DATA ANALYSIS

Introduction

A body of methods that help to describe facts,

detect patterns, develop explanations, and test

hypothesis.

Stages of data analysis

Coding Data entry Descriptive analysis Cross-tabulation

Testing relationships between variables

Coding

Summarised in numbers or letters for easy identification

Example of questionnaire
1. Gender
1 2 Male Female

4. Frequency of doing reference at library

1 2 3 Once a week 2-3 times a week 4-6 times a week Almost every day

2. Age___________ 3. Monthly expenditure

1 2 3 4 5 6 < RM 50 RM 50 - RM 100 RM 101 - RM 150 RM 151 - RM 200 RM 201 - RM 300 > RM 300

5. CGPA achievement__________________ 6. Do you participate in sports?

1 2 Yes No

Data entry

coded data is entered into computer

checked
cleaned/edited

Variable view

Measurement scale: nominal, ordinal and ratio

1. Jantina
1 2 Lelaki Perempuan

4. Kekerapan membuat rujukan di perpustakaan

1 2 Sekali seminggu 2-3 kali seminggu 4-6 kali seminggu Hampir setiap hari

NOMINAL

3 4

2. Umur___________
RATIO 3. Perbelanjaan bulanan
1 2 3 4 5 6 < RM 50 RM 50 - RM 100 RM 101 - RM 150

ORDINAL

5. Pencapaian CGPA__________________ RATIO 6. Aktif dalam aktiviti sukan

1 Ya Tidak

RM 151 - RM 200 2 RM 201 - RM 300 > RM 300

NOMINAL ORDINAL

Strategy for data analysis

QUALITATIVE
Percentage, mode, median, charts and tables (No measure for variation) Cross-tabulation, nonparametric measure of association Non parametric methods FREQUENCY DISTRIBUTION FOR EVERY VARIABLE DESCRIPTIVE STATISTICS MEASURES OF ASSOCIATION HYPOTHESIS TESTING/ MODELLING ESTIMATION, PREDICTION, FORECASTING Parametric methods QUANTITATIVE Mean, median, variance, graphs and many more Correlation Analysis

Descriptive analysis

frequency distribution
nominal and ordinal frequencies ratio/scale descriptives

mean mode median

Cross-tabulation

relationship between 2 or more variables

use 3M method

3M to select suitable analysis procedure

model
identify independent and dependent variables measurement

identify level of measurement: norminal, ordinal, ratio method

Model
Frequency of doing reference at the library

CGPA

INDEPENDENT ORDINAL

DEPENDENT RATIO / SCALE

Measurement scale
RATIO Has a meaningful zero point (physical) INTERVAL
Height Weight Time Speed

Predetermine equal intervals, no true zero point

ORDINAL Classifies and ranks objects NOMINAL Classifies objects into categories

Scores-achievement, aptitude, IQ test Temperature

Opinion Academic level Income level Liking

True Categories Artificial Categories

Taste Gender Ethnicity Color Area Opinion

Multivariate Analysis
INDEPENDENT DEPENDENT Nominal Ordinal Ratio / Scale

Nominal

CROSSTABS

CROSSTABS

CROSSTABS MEAN DESCRIMINANT

Ordinal

CROSSTABS

CROSSTABS

CROSSTABS NONPAR CORR DESCRIMINANT

Ratio/ scale

MEANS ANOVA

MEANS ANOVA

CORRELATION GRAPH

MANOVA

MANOVA

PARTIAL CORR
MANOVA REGRESSION

Analysis procedure for SPSS

ANOVA: Analyse-General Linear Model-Univariate CORRELATE: Analyse-Correlate-Bivariate CROSSTABS: Analyse-Descriptive Statistics-Crosstabs DESCRIPTIVES: Analyse-Descriptive Statistics-Descriptive DESCRIMINANT: Analyse-Classify-Descriminant FREQUENCIES: Analyse-Descriptive Statistics-Frequencies MEAN: Analyse-compare mean-mean MANOVA: Analyse-General Linear Model-Multivariate SPEARMAN CORRELATION: Analyse-Correlate Bivariate (Method=Spearman)

REGRESSION: Analyse regression-Linear (Method-Backward)

RECODE: Transform-Recode-Into Different Variable

Testing relationships between 2 variables

Study relationship to determine indicator of the study concept Significant relationships Degree/strengthM of relationship

Statistics that asses the strength of the relationship between variables or the amount of variation in the dependent variable that can be explained by the independent variable.

 Statistics that allow the researcher to determine or not a relationship is statistically significant (basically, an assessment or whether it is real or or could have occurred by chance)
(Chadwick, 1984)

Depends on

Strength of relationship & Sample size

Statistical terms
ANALYSIS OF VARIANCE (ANOVA) A procedure for deciding if a ratio level dependent variable is significantly associate with the nominal or dominant independent CHI-SQUARE TEST A test of statistical significance associate with contingency table analysis, where the dependent variable is a nominal one CORRELATION ANALYSIS A procedure for measuring how closely two ratio level variable co-vary together

CORRELATION COEFFICIENT (r) A measure of the strength of association between two variable, a correlation may vary from + 1 to -1 DEGREE OF FREEDOM Degree of freedom (df)= (Number of rows- 1) x (Number of column-1)
DEPENDENT VARIABLE A variable that is viewed as being influenced by other variable. Is the effect in a cause-effect relationship INDEPENDENT VARIABLE This variable is a cause in a cause effect relationship. It is a variable which has been selected as a possible influence on variations in a dependent variable

MEAN The mean is computed by summing the values of a variable and dividing the result by the total number of cases
MEDIAN The median represents the mid-point of a distribution MODE The most frequent occurring responses to a nominal variable MULTIPLE REGRESSION ANALYSIS An analysis that identifies an equation to predict variations in an a dependent variables MULTIVARIATE ANALYSIS OF VARIENCE (MANOVA) MANOVA is used in examining relations between a ratio level dependent variable and dependent variables measured at any level. It also permit the researcher to examine multiple dependent variables simultaneously.

NOMINAL MEASUREMENT A quantitative measure where the numbers are arbitrarily assigned to categories of the variable ORDINAL MEASUREMENT Measure where there is an order in the values assigned, but where the intervals between the values are not equal RATIO MEASUREMENT A qualitative measurement where intervals are equal and there is a true zero point. REGRESSION ANALYSIS A method for analyzing the relation between a ratio level dependent variable and independent variables. This form of analysis provides weightings that may be used in an equation to describe the relationship, standardized weightings provide a means for estimating the relative impact of independent variables on the dependent one REGRESSION LINE A straight line describing the relation between an independent and dependent variable drawn so that the vertical deviations of the points above the line equal the vertical deviations below the line.

SPEARMAN CORRELATION The statistic is used to measure the strength of association between two ordinal level variables PEARSON CORRELATION This statistics is used to measure the strength of association between two ratio level variables

STANDARD DEVIATION A measure that reflects the average amount of deviation from the mean value of the variable
TEST OF SIGNIFICANT A test reporting the probability that an observed association or difference is the result of sampling fluctuations, and not reflective of some real difference in the population from which the sample has been taken. VARIABLES Those concepts that we intend to measure VARIANCE Variance reflects the average amount of deviation from the mean value of the variable, it is the standard deviation squared

Technical Report

Title page Acknowledgement TOC Abstract Introduction Methodology Results and analysis

Discussion of findings

Recommendations
References Appendix

THE END