Quantitative Data Analysis Assignment (Recovered)

Analyzing quantitative dataI. Descriptive statisticsThe first step in data analysis is to describe or summarize the data using descriptive statistics. The purpose of descriptive statistics is to enable the researcher to meaningfully describe a distribution of scores or measurements using a few indices or statistics. Each statistics or indices used depend on the type of variables in the study and the scale of measurements used (ratio, interval, and ordinal, nominal). (Mugenda and Mugenda, 2003)Techniques of quantitative data analysis1.1 TabulationWhen a mass of data has been assembled, it becomes necessary for the researcher to arrange the same in some kind of concise and logical order. This procedure is referred to as tabulation. Thus tabulation is the process of summarizing raw data and displaying the same in compact orderly form for further analysis. Tabulation is essential because of the following reasons;a. It conserves space and reduces explanatory and descriptive statement to a minimum.b. It facilitates the process of comparisonc. It facilitates the summation of items and the detection of errors and omissions.d. It provides a basis for various statistical comparisonsTabulation can be done by hand or by mechanical or electronic devices. Hand tabulation is usually preferred in case of small inquiries where the number of questionnaires is small and they are relatively of short length. It may be done using the direct tally, the list and tally or the card sort and count method.1.2 Graphical representationA graph should be well labeled on both the vertical and horizontal axes. A graph should also have a title. There are three types of graphs commonly used to present data in research reports. These are histograms, frequency polygons and bar charts.1.2.1 HistogramsIt comprises of a series of adjacent bars whose heights (y-axis) represent the number of subjects obtaining a particular score or number of respondents belonging to a particular category. The scores are usually on the horizontal axis (x-axis).

1.2.2 Frequency PolygonsA polygon is a many sided figure, hence the name frequency distribution. In plotting a frequency polygon, one must establish the midpoint of the class interval. The midpoint is established by summing up the lower and the upper class limits of each class interval and dividing by two. e.g.Lower class limit 20.5Upper class limit 25.5Mid point (20.5+25.5) + 2 = 23The midpoints are then plotted against the frequencies and the points joined using straight lines. A frequency polygon is closed figure and so the ends meet the horizontal line one unit after the highest score and one unit from the lowest score.Commuting Distance of Employees0246810121416024681012141618DistanceNumber of Employees

1.2.3 Bar chartsBar charts are preferred when data is discrete or categorical or when the scale is nominal or none ordered. This is mainly because the categories in a nominal scale do not apply any order. The bar chart is very much like the histogram except that spaces are left between the bars to signify a lack of continuity or flow between categories.e.g1.2.4 Pie chartA pie chart can also be used to represent the data. It represents the data in proportions of the degrees in a circle, at a glance one is able to tell which group has dominated the representation.

1.2.5 Scatter diagramsThey are used to represent a relationship between two independent variables in a graph

1.3 NUMERICAL1.3.1 Measures of central tendencyMeasures of central tendency are numbers that define the location of a distribution. For example if we regard all measurements as being attempts to give us the true value of a particular phenomenon, we can regard the center of the distribution of a set of measurements an estimate of that true value. They include the mean, mode and the median.1.3.1.1 The meanTo obtain it we add up the scores and divide by the number of scores. The more common term is the average, the more technical term, used here, is the mean. Two features of the mean should be mentioned .The first is technical: it is the point in a distribution about which the sum of the squared deviations is at a minimum. This makes it important for estimating variance, and for least squares analysis. The second feature is that the mean is a very effective statistic where scores within a distribution do not vary too much, but it is not so effective when there is great variance. Therefore it is important to know how much spread or variability there is in a set of scores in order to interpret the mean correctly. (Keith ,2003). It is the average of asset of scores or measurements. (Mugenda and Mugenda ,2003). According to (Kothari, 2011), mean also known as arithmetic average, is the most common measure of central tendency and may be defined as the value which we get by dividing the total of the values of various given items in a series by the total number of items. The researcher uses the mean as general value that represents the phenomena.

