Sathyendran S A

Ways which are useful for drawing inferences about a population from a probability sample are called inferential statistics. They are used to describe a population using merely information from observations on a probability sample of cases from the population, Thus, the same statistic can be descriptive or inferential or both, depending on its use. 1.2 Significance Statistical tests look for significance, a concept that measures the degree to which your results can be obtained due to chance. In social science/educational research the term α = .05 is often used. This means there is a 5% or less chance that the results are due to chance. 1.3 Hypothesis When we conduct quantitative research we will often be concerned with finding evidence to either support or contradict, an idea or hypothesis. For example, you might propose that if you give a student training in how to use a search engine it will improve their success in finding information on the Internet. In hypothesis testing we generally have two hypotheses: 1) a null hypothesis (which usually indicates no change or no effect) and 2) an alternative hypothesis (which is usually our experimental hypothesis). The evidence from the sample is taken to support either the null or the alternative hypothesis. When a researcher is interested in hypothesis testing they will conduct an experiment to gather their data. So, we could take one sample from our population of students, give them some training in how to search and then ask them to find some specific information. We ask another sample of students to search for the same specific information but don't give them training - and we see which group did better through a variety of different measures, some subjective and some objective. So, does the data we gather contain evidence that agrees with the alternative (experimental) hypothesis or the null hypothesis? In testing a hypothesis we never actually prove or disprove a hypothesis, all we ever get is evidence from a sample that either 1) supports a hypothesis or 2) contradicts a hypothesis. The Hypothesis contains concepts which need to be measured. To do this we need to:  translate concepts into measurable factors  take these measurable factors and treat them as variables  identify measurement scales to quantify variables 1.4 Causality - cause & effect This is essentially concerned with showing how things come to be the way they are. To do this we need to identify our variables:  Independent variable - the variable that is deliberately manipulated by the researcher  Dependent variable - the variable that is measured to find out the effect of the manipulated (independent) variable  Control variables - may be potential independent variables, but are held constant during the experiment So, following on with our example, students are timed whilst searching for information to assess the effectiveness of their searching behavior, some were given prior search training  Independent variable = training - manipulate by varying training given to different students  Dependent = time taken to find information - which we can measure by timing how long to search  Control = searching behavior may be affected by previous use, age, educational level, and even time of day. Some of these may be controllable but others may not be, e.g. degree of frustration So, our experimental (or alternative) hypothesis is that if we give more training it will take less time to search and conversely if we give less training it will take more time to search we have a cause (training) and effect (time taken). The null hypothesis is that there will be no change or effect. Independent variables are assumed to have a causal impact on the dependent variable Wikipedia defines causality as ' a necessary relationship between one event (called cause) and another event (called effect) which is the direct consequence (result) of the first' 1.5 Generalizability or external validity Generalizability or external validity involves the extent to which the results of a study can be generalized (applied) beyond the sample to the larger population. In other words, can you apply what you found in your study to other people (population validity) or settings (ecological validity). For example, a study of postgraduate Masters Students in a UK university that found one method of teaching statistics was superior to another may not be applicable with first year undergraduate students (population) in an American university (ecological). 1.6 Reliability or internal validity Reliability or internal validity is concerned with repeating a piece of research in order to establish the reliability of its findings. Reliability is the consistency and dependency of a measure. Sometimes it is referred to as the repeatability or the test-retest reliability. This means that a reliable test should produce the same results on successive trials. 2. Some techniques for quantitative data analysis Data Analysis Technique 1: Distribution of Data Set The data that we want to analyze can be displayed in a rectangular or matrix form, often called a data sheet (see table 1).