Transcript Types of Statistics

Types of Statistics
Descriptive
Means
Medians
Modes
Percentages
Variation
Distributions
Inferential
Draws conclusions
Assigns confidence to conclusions
Allows probability calculations
Wang, Schembri and Hall JMBE 14:12-24 (2013)
FIGURE 5. Student performance in
(A) midsemester and (B) final
exams across 2010 (n = 265) and
2011 (n = 264) offerings of
MICR2000.
FIGURE 6. Student Evaluation of Course
and Teaching (SECaT) scores across
2010 and 2011 offerings of MICR2000.
Students were invited to voluntarily
respond to surveys regarding their
evaluation of teaching within
MICR2000 in 2010 (n = 108) and 2011
(n = 87) using a standardized
University-Wide Student Evaluation of
Course and Teaching (SECaT) survey
instrument. Student responses
corresponded to a 5 -point Likert scale
and quantified as follows: 1 = Strongly
Disagree; 2 = Disagree; 3 = Neutral; 4 =
Agree; 5 = Strongly Agree. Bars
represent mean +/– standard error of
the mean (SEM). *Denotes a
statistically significant difference
between student responses for 2010
and 2011 offerings of MICR2000, as
determined by the Mann-Whitney U
test (p < 0.05).
Wang, Schembri and Hall JMBE 14:12-24 (2013)
Three Kinds of Data
Nominal
Categorical
No mean
ex:
● Marriage status
● Gender
Sounds like “NAME”
Ordinal
Natural ordering
Unequal intervals
ex:
● Rankings
● Survey data
Sounds like “ORDER”
Interval
Extends ordinal data
Equal intervals
ex:
● Temperature
● Time
Sounds like what it is
Borgon et al., JMBE 13:35-46 (2013)
Hurney JMBE 13:133-141 (2012)
Boone and Boone Journal of Extension 50:2TOT2 (April 2012)
Darland and Carmichael JMBE 13:125-132 (2012)
Problem (Theory)
Question (Hypothesis)
Methods (treatment, control groups)
Intervention
Data (Triangulation)
Conclusions
Change practice
One category
Frequency, %,
Goodness-of-fit, 𝑥 2
Two categories
Frequency, %,
Contingency table, Test
of Association, 𝑥 2
Nominal or Ordinal
(Qualitative)
Continuous
One
Type of
Data
Relationships
Ranks
Multiple
Pearson
Correlation
Form of
Relationship
Linear
Regression
Primary
Interest
Measurement
Number of
Predictors
Degree of
Relationship
Spearman’s
rS
Multiple
Regression
Independent
samples t
Independent
Interval
(Quantitative)
Type of
Question
Two
MannWhitney U
Relation
Between
Groups
Paired
Samples t
Dependent
Differences
Wilcoxon
Number of
Groups
One
Independent
Multiple
Adapted from D.C. Howell, Fundamental Statistics for the Behavioral
Sciences (6th ed.) Wadsworth Cengage Learning (2008)
One-Way
ANOVA
KruskalWallis
Number of
Indep. Var.
Multiple
Relation
Between
Groups
Dependent
Repeated
Measures
ANOVA
Friedman
Factorial
ANOVA
1. Collect student demographic data
a) Want to discover if students between treatment
and control groups had the similar ethnic
backgrounds
2. Collect test grades before and after intervention
a) Want to see if your teaching intervention resulted
in a significant difference in test scores between
control and treated groups
3. Survey students on their own perceptions of learning
a) Want to see if your teaching intervention resulted
in a significant increase among responses to
Likert-scale questions regarding student learning
gains between control and treated groups
Graduate school level: You have categorized your
students into three performance groups; novice,
developing, and expert based on high school GPA
and SAT data. You want to compare the
performance of these groups on a critical
thinking assessment before and after your
teaching intervention.