45
Statistical Tools in Evaluation Part I

Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Embed Size (px)

Citation preview

Page 1: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Statistical Tools in Evaluation

Part I

Page 2: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Statistical Tools in Evaluation

• What are statistics?– Organization and analysis of numerical data

– Methods used involve calculations and graphical displays of data

– Formulas used can reveal the “true” nature of the data as well as critical relationships between variables (targets of study)

Page 3: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Statistical Tools in Evaluation

• Why Use Statistics?– Analyze and interpret data

– Standardize test scores

– Interpret research in your field

Page 4: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Problem:

Not all scoring / quantifying systems are the same.

Vary by:

Scores

Scales

Page 5: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Types of Scores

Continuous Scores that can be recorded in an infinite number of

values (decimal figures; greater and greater accuracy)

Examples: time, distance

Page 6: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Types of Scores

Discrete Scores that are whole numbers only

Examples: wins, losses, home runs, touchdowns

Page 7: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Types of Scales

• Nominal Scale– Lowest and most elementary scale

– Generally represents categories

– Something is in a category or it is not

– Examples: sex, state of origin, eye color

Page 8: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Types of Scales

• Ordinal Scale (order)– Generally refers to rank or order of a variable

– Does not tell how big or small the difference between ranks is

– Examples: • finish order in a race – 1st,2nd,3rd

• tennis team ladder of “best to worst”

• season ranking of a team

Page 9: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Types of Scales

• Interval Scale– Also provides order of variable, but additionally

provides information about how far one measure is from another

– Equal units of measure are used on the scale

– No true zero point that means absence

– Examples: temperature, year, IQ

Page 10: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Types of Scales

• Ratio Scale– Same as interval, but has a true zero point (absolute

absence or completely nothing)

– Examples: height, weight, time - *type of score?

Page 11: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Once you have scores (data) what is the first thing you do with them?

• Find out how they are distributed

Page 12: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Distribution of Data

Page 13: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Simple Score Ranking

• List scores in descending or ascending order depending on quality*

• Number scores from best – first, to worst – last

• Identical scores should have the same rank

– average the rank– or determine midpoint and assign same rank

Page 14: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Example of Simple Ranking

Page 15: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Frequency Distribution• Once data have been collected (numbers given

to a measurement), it is best to organize them in a sensible order

– Best at top of list

• highest to lowest – jump height, throw dist.

• lowest to highest – swim time, golf score

– Calculate frequencies of scores – how many of each score are present

Page 16: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Frequency Distribution

• Frequency distribution can tell:– frequency of a score (f) – how many of each score

– cumulative frequency (cf) – how many through that score

– cumulative percentage (c%) - % occurring above and below a score

Page 17: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Examples and Practice Problems

Page 18: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Graphing the Frequency Distribution

• Frequency of scores on y axis (ordinate)

• Scores from low to high on x axis (abscissa)

• Intersection of ordinate and abscissa is zero (0) point for both axes

0 Scores

Fre

quen

cy

Page 19: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Graphing the Frequency Distribution

• Frequency Polygon– Midpoints of intervals are plotted against

frequencies

– Straight lines drawn between points

• Histogram– Bars are used to represent the frequencies of scores

• Curve– Curved line represents the frequency of scores

Page 20: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

What else can grouped scores tell us?

How all scores compare to the average score

= Measures of Central Tendency

Page 21: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Measures of Central Tendency

• Statistics that describe middle characteristics of scores– Mode (Mo) The most frequently occurring score

• There can be more than one mode - bimodal

– Determination: Find the score that occurs most frequently !

Page 22: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Measures of Central Tendency

– Median - Median (Mdn, P50) - represents the exact middle of a distribution (50th percentile)

• The Mdn is the best measure of central tendency when you have extreme scores and skewed distributions

Page 23: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Median

• Median Calculations:– Determining position of approximate

median:• “Simple counting method”

• Formula - Mdn = (n + 1) / 2

(n = total number)

Page 24: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

• Ranks tell the position of a score relative to other scores in a group.

Percentile Rank- The percentage of total scores that fall below a given score.

Percentile - refers to a point in a distribution of scores in which a given percent of the scores fall (percentile is the location of the score).

25th percentile (quartile), 75th percentile, 90th percentile, etc.

Page 25: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Examples and Practice Problems

Page 26: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Measures of Central Tendency

mean (X): average score• most sensitive• affected by extreme scores• best for interval and ratio scale• probably most often used

Page 27: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Measures of Central Tendency

mean (X): average score.• most sensitive• affected by extreme scores• best for interval and ratio scale• probably most often used

– Calculation:X = X / n ( = sum; X = sum of scores)

Page 28: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Remember Curves?

• What types of curves are there and what do they mean?

– Normal curve

– Skewed curve

Page 29: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Characteristics of the Normal Curve

• Bell-shaped

• Symmetrical

• Greatest number of scores found in middle

• Mean, median, and mode at same point in the middle of the curve.

Page 30: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Characteristics of the “not-so-normal curve”

• Irregular curves represent different types of distributions

leptokurtic platykurtic

bimodal

positive skew negative skew

Page 31: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

• Normal curve - X, Mdn, and Mo are all the same value (location)

X

Mo

Mdn

Page 32: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

• Skewed curves - Mo is opposite end of the tail, Mdn is in the middle, and X is toward the tail

Mo Mdn X

Positive Skew

Page 33: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

• Skewed curves - Mo is opposite end of the tail, Mdn is in the middle, and X is toward the tail

MoMdnX

Negative Skew

Page 34: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

• Question: Why do these variables fall this way on a skewed distribution of scores?

• Question: Can you see the impact of extreme scores on these variables?

Page 35: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Measures of Variability

• Variability refers to how much individual scores deviate from a measure of central tendency; how heterogeneous the group is.

Page 36: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Measures of Variability

• Range (R) - Represents the difference between the low and high score.

• Simplest measure of variability; used with the mode or median.

• Calculation: R = High – Low

Page 37: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Measures of Variability

• Standard Deviation (SD, s) - Describes how far the scores as a group deviate from the X.

• It is the most useful descriptive statistic of variability.

Page 38: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

• SD calculations:

SD = (X - X )2

N

Page 39: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Examples and Practice Problems

Page 40: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Relationship Between Normal Curve and SD:

1 SD = 68.26% of all scores (34.13% above and below X)

2 SD = 95.44% of all scores (47.72% above and below X)

3 SD = 99.73% of all scores (49.86% above and below X)

Page 41: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used
Page 42: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

How Alike are Scores in a Normal Curve?

• Homogeneity = Near the mean - alike

• Heterogeneity = Away from the mean - different

Page 43: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used
Page 44: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

Questions?

Page 45: Statistical Tools in Evaluation Part I. Statistical Tools in Evaluation What are statistics? –Organization and analysis of numerical data –Methods used

End Part I