1.3.1.1 Medianit is the value of the middle item of series when it is arranged in ascending or descending order of magnitude. It divides the series into two halves; in one half all items are less than median, whereas in the other half all values have values higher than the median. (Kothari, 2011)In other words the median is the point below and above which 50% of the scores fall.The median can also used as the value that represents the data in the research.1.3.1.2 MODE it is the most commonly or frequently occurring value in a series. The mode in a distribution is the item around which there is maximum concentration. In general, mode is the size of the item which has the maximum frequency. Like median, mode is a positional average and is not affected by the values of the extreme items. It is therefore useful in all situations where we want to eliminate the effects of the extreme items. Mode is particularly useful in the study of popular sizes. E.g a manufacturer of shoes is usually interested in finding out the size most in demand so that he may manufacture a larger quantity of that size.1.3.4 Measures of variability/dispersionvariability is the distribution of scores around a particular central score or value. Purpose of variabilityMeasures of variability help the researcher to see how spread out the scores or measures of each variable are. (Mugenda and Mugenda, 2011). The various ways we can describe this spread is called the dispersion. These measures these measures quantify the variability of the distribution. They consist of:1.3.4.1 RangeRange is the simplest possible measure of dispersion and is defined as the difference between the values of the extreme items of a series.Thus range =highest value of an item in a series-lowest value of an item in a seriesExampleScores: 78,79,80,81,82,85The range is 85 78 = 7A big weakness of the range as a measure of variability is that it only involves two scores that is the highest and the lowest scores, it is therefore not sensitive to the total distribution.When the range is small then the items of the distribution in the phenomena are said to be homogenous. When the range is large then the items of the distribution are not uniform.1.3.4.2 Mean deviationIt is the average of difference of the values of items from some average of the series. Such a difference is technically described as deviation.1.3.4.3 Standard Deviation and the varianceThe standard deviation is defined as the extent to which scores in a distribution deviate from their mean or average. The standard deviation, therefore involves subtracting the mean from each score to obtain the deviation. If we square each deviation, sum the squared deviations and then divide this total by the degrees of freedom, we have a measure of variability called variance. If the value is small, it implies that the variance is small. This means that the scores are close together. If the value is large, it implies large variance and therefore the scores are more spread out.By taking the square root of the variance, one gets the standard deviation. The bigger the value derived by calculating the standard deviation, the larger the value derived by calculating the standard deviation, the larger the deviations from the mean denoting greater variability. A small standard deviation denotes less variability of scores in the distribution.1.3.4.4 Frequency distributions(Keith ,2003) Simple frequency distributions are a useful way to summarize and understand data. The individual scores in the distribution are tabulated according to how many respondents achieved each score or gave each response, or fell into each category. They also help the researcher to stay close to the data, at least in the initial stages of the analysis. There is great benefit to getting a hands-on feel of the data, especially when the availability of computerized programs makes it so easy for the researcher to be removed from the data. Mugenda and Mugenda (2004),in social science research, frequency may also refer to the number of subjects in a given category. For example, a frequency distribution of the variable marital statusScoresFrequency (f)

Single30

Married60

Divorced20

Separated10

n =20

Frequency distribution of marital status

Grouped FrequencyWhen intervals are given for each class interval, these are referred to as grouped frequency.The number of class intervals for each distribution depends on the sample size and the range of scores obtained. However, class intervals should be between 10 and 15 in number.Scores limit and exact limitsSome limits refer to situations where whole numbers are used to define the limits of class intervals.ExampleExample: NewspapersThese are the numbers of newspapers sold at a local shop over the last 10 days:22, 20, 18, 23, 20, 25, 22, 20, 18, 20Let us count how many of each number there is:Papers SoldFrequency

182

190

204

210

222

231

240

251

It is also possible to group the values. Here they are grouped in 5s:Papers SoldFrequency

15-192

20-247

25-291

Principles governing conversion of raw scores into grouped frequenciesi. All intervals should be the same width. Unequal class intervals cause problems when advanced statistical work is needed.ii. Intervals should be continuous throughout the distribution. That is even if there are no scores in a particular class interval, that class interval should be retained and a frequency of zero indicated against it.iii. Too few class interval lead loss of accuracy and too many class intervals result in inconveniences. Normally a class interval should range between 10 and 15 in number.

