully. Measures of central tendency (mean, median, mode), and measures of variability (standard deviation, range and variance) are commonly used. These are explained in greater detail below. Basic statistics • The mean is the average. It is obtained by adding all the scores and dividing by the sample size. • The median is the score that splits the sample into two halves, i.e., it is the number that appears in the middle of the distribution. • The mode is the most frequently occurring score. Once you have the mean, you can determine how much the scores spread out around the mean (average amount of dispersion of the scores in the distribution): 1. for each score, subtract the mean from the score, square this figure 2. add all...