Please enable JavaScript.
Coggle requires JavaScript to display documents.
KMF6013 Statistics for Social Sciences - Coggle Diagram
KMF6013 Statistics for Social Sciences
Unit 4: Descriptive statistics
Central tendency (3M):
Mean, Mode, Median
Dispersion:
Range: :check: difference between the highest and lowest number (maximum)
Variance: :check: How far the number from the mean
Std Deviation: Average :check:number differs from the mean (Square root of variance)
Normal Distribution
Standard normal distribution called= "standard score" or "z score"
Skewness and Kurtosis
:star: Descriptive statistics is to describe or summarize data in meaningful ways
Consist of collection, organization, summarization, & presentation of data e.g pie chart
Pair Sample T-test
Techniques M. A. S. R: :check: Mapping :check: Assumptions :check: Step by step :check: Reporting
Mapping : :check: normality assumption (Paired sample t-test)
:check: Non parametric
alternative (Wilcoxon Signed Rank test)
:!!: Type of variable : Independent variable (IV)
:check: p = .000 < 0.05
:check: p-value is
SIGNIFICANT
:check: Fail to Reject Ha
To compare the differences between mean of fear of statistics time 1 & time 2
Unit 3: Reliability Test
:check: Reliability relates to the quality of measurement
:check: Reliability is the "consistency" or "repeatability" of the study measure
Factor analysis
:star: Exploratory factor analysis & :star: confirmatory factor analysis
Factor extraction: Determining the smallest number of factors to represents the interrelationships among sets of variables.
:check: Kaiser's criterion
:check: Scree test
:check: Parallel analysis
3 main step involved in factor analysis:1) Assessment of the suitability of the data 2) Factor extraction : :check: Kaiser's criterion :check: Scree test :check: Parallel analysis 3) Factor rotation
:!: Strength of relationship: a) Correlation matrix : Greater than .3, b) KMO- Kaiser Meyer Olkin : : .6 and above, c) Bartlett's test : Significant (p < .05)
Factor rotation and interpretation: Orthogonal (varimax method) -minimize the number of variables that have high loadings on each factors
Normality test
Parametric test :check: ANOVA :check: T-test :check: Linear regression
Criteria :
a. Skewness and Kurtosis Z-value Should be in the span of (-1.96 to +1.96) Skewness/ Kurtosis Formula, Z-value = statistic / Std error
d. Stem and Leaf Should visually indicate that your data are approximately normally distributed (normal curve)
e. Normal Q-Q plot Should visually indicate that your data are approximately normally distributed (Dot along the line)
f. Box Plot Should visually indicate that your data are approximately normally distributed (Symmetrical)
b. Kolmogrov Smirnov/Shapiro Wilk p-value = .05 and above
c. Histogram Should visually indicate that your data are approximately normally distributed (normal curve)
ANOVA Test
Techniques M. A. S. R: :check: Mapping :check: Assumptions :check: Step by step :check: Reporting
Mapping : :check: Normality assumption (One Way ANOVA) :check: Non parametric alternative (Kruskal Wallis)
:!!: Type of variable : Independent variable (IV) & Dependent variable (DV)
:!!: Quick guide: :check: p-value <= (critical value) : Reject the null hypothesis of the statistical test
:check: p-value > (critical value) : Fail to reject the null hypothesis of the statistical test
To compare the differences QoL score by Gender.
:check:IV: Nominal/Ordinal
DV: Interval/Ratio
:check: p=.151>.05
:check: p value is NOT SIGNIFICANT
:check: Fail to Reject Ho
Independent T-Test
Mapping: :check: Normality assumption (Independent T-test) :check: Non parametric alternative (Mann-Whitney) :check: Type of variable : Independent variable (IV) & Dependent variable (DV) :check: IV: Nominal/Ordinal
DV: Interval/Ratio
:!!: Quick guide: Same with ANOVA
To compare the differences QoL score by Gender.
:check: p value is NOT
SIGNIFICANT
:check: Fail to Reject Ho
1) An independent t-test was run to determine if there were differences in QoL score by gender.
Effect Size: :star: Weak (.03)
:star: Moderate (.06)
:star: High (.12)
3 p value [purpose]:a. :star: a) Normality of distribution
Kolmogrov Smirnow
p>.05 (Consider data met the normality requirement) -OK
:star: b. Homogeneity
Levene Test
p>.05 ( Equal Variance Assume) -OK
:star: c. Significant Level
p>.05 (Not Significant) = NO differences
p<.05 (Significant) =Differences
Two Way ANOVA
Mapping: :check: Normality assumption (Two Way ANOVA) :check: Non parametric alternative (Kruskal Wallis) :!!:Type of variable : Independent variable (IV) & Dependent variable (DV)
:!!: Quick guide : Same with ANOVA test
To compare the differences QoL score by type of disabilities & Gender
:check: p=.355>0.05
:check:p value is NOT
SIGNIFICANT
:check: Fail to Reject Ho
:check: IV: Nominal/Ordinal
DV: Interval/Ratio
Type of p:
:check: Normality
(Kolmogrov Smirnow) = p>.05 [DATA NORMAL ]
Your data is normally distributed
:check:Levene’s Test
= p>.05
Your data is homogeneous
:check: Significant Level
a) p>.05
No significant difference between group
b) p<.05
Significant differences between group
A one way between groups ANOVA was conducted to examine the differences on QoL Score by type of disabilities.
Correlation test
5 steps in correlation: :red_flag: State Ho & Ha :red_flag: Set confidence level (a) :red_flag: Report r and sig-r :red_flag: decision :red_flag: conclusion
Guilford rule of thumbs r= 1< r< 1
Multiple Linear Regression (MLR)
:star:Determine relationship between one or more IVs and one DV
:star:Predict value of the dependent variable (Y) based on value of independent variables (X’s)
:star:Regression analysis is to derive prediction equation
:fire: MLR(ENTER) - What are the factors contribute to the Subjective Career Success?
:fire: MLR(STEPWISE) - Which predictor variables significantly contributed towards the Subjective Career Success ?
a= .05 (significant) If alpha is lower than .05 (the significant is good)
Tools of analysis
:red_flag: F-test :red_flag: T-test :red_flag: Regression
Inferential statistics tools
hypothesis tests, confidence intervals, and regression analysis
Chi- square test (for association)
Determines whether there is an association between categorical variables (e.g. variables or independent variables)
Two categorical variables
Two or more group
Correlation & Regression: :check: Correlation is a statistical method used to determine whether a linear relationship between variables exists.
:check: Regression is a statistical method used to describe the nature of the relationship between variables—that is, positive or negative, linear or nonlinear.
:red_flag:Non- parametric test: 1) Mann-Whitney (T-test) - test significant differences ONLY 2 groups 2) Kruskal- Wallis (ANOVA) - test significant for MORE than 2 groups