2. Inferential statisticsThe ultimate purpose of research is to be able to generalize the results from samples to populations. We use hypothesis testing technique to generalize from the sample to the population. These techniques are often referred to as inferential statistics. Inferential statistics are concerned with determining how likely it is for the results obtained from a sample to be similar to results expected from the entire population.a. Choice of a test of hypothesisThere are various types of statistical procedures that are used in testing the hypothesis and the choice of the procedures to use depends on the following factors;b. Size of the sampleIn testing hypothesis, some data analysis procedures cannot be used if the sample size is too small e.g regression.c. Type of variable and measurement scaleThe type of data analysis procedures used sometimes depends whether the variables are continuous or discrete. Similarly the measurement scale (ratio, interval, ordinal, nominal) will determine the procedure one should use to analyze datad. Types of research designStatistical data analysis procedures also differ depending on the research design. E.g data from an experimental study that compares differences between two or more groups is best analyzed using analysis of variance (ANOVA). Relationships and predictions among variables are best determined using correlation and regression techniques.Statistical procedures used in inferential statisticsIt tells us the direction and strength of relationships between variables both how the variables are related, and how much they are related.Simple correlationThe most widely-used type of correlation coefficient is Pearson r, also called linear or product- moment correlation. Pearson correlation assumes that the two variables are measured on at least interval scales and it determines the extent to which values of the two variables are "proportional" to each other. The value of correlation (i.e., correlation coefficient) does not depend on the specific measurement units used; for example, the correlation between height and weight will be identical regardless of whether inches and pounds, or centimeters and kilograms are used as measurement units. Proportional means linearly related; that is, the correlation is high if it can be "summarized" by a straight line (sloped upwards or downwards).Correlation coefficientThe computation of correlation coefficient yields a statistic that ranges from -1 to 1. This statistic is called a correlation coefficient . The correlation coefficient tells the researcher:1. The magnitude of the relationship between two variables. The bigger the coefficient (absolute value) the stronger the association between the two variables.2. The direction of the relationship between the two variables. If the correlation coefficient is positive (+), it means that there is a positive relationship between the two variables. A positive relationship means that as variable x1 increases, variable x2 increases as well or as variable x1 decreases, variable x2 decreases. A negative relationship (-) means that as variable x1 decreases, variable x2 increases and vice versa.The importance of correlation1. Correlation analysis takes one a step further by examining how various variables are related.2. Determining the strength and directions of the association between two variables is very important because this piece of information forms the basis for selecting variables for further statistical analysis e.g regression analysis.

RegressionRegression analysis is a type of analysis used when a researcher is interested in finding out whether an independent variable predicts a given dependent variable. Regression analysis could be categorized into:I. Simple regressionII. Multiple regressionSimple regression is used when the researcher is dealing with only one independent variable and one dependent variable.Example A researcher might be interested in finding out whether education predicts the financial status of households. In this example, education is the independent variable and financial status is the dependent variable.Multiple regressionsMultiple regression attempts to determine whether a group of variables together predict a given dependent variable.ExampleA researcher might be interested in finding out whether age, education, household size, and marital status influence the financial status of households. The four independent variables are considered together as predictors of financial status.Chi testThe chi-square (x2) is a statistical technique which attempts to establish relationship between two variables both of which are categorical in nature. For example, we may want to test the hypothesis that there is a relationship between gender and the number of road accidents caused by drivers. The variable gender is categorized as male and female. The variable number of accidents is categorized as none, few or many.The chi square technique is therefore a form of count occurring in two or more mutually exclusive categories. The technique compares the proportion observed in each category with what would be expected under the assumption of independence between the two variables. The Analysis of Variance ANOVAAnalysis of variance is a data analysis procedure that is used to determine whether there are significant differences between two or more groups or samples at selected probability level. The questions to be answered by analysis of variance are: what is the probability that the variation among a group of sample means has occurred as a result of randomly selecting the samples from a common population? Are the differences among the groups due to the treatments given or to chance?One way analysis of varianceThis refers to analysis of variance where groups are being compared on only one variable but at different levels. In other words there is only one independent variable that is measured either ordinal or nominal levels. The dependent variable is measured at either the ratio or interval scale.ExampleA researcher might be interested in finding out whether teaching methods influence performance in class. For this perpose, a class might be randomly divided into three groups and a different method of teaching used for each group.Group 1: Lecture methodGroup 2: Discussion methodGroup 3: Individual studyIn this example, the independent variable (teaching method) is measured at the nominal scale and the performance, dependent variable is measured at interval scale. The hypothesis being tested here is whether type of teaching method makes a difference in performance among students.Two way analysis of varianceOften, researchers are in interested in comparing two or more groups in more than one variable. For example, a researcher might want to compare mastery of the history content of standard eight pupils. Comparisons of mean scores for males and females and also performances of different schools could be included in the same study.schoolgender

Mean scoreMean score

Hospital hill6070

St George8080

Muthaiga Primary5070

Nairobi Primary6090

Two ways analysis of variance enables the researcher to make three types of comparisons. In the above example, the mean score of females and males are compared keeping the schools they come from constant. The same analysis of variance yields a comparison of the mean scores of subjects from different schools keeping the gender constant. We can also obtain more information by comparing subjects on the two variables, namely gender and schools, and determining whether the interaction between the two variables is statistically significant.t-test problemsIt is a special case of ANOVA which is used to test whether there are significant differences between two means derived from two samples or groups at a specified probability level. For example , a researcher might want to compare IQ performance from rural and urban children.F-ratioAnalysis of variance yields the F-statistics. The researcher must decide on the probability level desired to determine whether the calculated F-statistic is significant or not. It should also be noted that there are two types of degrees of freedom associated with the F-statistic. These are the numerator and the denominator degrees of freedom.ReportingThe ANOVA statistic is reported like this:The results shows that sucking lollipops significantly increases IQ of college men, F(3,17) = 2.87,p= .007.Where:F - the test statistic3 - the model degrees of freedom (numerator)17 - residual degrees of freedom (denominator)2.87 - value of F, the test statistic.007 - the probability of H0 being trueTests of significanceOnce a researcher has obtained the results from various data analysis procedures, he or she must decide whether the results are statistically significant. The F-test and the t-test are commonly used to evaluate the significance of various statistical indices. The regression coefficient bi , the chi-test x2 , the correlation coefficient r and the coefficient of determination R2 all must be tested for statistical significance. The researcher must select and apply the appropriate test of significance. The test of significance helps us to decide whether we can reject the null hypothesis. In general terms, a test of significance helps the researcher to determine whether obtained results truly hold at a given confidence level.

Steps of testing the hypothesisStep A: Null and alternative hypothesesThe first step of statistical testing is to convert the research question into null and alternative forms. We use the notation H0 to represent the null hypothesis and H1 (or Ha) to denote the alternative hypothesis. H0 is a statement of no difference. This is the hypothesis that the researcher hopes to reject. H1 opposes H0. We retain the premise of the null hypothesis until proved otherwise. This has a basis in [quasi-]deduction and is analogous to the presumption of innocence in a criminal trial.Step B: Error threshold (a) If we wish to reach a yes-or-no decision, fixed level testing must be pursued. To pursue fixed-level testing, we set an error threshold for the decision. The error threshold, called alpha (a), is the probability the researcher is willing to take of incorrectly rejecting a true H0. For example, the researcher may be willing to take a 1% chance of incorrectly rejecting a true H0.Step C: Test StatisticA test statistic is calculated. There are different test statistics depending on the data being tested and question being asked. In this chapter, we introduce tests of single means. For single means tests, the null hypothesis is H0: : = some value and the test statistic is either a zstat or tstat. These statistics are introduced below.Step D: ConclusionWe convert the test statistic to a p value by placing the test statistic on its appropriate probability distribution and determine the area under the curve beyond the test statistic. With fixed-level testing, the p value is compared to the " level and this simple decision rule is applied: With flexible significance testing, the p value answers the question: Thus, the smaller p value, the better the evidence against H0. As an initial rule-of-thumb we might say that we ought to take note of any p value approaching .05 (or less). In the parlance of statistics, such findings denote statistical significance.

THE ANALYSIS OF QUALITATIVE DATAData analysis refers to examining what has been collected in a survey or experiment and making deductions or inferences. It involves uncovering underlying structures; extracting important variables, detecting any anomalies and testing any underlying assumptions.It involves scrutinizing the acquired information and making inferences.DIVERSITY IN QUALITATIVE ANALYSISQualitative research concentrates on the study of social life in natural settings. Its richness and complexity mean that there are different ways of looking at and analyzing social life, and therefore multiple perspectives and practices in the analysis of qualitative data.There is variety in techniques because there are different questions to be addressed and different versions of social reality that can be elaborated (Coffey and Atkinson, 1996: 14)The variety and diversity in approaches underlies the point that there is no single right way to do qualitative data analysis- no single methodological framework. For example, Miles and Huberman (1994:9) suggest a fairly classic set of six moves common across different types of analysis. Likewise, Tesch (1990) while concluding that no characteristics are common to all types of analysis, nevertheless identifies ten principles and practices which hold true for most types of qualitative analysis.Much depends on the purposes of the research, and it is important that the method of analysis is integrated from the start with other parts of the research, rather than being an afterthought. The researcher should decide how to analyze data before going to the field as this will determine the recording technique that will be used during the data collecting exercise.Methods for the analysis of data need to be systematic, disciplined and able to be seen and described. A key question in assessing a piece of research is: how did the researcher get to these conclusions from these data? An answer should be given in order to have confidence in the findings put forward.The analysis will vary with the purposes of the research, the complexity of the research design and the extent to which conclusions can be reached easily (Orodho and Kombo, 2002: 116)ANALYTICAL INDUCTIONAnalytical induction was developed by Znaniecki (1934), and was originally identified with the search for universals in social life.Currently it is often used to refer to the systematic examination of similarities between cases to develop concepts or ideas.Hammersley and Atkinson (1995) describe it using the following steps; An initial definition of the phenomenon to be explained is formulated Some cases of this phenomenon are investigated, documenting potential explanatory features A hypothetical explanation is framed on the basis of the analysis of data, designed to identify common factors across the cases. Further cases are investigated to test the hypothesis. If the hypothesis does not fit the facts from these new cases, either the hypothesis is reformulated or the phenomenon to be explained is redefined so that negative cases are excluded. This procedure of examining cases, reformulating the hypothesis, and /or redefining the phenomenon is continued until new cases continually confirm the validity of the hypothesis, at which point it may be concluded that the hypothesis is correct, though not with absolute certainty.

A QUICK IMPRESSIONIST SUMMARYIn qualitative research, data can be analyzed by a quick impressionist summary. This involves: Summarizing key findings. E.g. in FGDs, the researcher notes down the different responses of the participants on various issues. Explanation. Interpretation and conclusion.This rapid data analysis technique is mainly used in situations that require urgent information to make decisions for a program for example in places where there is an outbreak in cholera and vital information is needed for intervention.This technique can also be used when the results already generated are obvious, making further analysis of data unwarranted. For example if a researcher finds out that 80% of respondents give similar answers to what caused a fire outbreak doing further analysis is unwarranted.This form of analysis does not require data transcription as the researcher records key issues of the discussions with respondents. A narrative report is written enriched with quotations from key informants and other respondents.

GROUNDED THEORY ANALYSISGrounded theory is a research strategy whose purpose is to generate theory inductively from data. This analysis aims directly at generating abstract theory to explain what is central in the data.At the heart of grounded theory analysis is coding: open coding, axial coding and selective codingOpen codingOpen coding constitutes a first level of conceptual analysis with the data. The analyst begins by breaking open the data that is opening up the theoretical possibilities in the data. The purpose is to use the data to generate abstract conceptual categories- more abstract than the data they describe- for later use in theory building. These are substantive codes- the initial conceptual categories in the data.Open coding involves a close examination of some of the data, identifying conceptual categories to account for the data being studied.Open coding is the part of analysis that pertains specifically to the naming and categorizing of phenomenon through close examination of data. At this point, data are broken down into discrete parts, closely examined, compared for similarities and differences, and questions are asked about the phenomena as reflected in the data.Axial (theoretical) codingAxial coding is the name given to the second stage, where the main categories which have emerged from open coding of the data are interconnected with each other. Open coding breaks the data apart, or runs the data open (Glaser, 1978), in order to expose their theoretical possibilities and categories then axial coding puts categories back together again, but in conceptually different ways. This is about inter-relating the substantive categories which open coding has developed.Strauss and Corbin (1990) write about the inter-actionist coding paradigm. This identifies causal conditions, phenomenon, context, intervening conditions action/ interaction strategies, and consequences as a way of interrelating categories in the data. Thus if the inter-actionist paradigm is used, the outcome of axial coding is an understanding of the central phenomenon in the data in terms of the conditions which give rise to it, the context to which it is embedded, the action/ interaction strategies by which it is handled, managed or carried out, and the consequences of those strategies.

Selective codingSelective coding builds on the propositions produced by axial coding. The objective here is to integrate and pull together the developing analysis.The theory to be developed must have a central focus around which it is integrated. This will be the core category of the theory and must be a central theme in the data, and should also be seen as central by the participants whose behavior is being studied.In order to integrate the other categories in the data, the core category will have to be at a higher level of abstraction. Selective coding will then aim at developing the abstract, condensed, integrated and grounded picture of the data. This helps move from a description point of view to a more conceptual category abstract enough to encompass what has been described.Limitation of grounded theoryIt faces a dilemma on how to be subjective, interpretive and scientific at the same time.

THEMATIC ANALYSISThemes refer to topics or major studies that come up in discussions. This analysis categorizes related topics where major concepts or themes are identified.In this form of analysis, the researcher does the following: Peruses the collected data and identifies information that is relevant to the research questions and objectives. Develops a coding system based on samples of collected data. Classifies major issues or topics covered. Rereads the text and highlights key quotations/ insights and interpretations. Indicates the major themes in the margins. Places the coded materials under the major themes or topics identified. All materials relevant to a topic are placed together. Develops a summary report identifying major themes and the association between them. Uses graphics and direct quotations to present the findings. Reports the intensity, which refers to the number of times certain words or phrases or descriptions are used in the discussion.The frequency with which an idea, word or description appears is used to interpret the importance, attention or emphasis.

Weakness: The thematic method tends to rely heavily on the judgment of a single analyst. This may lead to high levels of subjectivity and bias. It may be necessary to have two or more analysts to code the transcripts independently and compare notes.

NARRATIVE ANALYSIS AND MEANINGIn narrative analysis, form and content can be studied together, and a concern with narrative can illuminate how informants use language to convey particular meanings and experiences.Coffey and Atkinson (1996) show how analysis can explore participants use of imagery, and how such devices as metaphors reveal shared meanings and understandingsPeople use metaphors as a way of making sense of experiences, and of expressing and conveying its meaning. Qualitative analysts will often do the same thing in making sense of data.Metaphors are one important way of figurative language. They are a major type of literary device (trope), comparing two things using their similarities but ignoring their differences. Others can be irony (the view from the opposite or paradoxical), synecdoche (linking instances to a larger concept) and metonymy (representing a whole in terms of one of its parts).

CONTENT ANALYSISContent analysis examines the intensity with which certain words have been used. It systematically describes the form or content of written and/ or spoken material.In content analysis a classification system is developed to record the information. In interpreting results, the frequency with which a symbol or idea appears may be interpreted as a measure of importance, attention or emphasis. The relative balance of favorable attributes regarding a symbol or an idea may be interpreted as a measure of direction or bias. In content analysis, the first step is to select the data source to be studied, then develop a classification system to record the information. There are various forms of content analysis as follows:Pragmatic content analysis: classifies signs according to their probable causes and effects. The emphasis is on why something is said. This could be used to understand peoples perceptions and beliefs.Systematic content analysis: classifies signs according to meaning.Designation analysis: determines the frequency with which certain objects or persons, institutions or concepts are mentioned. This is a simple counting exercise.Attribution analysis: examines the frequency with which certain characterization or descriptions are used. The emphasis is on the adjectives, verbs and descriptive phrases and qualifiersAssertion analysis: provides the frequency with which certain objectives are characterized in a particular way. Such an analysis often takes the form of a matrix with objects as columns and descriptors as rows.

ETHNOMETHODOLOGY AND CONVERSATION ANALYSISThe fundamental assumption of ethno-methodology is that people within a culture have procedures for making sense of their daily life. The primary focus is on how central features of a culture, its shared meanings and social norms, are developed, maintained and changed, rather than on the content of those meanings and norms.Conversation analysis becomes a central concern, as ethno-methodologists seek to understand peoples methods for producing orderly social interaction.The general purpose of this study is to understand the social organization of ordinary, naturally occurring human conduct, in which talk is a primary vehicle for the production and intelligibility of human action. When talk is analyzed, verbatim transcripts of actual conversations are used.Silverman (1993: 125) gives account of three fundamental assumptions of conversation analysis. They concern the structural organization of talk, the sequential organization of talk, and the need for the empirical grounding of the analysis. Following these assumptions, conversation analysis studies the situated production and organization of talk, developing a bottom-up understanding of how context influences participants production of the social reality.Conversation analysis generates significant implications from the analysis of previously unnoticed interactional forms. Heath and Luff (1996: 324) conclude that the naturalistic analysis of conversation and interaction has developed a substantial body of findings which delineate the interlocking social organization of a wide range of ordinary social actions and activities.

DISCOURSE ANALYSISDiscourse refers to the general framework or perspective within which ideas are formulated. It focuses attention on the way language is used, what it is used for, and the social context in which it is used.Analysts see speech as a performance; it performs an action rather than describes a specific state of affairs or specific state of mind. Much of this analysis is intuitive and reflective, but may also involve some form of counting, such as counting instances of turn-taking and their influence on the conversation and the way in which people speak to others.Features of discourse analysisIt is concerned with talk and texts as social practices; and as such it pays close attention to features which traditionally would be classed as linguistic content- meanings and topics- as well as attending to features of linguistic form such as grammar and cohesion.It has a triple concern with action, construction and variability (Potter and Wetherell, 1987). People perform actions of different kinds through their talk and writing, and they accomplish the nature of these actions partly through constructing their discourse out of a range of styles, linguistic resources and rhetorical devices.It is concerned with the rhetorical or argumentative organization of talk and texts.

Discourse analysis is sensitive to how spoken and written language is used, and how accounts and descriptions are constructed. At the microscopic level, it shares much in common with conversation analysis. In a more macroscopic perspective, it emphasizes the interrelationships between accounts and hierarchies. At this level, it is similar to deconstruction, in dismantling constructed accounts to show connections with power and ideology.

SEMIOTICSSemiotics, or the science of signs, lays out assumptions, concepts and methods for the analysis of sign systems. Eco (1976) points out that semiotics is concerned with everything that can be taken as a sign. It is based squarely on language, in line with the view that human linguistic communication can be seen as a display of signs, or a text to be read.Semiotics can be used in the analysis of texts and also narrative structures. With its focus on linguistic structures and categories, it can be used to develop a theory of texts and their constituent elements. It gives deeper meaning in the system of rules that structures the text as a whole. It is this underlying structure and the rules it embodies that can tell the researcher what its cultural and social message is.

DOCUMENTARY AND TEXTUAL ANALYSISThere is richness in documentary data for social research. The analysis of such data shares characteristics with the approaches described but also has distinctive themes.One theme focuses on the social production of the document, starting with how the document came into being. All documentary sources are the result of human activity, produced on the basis of certain ideas, theories and principles.Documents and texts studied in isolation from their social context are deprived of their real meaning. Thus an understanding on the social production and context of the document affects its interpretation.A second related theme is the social organization of the document. It asks questions such as, How are documents written? For what purposes? What is recorded/ what is omitted? What does the writer seem to take for granted about the readers? What do readers need to know in order to make sense of them?These questions are used to study the social organization of documents, irrespective of their truth or error.A third theme concerns the more direct analysis meaning of text for meaning, this time including questions of truth and error. It can focus on the surface or literal meaning, or on the deeper meaning. Methods used range from interpretive understanding to more structural approaches.A fourth theme would be the application of different theoretical perspectives to the analysis of texts and documents. This can also incorporate deconstruction a s used in discourse analysis.

COMPARATIVE ANALYSISUsing this method, data from different people is compared and contrasted and the process continues until the researcher is satisfied that no new issues are arising. The researcher moves backwards and forwards between transcripts, memos, notes and the research literature.Comparative analysis examines similarities and differences in events during different time periods.

THEORETICAL AND PHILOSOPHICAL ANALYSISThis utilizes historical parallels, past trends, and sequences of events to suggest the past, present and future of the topic being researched.Findings would be used to develop a theory or philosophy of leisure. For example, an analysis of public recreation agency goals and objectives of previous eras can be used to describe the future in the context of social, political, economic, technological, and cultural changes in society.

CHALLENGES FACED IN DATA ANALYSISThe researcher should ensure the following: Understand the assumptions of their statistical procedures. Be sure to use the best measurement tools available. If measures have errors, then that fact should be considered. Beware of multiple comparisons. If one has to do many tests, replacement or cross- validation should be done to verify the results. Keep in mind what one is trying to discover. One should look at the magnitude rather than the values. Use numerical notation in a rational way. One should not confuse precision with accuracy. Be sure to understand the conditions for causal inferences. If one needs to make inference, then he/she should try to use random assignment. Be sure the graphs are accurate and reflect the data variation clearly.

ETHICAL ISSUESIn data analysis, a researcher should maintain integrity. This is particularly in the application of statistical skills to problems where private interests may inappropriately affect the development or application of statistical knowledge. For these reasons, researchers should: Present their findings and interpretations honestly and objectively. Avoid untrue, deceptive, or doctored results. Disclose any financial or other interests that may affect, or appear to affect their analysis. Delineate the boundaries of the inquiry as well as the boundaries of the statistical inferences which can be derived from it. Make the data available for analysis by other responsible parties with appropriate safeguards for privacy concerns. Recognize that selection of a statistical procedure may to some extent be a matter of personal judgment and that other statisticians may select alternative procedures. Direct any criticism of a statistical inquiry to the inquiry itself and not to the individuals conducting it. Apply statistical procedures without concern for a favorable outcome.CONCLUSIONIn data analysis, the researcher has, according to Cohen (1993) to be sure of the following: Be sure the analysis sample is representative of the population in which the researcher is interested with Be sure to understand the assumptions of statistical procedures, and be sure they are clearly defined. Be sure to use the best measurement tools available. If measures have errors, then that fact should be taken into account. Be clear of what he is trying to discover. Be sure the graphs are accurate and reflect the data variation clearly.

REFERENCESDawson, C. (2002) Practical Research Methods:A User- friendly Guide to Mastering Research. Oxford: How to Books Ltd.Kombo, D.A. and Tromp, D. L. A. (2006). Proposal and Thesis Writing: An Introduction. Nairobi: Paulines Publications Africa.

Orodho, A. J. and Kombo, D. K. (2002). Research Methods. Nairobi: Kenyatta University, Institute of Open Learning.

Punch, K. F. (1998). Introduction to Social Research: Quantitative and Qualitative Approaches. London: Sage Publications Ltd.

Olive M Mugenda & Abel G Mugenda(2003). Research Methods: Quantitative and Qualitative Approaches. Nairobi: Laba Graphic Services Ltd.C.R Kothari(2011). Research Methods: Methods and Techniques. India: new age international (p) Ltd.

23